Working Document No. 116 "'-A.I00~W~ [p[f@®[(~WO íJ ®®Z70íJ ®®ffJ~-, CC~~D COl ECC/ON HISTORlCA Oclober 1992 CC6LATI Centro Internacional de Agricultura Tropical ¡ j CONTENTS EXECUTIVE SUMMARY INTRODUCTION 1. THE CIAT CASSAVA PROGRAM 1.1 History of CIA Ts Cassava Program 1.2 Cassava Program 1987-1991 1.3 Program Organization 1.4 Human and Financial Resources HEADQUARTERS BASED RESEARCH 2. GERMPLASM MANAGEMENT AND IMPROVEMENT 2.1 Breeding Strategies and Technical Advances 2.2 National Program Col1aboration 2.3 Futura Perspectivas for Germplasm Development 3. PHYSIOLOGY 3.1 Yield Stability During Prolonged Midseason Water Stress 3.2 Screening for Drought Tolerance and Low HCN Content 3.3 Potential Photosynthesis of Cassava and its Aelation to Productivity 4. QUALlTY 4.1 Objectives 4.2 Research Priorities 4.3 Methodologies 4.4 Results 4.5 Future Priorities 5. BIOTECHNOLOGY 5.1 Analysis and Conservation of Genetic Diversity in Manihot 5.2 Crop Productivity Research 5.3 Institution Building 1 2 5 7 11 15 16 35 41 43 43 49 55 65 65 65 67 70 82 83 B4 87 94 6. PATHOLOGY 97 6.1 Etiological Studies 97 6.2 Epidemiological Studies 99 6.3 Witches' Broom (WB) Diseasa 99 6.4 Superelongation Disease 102 6.5 Cultural Practiees to Control CBB 104 6.6 Root Rot Problems 104 6.7 Biological Control of Cassava Diseases 109 6.8 Cassava Endophytes 116 6.9 New Method for Interehanging Indexed Vegetative Planting Material of Cassava 120 6.10 Storage of vegetative Planting Material 120 6.11 Geographie Distribution and Potential Risk of Six Cassava Diseases 122 7. VIROLOGY 125 7.1 Frogskin and Caribbean Mosaie Diseases 125 7.2 Cassava Common Mosaie Virus 131 7.3 Cassava Vein Mosaie Virus 132 7.4 Afriean and Indian Cassava Mosaie Viruses 133 7.5 Latent Viruses of Cassava in Latin America 134 7.6 Conclusions and Future Research Objectlves 135 8. ENTOMOLOGY ANO ACAROLOGY 137 8.1 Mites 138 8.2 The Cassava Hornworm Erínnyis ello (L) 149 8.3 Cassava Mealybug 153 8.4 Cassava Root Mealybug: Pseudococcus mandlo 159 8.5 The Cassava Burrowing Bug 160 8.6 Whiteflies 163 8.7 Inseets Attacking Sexual Seed 166 8.8 Insects Attacking Drled Cassava 166 9. CROPPING SYSTEMS 169 9.1 Intereropping Cassava and Maize 170 9.2 Intereropping Cassava and Cowpeas 182 9.3 Technology Validation with Farmers 184 10. PLANT NUTRITION ANO SOIL MANAGEMENT 10.1 Screening Germplasm for Adaptation to Low-P Soils (Santander de Quilichao) 10.2 Relationships Between Leaf Gas Exchange, Total Biomass, RY and Adaptation to Low-P Soils 10.3 Long-Term Response to NPK Fertilizer in Infertile Acld Soil (Santander de Quílichao) 10.4 Response to NPK Fertilizer in Sandy SOU (Media Luna) 10.5 SoU Management Systems at Media Luna 10.6 Verietal Oifferences in Response to P 11. SOIL CONSERVATION ANO PROOUCTIVITY 11.1 1987-1989 Trials 11.2 Potential of Forage Legumes as Ground Caver in Relation to Soil Conservation and Cassava ProductMty 11.3 Soil Losses and Runoff in Cassava-Legume Association vs. Other Practices 11.4 Productivity in Cassava-Legume Association VS. Other Practices 11.5 SoU Properties as Affected by SoU Erosion and Cassava Cropping Practices 11.6 Conclusions 12. PROCESS ANO PROOUCT OEVELOPMENT 12.1 Objective 12.2 Methodologies 12.3 Priorities 12.4 Oried Cassava tor Animal Feed 12.5 Fresh Cassava Conservation 12.6 Cassava Flour tor Human Consumption 12.7 Cassava Starch 12.8 Future Activíties REGIONAL COLLABORATION 13. REGIONAL COLLABORATION IN LATIN AMERICA 13.1 Modes of Coflaboration 13.2 Training and Networks 13.3 Country Reports 193 193 195 197 197 200 203 209 210 215 215 220 221 221 223 223 223 224 225 229 234 240 247 249 249 252 254 14. COLOMBIA 259 14.1 DRI/CIAT Integrated Project. Atlantic Coast of Colombia 259 14.2 PNR/CIAT Collaborative Project 263 14.3 Cauca Starch Praject 269 14.4 Other Regions of Colombia 270 14.5 Impact 271 15. ECUADOR INTEGRATED CASSAVA PROJECT 273 15.1 Cassava in Ecuador 273 15.2 Principies of the Ecuador Intagrated Cassava Projact 274 15.3 The Cassava Integrated Project in Manabr Provinca 275 15.4 Grawth and Evolution of UAPPY Production and Marketing 278 15.5 Differences Among APPYs: Results of Monitoring Studies 283 15.6 Training 287 15.7 Integratad Cassava Project Rasearch: 1990-91 Highlights 289 15.8 Future Directions 294 16. BRAZIL 297 16.1 Activities Laading up to Project Execution 297 16.2 The Ceará Integrated Cassava Development Project 299 16.3 Progress Toward Outcomes 299 17. PARAGUAY 313 17.1 Cassava Production 313 17.2 Postharvest Activities 320 18. SEED SUPPL Y SYSTEMS 325 18.1 Problem Definition 325 18.2 Activities 327 18.3 Challenges far the Future 329 19. REGIONAL COLLABORATION IN ASIA 335 19.1 Objective and Priority Setting 335 19.2 Areas of Collaboration 336 19.3 Achievements and Future Emphasis 338 19.4 Lessons Learnt 341 20. GERMPLASMIMPROVEMENT 343 20.1 Generation of Breeding Materíals 343 20.2 Oístribution of Breedíng Materíals to Asian Programs 346 20.3 Varietal Selection 348 20.4 Variatal Release 352 20.5 Socioeconomic Effects 353 20.6 Futura Emphasís 353 20.7 Lassons Laarnt 354 21. SOIL CONSERVATION ANO FERTIUTY MAINTENANCE RESEARCH IN ASIA 355 21.1 Cassava Soils and their Major Constraints in Asia 355 21.2 Erosion Control 357 21.3 Soll Fertílity Maintenanca 365 21.4 Conclusions 374 22. COLLABORATION WITH liTA IN SUPPORT OF AFRICAN NATIONAL PROGRAMS 375 22.1 Importance and Trends of Cassava in Atrica 375 22.2 Araas of Collaboration 376 23. EXPANSION OF THE GERMPLASM BASE FOR AFRICA 379 , 23.1 Strategy and Mathodology 379 23.2 Raactíon to Pests 383 23.3 Agronomic Characteristics 387 24. COLLABORATIVE STUOY OF CASSAVA IN AFRICA 399 24.1 Site Selection for Phase One 399 24.2 Data Analysis 402 24.3 Further Analysís of the Africen Geographíc Informatíon Systems tor Cassava 404 25. ADOPTION ANO IMPACT STUOIES 407 25.1 Integrated Cassava Projects 407 25.2 Varietal Adoption in Asia 425 THE WAY AHEAD 26. THE CASSAVA PROGRAM IN THE 19905 26.1 Cassava's Socioeconomic and Political Environment 26.2 Cas5ava Within the Context of International Agricultural Research 26.3 The Institutional Environment 26.4 Cassava Program Strategies in the 1990s 26.5 Program Organization and Resource Allocation 26.6 Complementary Activities ACRONYMS ABBREVIATIONS APPENDICES CASSAVA PROGRAM PUBLlCATIONS 1987-1991 STAFF L1ST 427 427 431 432 433 437 438 441 445 449 473 EXECUTIVE SUMMARY CASSAVA PROGRAM REPORT 1987-1991 EXECUTIVE SUMMARY This .J3Port presents the work and achievemems of CIAT's Cassava Program over the period 1987-1991. The period corresponds to one in which total world cassava production rose from 137 million t in 1986 to 150 million t in 1990, a growth rate of 2.3% per annum. The Cassava Program at CIAT forms part of a global cassava research and development system. Together with its sister program at the International Institute of Tropical Agriculture (liTA), the Program generates basíc knowledge on the crop and develops improved component technologies, trains national cassava researchers and extensíon leaders and prevides an important link between institutions in !he developing and developed world. At headquarters the Cassava Program undertakes strategic and applied research of global signíficance. The Program ensures !he integration of its actMties with !hose national cassava R&D systems in the Americas, Asia and with liTA in Afríca through regional collabcrative programs and projects. During !he period 1987-1991, tha number of senior staft positions within the Program increased from 9 to 11. These human resources wara complemented by sciantific personnel contracted through a numbar of special projacts. HEADQUARTERS-BASED RESEARCH Cassava Germplasm Collection, characterization and conservation. The receipt of cassava germplasm accessions, from liTA and from Asian national programs, constitutes tha basis far cooperation in Atrica and Asia. Recent germplasr:n introductions from Brazil will giva proper represantation to one of the richest countries in terms of genetic diversity. A complete morphological and biochemical characterization of accessions in tha garmplasm collection is allowing tha detection and elimination of duplicates. A core collection rapresenting the overall genetic diversity has been defined and will be available in 1992 for duplication in another institution. An In vitro and fíeld collection of wifd Manihot relatives has been initiated. Some of !he species are already included in crossability studies, biochemical and molecular characterization. Breeding. The ecosystem approach to cassava breeding has be en expanded to include semiarid areas, where cassava can contribute to the alleviation of food shortages. As a result of a cyelie process of selection and recombination over more than ten years, a large number of elite clones for humid and subhumíd lowland tropics and highland tropics have been identified. Over the last five years 15 varieties derived from CIAT's gene pools have been released in 9 countries in Latín America and Asia. A wide range of geneec díversity has been introduced to Africa from CIAT, increasing the potential impact of IITA's cassava breeding program on that continent. A promising collaborative effort with Brazilian institutions to develop cassava germplasm for semiarid and subtropical regions of the world has been initiated. A participatory research model has been developed, in cooperation with national institutions in Colombia, which takes into consideration farmers' criteria for selecting and adopting new cassava varieties; this has contributed to the refinement 01 breeder's sefection criteria. Prelíminary experiments on seed treatment and germination have been undertaken as part of a multidisciplinary project to explore the development of true cassava seed as a production alternative. Physlology Water Stress. Research on cassava tolerance to prolonged mid-season water stress have reveaJed that eertain physiofogical and morphological characteristics are associated with tolerance to water stress and with yield stability. Most notably, the ability to rapidly form and maintain leaf area is of paramount importance. Second, genotypic characteristics related to better partitioning of biomass between leaf and stem such as high speclfic leaf area and high leaf area ratio may lead to higher leaf area index without adverse effect on root yield. Cassava appears to tolerate prolonged drought by restricting leaf canopy and top growth and by partially closing its stomata while maintaining reasonable CO. uptake rates, hence reducing water loss and continuing to accumulate dry matter into storage roots. Moreover, cassava is capable of extracting deep soil water slowly when available. These characteristies are advantageous in drought-prone areas such as Sub-Saharan Africa, Northeast Thailand and Northeast Brazil. Under dry conditions, cassava can produce reasonable yields while most other tood crops would fail. The ability of cassava to survive and produce under prolonged drought of several months is based on its high water use efficiency as compared to other food crops. Screening for low HCN content under water stress has revealed that genotypic differences exist with some clones maintaining low HCN levels under both wet and stress conditions. Maintaining low HCN under stress is of a paramount importance when fresh cassava is used for human consumption. Photosynthesis. Research on cassava photosynthesis in relation to crop productivity has reveaJed that cassava's high photosynthetic potential underlies its high productivity under favorable conditions as well as its tolerance to stressful environments. Cassava requires high ambient temperature and high solar radiation far optimal leaf development and for the expression of its photosynthetic potential. The high sensitivity of photosynthesis to temperature suggests the need for genotypes more tolerant to low temperatures. which ji could be used in the highland tropics and in the subtropics. Moreover, the positive association of photosynthesis with productivity suggests that seleetion for high photosynthesis ín parental materíals may lead to higher yíeld when combined with other yield determínants. In addítíon to hígh harvest index, leaf area duration and number of storage roots are among the most important yield determinants. Wide ranges of variability in these charaeteristics exist among cassava clones which could be exploited in cassava breeding programs. Quallty Cyanide. The use of the improved enzymatic methad of cyanide analysis has resulted in a cheaper and more flexible assay for this important chemical constituent. No significant correlation was found to exist between the enzymatic and picric acid methads of cyanide determinatíon. A priority has thus besn placed on the davelopment of a new method for rapid cyanide estimatíon for germplasm screening purposes. Analysís of some wild Manihot species showed high levels of cyanide in both root and leat samples. Starcb guality. In addition to starch and sugar content analyses, the methodologies for amylose contents, microscopic evaluation of starch grains, x-ray analysis. differential scanning calorimetry. Brabender viscoamylograms and starch solubirrty and swelling power have all been developed. either at CIAT or in collaborating Colombian institutions. Varietal differences in amylose content (range 16-25%) and tunetional properties have become apparent. M Col 1522, which produces high quality sour starch has different struetural. physical and functional properties from the other varieties so tar avaluated. Eating quality. An expert taste panel for eating quality evaluation identified the charaeteristics of high quality cassava. Cassava taste and hardness (texture) of fresh. bailed cassava were the two most related to preference, while no relationship was found between starch properties and eating quality. The panel identified significant changes in eating quality due to the pre-harvest environment. Effeet of pre-harvest environment on quality. Soil fertility significantly affected root quality. The application of potassium reduced cyanide and increased dry matter contento while phosphorus tended to have the opposite effeet. Orought stress affeeted root dry matter and cyanide contents. with large varietal differences in response. Blotechnology Molecular fingerprinting. Electrophoretic analysis of Esterase (EST) isozymes were developed to identify cassava genotypes. Recently mora powerful. ONA-based, techniques for the analysis of Manihot genetic diversity have been developed. Several random amplified porymorphic ONA markers (RAPO) and hte phage M13 probe haya been effeetive in differentiating cassava genotypes. EST fingerprinting of cassava is now iii routine activity in the GRU. Both EST and DNA fingerprinting will be useful for genotype identification, to assess germplasm relatedness and gene pool origino In vitro Active Gene Bank flVAG) and Cryopreservation. The conservation of cassava germplasm under reduced growth conditions has been developed in the BRU and recently transferred to the GRU. The cassava IVAG at CIAT is probably the largest and most complete for any crop in the world; over 4500 accessions are currently maintained. This year, consistent recovery of plants from cassava shoot tips, cryopreserved in liquid nitrogen (-196°C), has been achieved. This break through opens the way to a long-term, gene bank storage of cassava. Genetic transformatiQn. Plant regeneration is necessary for a transformation system in cassava. Plant regeneration through somatic embryogenesis on immature leaves and meristem tips has been developed. Using a plasmid construct with two selectable markers (bar gene and NPT 11 gene) and one reporter maker (gus A gene), the transient expression of GUS activity in somatic embryos after bombardment of mEÍtallic particles coated with the plasmid on early stage embryogenic cailus has been obtained. This is an important first step towards a transformation system in cassava. Molecular mapping. A special research project to construct saturated molecular and physical maps of cassava has been initiated. Such maps will be useful to tag, and eventually isolate and clone, cassava genes. Pst I and Hind 111 genomic librarles provided probes with the highest polymorphisms. Polymorphism between a cassava cv. and a wild Manihot sp. was dramatically higher than within varieties. The Gassava Biotechnology Network (C8N). The CBN was founded in a workshop at CIAT in September 1988. The general goal of the network is to contribute to the solution of priority constraints in cassava production and utilization which have proved recalcltrant to the application of traditional methodologies. The CBN has received wide acceptance. This is shown by the number of projects underway on research constraints and technological bottlenecks, from 5 in 1988 to 22 in 1991. A proposal for funding critical network activities has been presented to the oolS, the Netherlands. The project ineludes funding for a coordinator, scientific and steering committee meetings, training for developing country scientists and bridging funds far critical research. Pathology Etiological studies. The following unreported pathogens of cassava were characterized: Scytalídium sp. and Verticillium dahliae. Fusarium so/ani and F. oxysporum were also reported far the first time as cassava root pathogens. Epídemialogical studies. The relationships between mycorrhiza-Phytophthora and flooding-Phytophthora incidence were investigated. iv Control of important cassava disaases. Production recommandations haya been formulated for the control of cassava bacterial blight (CB8), the Phytophthora/Fusarium root rot complex, Diplodia root rot, and the witches' broom mycoplasma disease which cause severe epidemics in 4 important cassava growing areas. These recommandations integrate several control approaches to reduce bacterial or fungal infections of planting material and son, and pathogen dissemination. The release of two resistant clones in Brazll with production recommendations has been an outstanding success. Biological control. Research on novel biological control strategies is leading to the development of practical applications based on the use of microorganisms for the control of foliar pathogens (CBB and the superelongation disease) and preharvest or postharvest root rots, as well as microorganisms capable of producing growth regulators that can promote yield production. Advances have been made in development of the cultural practices which promote beneficial microbial residents in different ecological zones. Endophytes. The existence of unreported, potentially deleterius endophytes in improved clones was demonstrated for the first time. This finding partially explains yield instability in most cassava, and opens up a new potential far development of control measures. Simílarly, the identification of beneficiai endophytes could lead to their use as biological control agents, plant growth stimulants, and indueers of drought resistance. Stake storage. A system for the effective storage of cassava stakes was designed far solving problems of establishment and production in areas where this agronomíc practice is neeessary. , Method for interchanging indexad vegetative planting material. A new method for the interchange of virus-indexed vegetative planting material 01 cassava was developed to assure establishment of introduced genotypes. Geographie distribution of cassava diseases. The geographie distribution and areas of potential risks have been determined for Phytophthora root rot, Fusarium stem and root rot; the witches' broom myeoplasma, superelongation disease, Diplodia root rot, and cassava bacterial blight, the most important diseases of cassava in Latin America. These were obtained by extrapolating epidemiology studies, surveys and agroeeologieal data bases. This information will assist scientists in planning researeh projects, strategies for disease control and quarantine regulations. Virology Frogskin (FSD) and Caribbean mosaie disease (eMD). Progress has been made on the identification of phytoreovirus-like agents assoeiated with FSD and CMD. Virus-fike partides and viroplasm-fike bodies have been found in affected plants. Nine ds-RNA segments are eonsistently found in affected plants. Hybridization studies provida evidenee that the ds-RNAs associatad with FSD and CMD are either identical or elosely v related. The whitefly Bemísia tuberculata appears to be the vector of these phytoreovirus- like agents. A cONA probe has been developed to identify rapidly tha ds-RNAs associated with this disease complex. Research on FSO and CMO continues and is centerad on confirming the association of the phytoreovirus-like agents and the complex of disease symptoms. Cassava vejn masaje virus (CCMV). The sequencing of CCMV is nearly complete. The virus Is 6400 bases in length and it Is most closely related to potato virus X (PVX). Most of the sequencing of CCMV was done at the VRU in CIAT and this is perhaps the first plant virus to be sequenced in Latin America. Cassava vein mosaic virus is most prevalent in the northeastern states of Brazil, especially in the hot semiarid zones where it Is not unusual to find more than 50% of the plants infected with CVMV. Since the virus is not presant in Colombia, all work on this virus had to be done in Brazil. A cONA clone to CVMV has been obtained, and this will facilitate efforts at molecular characterizatíon and the development of rapid diagnostic tests. Afriean Cassava mosaic virus (ACMV). Afriean cassava mosaic virus is the mast destructive cassava virus in the world. Efforts are being made to find new sources of resistance through the CIAT -liTA cooperative germplasm exchange project. The identification of ACMV resistant germplasm adapted to the Americas will be a safeguard against the possible establishment of ACMV in this hemisphere. Other viruses. Beside the majar diseasas, thera are seven other known viruses that ¡nfect cassava most of which ara symptomless viruses that are not known to eausa disease. Oiagnostic methods are available at CIAT to the four viruses that cause disease and to the thrae symptomless viruses found in Latin Ameriea. These diagnostic methods help assure the safa movement of eassava germplasm, and research will continue to develop more sensitive detection methods. The control of viral diseases requires either the identifieation of resistant germplasm or the implementation of cultural practices that mitigate losses. Most viral diseases are controllable with current technology, and continued development of rapid diagnostic techniques together with the deployment of resistant germplasm should further reduce the losses caused by viruses. Entomology and Acarology Host plant resistance. Sources of host plant resistance to mitas, mealybugs, whitaflies, thrips and lacebugs heve been identified, and partially characterized for miles and thrips. Cassava hybrids with rasistance to mites, mealybugs, whiteflies and thrips have been developed and several have been released by national programs. vi Cassava mea!ybugs. The geographical distribution of the most important species 01 cassava mealybugs has been determined. Their key natural enemies have been identified. Several species of parasites and predators have been evaluated in the laboratory and field for their potential as biological control agents, leading to the introduction and release in Colombia of the parasite, Aenasius near vexans, discovered in Venezuela. Several species of natural enemies have been sent to liTA in Africa for evaluation and release against the introduced pest, Phenacoccus manihotí. Cassava hornworm. Effective biological control of the cassava hornworm, a migratory lepidopteran which ceuses severe defoliation, is based on a hornworm-specific baculovirus. The timing, application frequency and optimal concentration of virus prepared from field-collected, diseased hornworms have been determined, and methods tor storage of the virus have been developed. . Application of the virus duríng the ¡nmal stages of a hornworm attack, when hornworm larvas are most susceptible results in better than 95% control. Cassava whiteflies. Because of recent increases in dlrect crop damage due to cassava white flies and in their potential importance and impact as vectors of virus diseases, higher príoríty has been given to research on whiteflies. High levels of resistance have been identified and incorporated into high yielding hybrids. Several species of natural enemies have been identified and are being studied. Burrowing OOgs. Chemicel control of burrowing bugs which attack cassava roots is feasible, but requlres the use of highly toxlc pesticldes. Ecologically sound alternatives based on cultural practices are being sought. Intercropping of cassava with crotalaria reduces pest damage through allelopathy, but has not been adoptad by farmars. COmmercially acceptabla cultural control systems based on a1lelopathy are being sought. The potential of entomophagous nematodes tor biological control 01 the burrowing bugs is under study, and preliminary results have been positive. True cassava seed and dried cassava. Research efforts on potential arthropod pesta of true cassava seed, wild Manihoti species and of dried, stored cassava have been initiated. Cassava green mite. CIATs contríbution to biological control of the Cassava Gresn Mite (CGM) in the Americas and Africa includes 1) extensive surveys for natural enemies in Colombia, Venezuela and Ecuador and smaller-scale surveys in Northeast Brazil, Trinidad & Tobago, Guyana, Peru, Paraguay, Mexlco, Cuba, Panama and Nicaragua; 2) development of culture, packing and shipplng methods far natural enemies; 3) ecological and biological characterization of predatory mitas, coleopteran predators, and the fungal pathogen, Neozygites sp.; and 4) estimation of field impact of natural anemies. As a consequence of the research conducted on bahalf of IrrA, CIAT characterized the biological and ecological nature of the CGM problem identified by EMBRAPA in Northeast Brazll, and developed a strategy for CGM management as parí of an integrated crOP protection effort. vii Cropping Systems Cassava-maiza intercropping. The evaluation of newly releasad maize varieties for performance in association with cassava indicates that anvironmental conditions during maize davalopment and the agronomic management of the maize are key factors determining the yield of cassava in the North Coast of Colombia. Land use is more efficient when maize and cassava are planted together than when either is planted alone. Cassava yields significantly more in association with improved than traditional maize varieties, thereby increasing land use efficiency further. In less favorable environments, yields of cassava in association with maize were significantly lower than in monoculture. Improved maize varieties allocate mora dry malter to the grain than traditional varieties. Total nutriant removal in improved maize is greater than in traditionai varieties. As new varieties replace traditional maize, nutrient balance within tha farm will be dífferent, since more nutrients will be exported from the system. Unless soil fertility is maintained by lengthening fallow periods or by other means, yields will decline. Maíza type did not influenca uptake of nutrients by cassava. Since the restoration of nutrients to the farm is vital for tha sustaínability of the cassava¡maize association, chemical fartilization Is a possible short term solution in some regions. In trials with low levals of fertilization improved maize yielded significantly more than traditional maize. Cassava yialds were not affected by the level of fertilization applied to maize. the highest marginal retum in intercropped cassava/maize was obtained with low levels of fertilization. Cassava-cowpaa jntercropping. Tha cassava/cowpea association is important in areas with prolonged dry seasons. Yields of intercropped cowpea are often similar to yields obtained in monoculture, however cassava yields ara negatively affected by cowpea competition, particularly if environmentaí conditions favor the early development of cowpea. More vigorous cassava clones are less affected by cowpea competition. Pre-production trials. Farming involves a yearly sequence of events beginning before the onset of planting and ending with post harvest activities. Testing of new production practices should be done in this context, particularly since the adoption of any technology component depends strongly on the interaction between new and existing technology and labor requirements, particularly during peak demand perlods. Improved technology ror the cassava/maize assocíation was testad by farmers over two years on a totaí of 76 plots. Malze yields with improved technology were superior and cassava yields were equal or s¡.¡perior to those obtained with traditional technology. Hand labor requirements of the improvad and traditionaí technology were similar. Totaí production costs of the improved technology were 8% aboye those of the traditionai technology. viii Plant nutrition and soil fertility Management af low fertilitV soils. Research an plant nutrition and soil fertility management in relatian to cassava productivity have revealed that cassava is tolerant to low fertility son provided that soil organic matter is high. Continuous cultivation of cassava for several years in acid soils high in organic matter did not resutt in large declines in yield in absence of phosphorus and nitrogen fertilizer. On the other hand, large yield responses to potassium fertilizer were notable in these soils. The removal 01 large amounts of potassium in the harvestable roots leads to gradual depletion of thís element. To sustain productivity, moderate amount of potassium fertilizer should be used. Where cassava is produced in sandy soils low in organic malter and in the absence of a fallow system, moderate levels of NPK fertilizer are required to sustain productivity. Yield responsa to NPK fertilizer was notable in these soíls. Alternatively, cassava productivity could be increased by application of surface plant mulch in poor sandy soils. Mulch application appears to be beneficial in improving the chemical and physical properties of the soil. Moreover, mulch can alleviate water stress by reducing water evaporation trom the surface soil exposed to high temperature. Reducing evaporation is important, sinca these sandy soils are characterized by a low water retention capacity. Another advantage of mulching is the large reduction in HCN content of cassava roots in tha absence of fertilizar application. Adaptation to Iow-phosphorus soil. Screening cassava germplasm for adaptation to low- phosphorus soils indicated a wide range of adaptation among the tested materials. Several varieties well-adapted to low-P soils were identified, including land races as wall as advancad breeding lines. Mest notably, the two CIAT clones CM 523-7 and CG 2177-2 recently released as commercial varieties for los Uanos Orientales of Colombia were among the highly adaptad lines to low-P soils. Research on mechanisms underlying varíetal response to P suggested that varietal differences were not closely related to P- uptake. On the other hand, internal use of absorbed phosphorus, growth habit and pattems of biamass allocation to tops and roots are more important. Varieties that partitioned more dry matter to roots, as compared to top growth, had higher Puse efficíency in terms 01 yield gains. It appears that adaptation to low-P soils coúld be enhanced by selection for both high fibrous root length density and high storage roots sink capacity. Sol! conservation Control of soil erosiono Research on soil conservation in cassava-based cropping systems on hillsides indicated a high level of soíl erosiono Annual soíl losses trom bare soils exceeded 100 t of dry soil/ha. Since cassava planting usually coincides with periods of intense ralnfall, soilloss trom steep lands might excaed the tolerable levels unless the soils were appropriately managed. Several cropping systems and cultural practices were tested in relation to soil erosion and crop productivity. Growing cassava in contour ridges or in combination with live grass barriers greatly reducad soil erosion while maintaining ix cassava productivíty as compared to traditional practica. On the other hand, growing cassava in down-slope ridges resulted in high levels 01 soil losses and more runoff than any other practices. Growing cassava in association with 10rage legumes was effective in reducing soil loss but cassava productivity varied with the degree of legume competition. Reduction in cassava yield ranged from 1 a to 40% depending on the legume used and on the intensity of the legume cover. The potential of cassavallorage legume systems in controlling soil erosion and in maintaining productivity requires further investigation taking into account both the short- and long-term consequencas. Process and product development Methodology. A 4-stage methodology for cassava process and product development, comprising identification of opportunities, lab and prototype research, pilot scale testing and commercial expansion has been developed in collaboration with national institutions. Orled cassava for animal feed. The commercial expansion of dry cassava chip production, introduced by CIAT in collaboration with the Integrated Rural Oevelopment Fund on Colombia's Atlantic Coast, is now self reliant and autonomous. Feedback from this project has resulted in research on improvements in dryíng efficíency and product quality. The production of cassava-based chicken feed rations was found to be a viable option at the level of small farmer cooperatives. Eresh cassava conservation. Pilot scale testing in the city of Barranquilla, Colombia has demonstrated the technical and economic feasibility of the fresh cassava storage technology developed by CIAT and N RI. Problems with urban distribution have frustrated large scale adoption of the storage technology. However, private entrepreneurs are now actively taking on distribution functions, and supplying supermarkets, restaurants and small shops. The storage technology is being successfully used commercially by a cooperative in Santander department Colombia and pilot testing has been successful in Paraguay. Cassava flour. The pilot stage of a project to develop high quality flour for human consumption is currently being executed. Market studies, including industrial tríals of the flour, have demonstrated that cassava flour will have both price and quality advantages over wheat flour in some market segmants (a.g. cookies and processed meats). A potential market of over 20,000 t/yr in Colombia was estimated. The pilot processing plant currently under evaluation is operatad by a small farmer cooperativa and employs artificial drying of chips produced from washed roots. The high quality chips are milled at a wheat flour mili with conversion rates of chips to flour of 90%. Current information suggests that the project is economically feas;ble in Colombia, and that the rate of return is improved if in-plant milling is adoptad. A small scale prototype mili has been designed which will permit in-plant production of flour. x Cassava starch. A CIRADjCEEMAT-CIAT research program on cassava starch started in 1989 focusing primarily on sour or fermented starch. Evaluation of existing traditional small scale production units in Colombia identified areas for process improvement to increase effioiency and improve product quality. Two pilot plants incorporating process improvements are now under avaluation. The characteristic "expansion power" of sour starch as measured by specific volume correlated significantly with organlc acid contents and certaln viscoamylogram oharacteristics. The natural fermentation process has been found to be predominantly lactio, with ce. and lactio acid production and amylolytic enzyme action pitting starch grains. Maximum viscosity of sour starch is lower than that of raw starch and gelling abílity is reduced. REGIONAL COLLABORA TION Lat'n Amerlca' Many countries in the Americas have become aware of the important role that cassava can play in providing a vehiele tor income and employment generation in the rural sector, consequently the last five years has seen a resurgence of interest in cassava activities. In addition to direct collaboration with national research programs, participation in multi- Institutional integrated cassava research and development projects in severa! countries has formed the basis of the Cassava Program's actMties in the region, wíth priority plaeed on Colombia, Brazil, Ecuador and Paraguay. Tbese projects aim to link cassava farmers with expanding markets through the introduction of novel or improved cassava processing alternativas, thus providing incentives for farmers to increase production.' Emphasis in training during the period has been on the support of in-country courses, with fewer CIAT·based production and utilization eourses. Training at CIAT has inereasingly focused on diseiplinary in-serviee spacialization. Networks, both speeialized and sub-regional in nature, have been eonsolidated or established; these inelude the Panamerican Cassava Breeders' Network, an Integrated Cassava Projacts Network, a Cassava Utilization Research Network and a network tor Cassava Devalopment in the Southern Cone countríes. These networks plays a key role in definition of regional researeh priorities and identification of opportunities for horizontal collaboration. Colombia As CIAT's host eountry, Colombia plays an important role in providing situations for the testing and adaptation of component teehnologies and participatory researeh and 1 For the purpose of this report the terro Latin Amerles ¡neludes Mexlco. Central Amerlca. the Caribbean and South Amerlca. xi technology transfer melhodologies. As such it provided Ihe site for the first integrated cassava project initiated in 1981 on the Atlantic Coast. Rapid and dynamic growth of both farmer cooperative and private cassava drying plants has occurred over the last two years. There are now over one hundred plants located on the Atlantic Coast and expansion of the cassava drying technology to other areas of the eountry has been aehieved through a join! National Rehabilitation Plan/CIAT project. The national produetion of dry eassava is now estimated at 25000 t with benefits accfuing to over 5000 families. The increased market tor eassava has stimulated demand from farmers for improved eassava production technologies. A study on the adoption of cassava production technology components in Ihe Atlantic Coast of Colombia was undertaken in 1991. Preliminary resuits of a sub-sample of the data shows that: • Cassava varieties "Venezolana" (M Col 2215) and "Verdecita" (M Col 1505) haya been adopted by 91% and 5% of eassava farmers, raspectively. Together they cover 44,000 ha in the three principal cassava producing departments of Colombia. • Stake Irealment and storage technologies have been adopted by 10% and 71% respectively. Planting density and weed control technologies have been adopted by 60% and 53%, respectively. • Technology adoption has been the principal factor far cassava yield increases. Sinee 1982 cassava yields (eassava/maize intercrop) increased by 52%, 56"'-&, and 76"'-& in the departments af Bolivar, Sucre and Cardaba,. respectively. • Cassava area has increased significantly as a reaetion to improved prices and demando Cassava farmers have increased area planted to cassava decreased fallow area and period, and 95% of farmers are harvesting the same area as planted rather than leaving cassava in Ihe ground until markets improve. • As a reactian to improved cassava prices and overall demand, cassava farmers have deereased on farm eassava consumption (as share of total production) by some 50% since 1982. Sales to drying plants currently constitute 22% of total eassava produetion. • Overall, 71 % of cassava farmers have adopted at least one production technology component, and 80% responded that they haya increased their incomes as a result of improved technology and the increased demand and improved market. The data clearly show that cassava drying plants have served as an effective vehicle for eassava teehnology diffusion. xii Ecuador Th9 Ecuador integrated cassava project operates primarily in two coastal provinces, Manabl and Esmeraldas. The goal of the project is to unite and integrate the efforts of local, national and international agricultural development institutions engaged in research, extension and education in order to identífy cassava production, processing and utilization technologies appropriate for low-resource cassava farmers. Current project beneficiaries are 18 farmer associations (APPYs) with 350 members in Manabl and 5 with 60 members in Esmeraldas. Among the APPYs, four have all women members, eight have only men and eleven have mixed membership. 111e women's APPYs produce cassava starch exclusively while the men's and mixed associations produce cassava chips which are milled ¡nto various flour products. The APPYs in each province are organized into unions (UAPPYs) which are responsible for providing the associations with credit, training and technical assistance, and handle the marketing of proeessed cassava praducts. Total output from the UAPPY-Manabr increased markedly from 50 t of cassava flour during the initial year of the project (85-86) to 1,346 t of flour and 104 t of starch during the 90-91 processing year. Research conducted within the project until 1989 was primarily focused on adapting cassava processing technology from the Atlantic Coast of Colombia to the agroecological and social conditions of Manabr and Esmeraldas. In 1989, an unexpected downturn in the demand for cassava flour as the agglutinant for making shrimp feed pellets causad a dramatic shift in the cassava' programo Farmer processors demanded assistance in identífying new markets for their existing products, technology for producing new products and methods to improve processing quality. An intensive market diversification effort was therefore initiated. Today the primary markets for the UAPPYs products inelude cardboard box factories, plywood milis and food industries as well as the shrimp feed industry. Farmers, now aware of the need for better fresh cassava quality, are demanding new varieties with higher dry matter, improved drought tolerance and earliness. The first new variety meeting these requirements, M Col 2215, introduced through CIAT in 1987, will be released by INIAP this year. Countries like Ecuador, where farmers are demanding improved production and processing technology and there Is no single institution with post-hervest research capability, require new mechanisms for conducting research. The formatlon of multi- institution and interdisciplinary teams with the active participation of trained UAPPY para- technicians is proving to be an efficient alternative approach which ensures the continuous involvement of farmer users in the research process. xiii Brazil Following intensive contacts with Brazilian research and extension agencies through training events and study tours in Colombia aod Ecuador, an integrated cassava development project was initiated in the State 01 Ceará in 1989. The project which is executed by the Ceará State Cassava Committee (CCC) is partially financed by the W.K. Kellogg Foundation. One of the principal activities of the project has been the organization of tarmers' groups tor the construction, operation and administration of cassava proeessing facilities. When the project ínitiated activities in May 1989 12 drying plants already existed. By the end of August 1991, the total number of small-scale processíng plants had risen to 59 and 1380 farmers were benefitting directiy from the project. The building up of local institutional capacity and support tor the project has progressed steadily and the role of the CCC as the coordinating body tor all activities related to eassava development has gained general recognition. In addition, five Regional Gassava Committees (RCC) have been established and are contributing to the rapid and efficient decentralization ot project activities. The identification of local financial rescurces tor expansion ot the project into new areas has been actively sought. The total value of resources obtained and allocated to farmers' groups to finance the construction of their cassava processing facilities now amounts to US$347,048. Marketing channels tor dry cassava chips have started to consclidate. The majn consumers of the dry cassava chips have been dairy farmers located in the vicinity of the drying plants. ·In 1990, the total number of purchasers was 410, with 19 (5%) of these buying 62% of the total production. The first results from the 15 pre-production plots planted in 1990 have shown that, at 15 months, average yields of cassava were 60% higher than those obtained by tarmers employing traditional production practices. Paraguay Paraguay is the largest per capita producer of cassava roots in the world and the erop is considered strategie in terms of the country's toad security. CIArs Cassava Program has been instrumental in orienting and supporting a young, unexperieneed but dedicated group of researchers and extension leaders in defining priorities, tormulating objectives and strategies and executing projects with the objective 01 maintaining cassava's position as a principal souree of carbohydrates for both human consumption and animal feed. The focus of cassava related activities in Paraguay has been centered on (a) sustaining and improving production, with particular emphasis on soil fertility and erosion control and xiv (b) making better use of the crop through improvements in post harvest handling, processing and marketing. Since 1985, a project partially financed by the International Oevelopment Research Centre, IORC, has concentrated on two important cassava growing areas: Paraguar!, a Oepartment close to the capital Asunción, where soil degradation has signíficantly reduced the quantity and quality of the cassava produced with the resuh that the area is a net importer of roots; and Caaguazú, a Oepartment where the native forest has been apenad for agricultura within the last twenty years and is the principal supplier of cassava for the Asunción market. The development of technological components to improve cassava production in Paraguar! and Caaguazú have led 10 the formulation of two complete technical recommendations for farmers. The fresh cassava atorage technology developed by CIAT INRI is being tested and adapted to conditions in Paraguay. Starch is used mainly to make a traditional bread known as "chipa"; work is underway té improve extraction efficiency and product quality and to introduce simple effluent treatment. Seed systems The opportunities for increased cassava root utílization described above and the demand from farmers for improved production components that will increase productivity and reduce costs, has highlighted the need to undertake activities oriented toward 1he development of organized cassava seed supply systems that will ensure the availability of high quality planting material of either local origin or from improved genotypes. This area is seEm as a major constraint in the evolution of the integrated cassava projects towards íncreased and more stable cassava production at lower costs. In collaboratíon with the Colombian Agricuhural Institute, ICA, the Seed Unit has been developíng pilot models for the organizatíon of seed production in different regions of 1he country whích differ according to the user groups and end uses of cassava. The experience gaíned will provide the basis for implanting similar modela in other countries. In Asia, cassava faces fewer market constraints as compared with tropical America, the crop having made the transition from being purely a starchy staple to a multipurpose earbohydrate souree in many countries. In addition, cassava research programs are relatively stronger with a low tumover in personnel. Priorities for research at the regional level have been focused on germplasm improvement and son fertility maintenance as key elements in ensuring highly productive eassava-based eropping systems. The promotion of horizontal exchange of information on post harvest processing and marketing has also reeeived attention. Training of personnel, the execution of joint projects through research contracts, and the formatlon of a regional cassava research network that meets every xv three years have been the principal mechanisms tor improving national programs' research capacity. As a result of CIAT's collaboration, the research capacity of cassava programs in Thailand, Indonesia, China, the Philippines, and Malaysia has been greatly strengthended and new cassava research programs established in Vietnam and Myanmar. Through CIAT involvement, interinstitutional cooperation at a national level has improved in Vietnam, China and Indonesia. Thailand is contributing its best breeding materials to other Asian countries through CIAT. Cassava variatal Improvement in Asia Generation of breeding materíals. Cassava breeding programs in Asia have benefitted significantly from 1he availability of selected and upgraded genetic materials from CIAT /Colombia, which are characterized by improved harvest índex (HI) and tolerance to biotíc and abiotic constraints. Since 1975, 274,196 hybrid seeds from CIAT /Colombia have been distributed to 9 countries. The establishment of a joínt Thai-CIAT cassava breeding program has brought about further yield improvement through higher bio-mass and root dry matter content as well as adaptation to semiarid lowland tropícs and ímprovement plant type. From thís program 47,224 hybrid seeds have been dístributed to 9 countries since 1965 and 215 clones have been shipped to 11 countries since 1988. Varíetal selection. Steady progress in varietal selection has been made by the cassava breedíng programs in Thailand, China, Vietnam, Malaysia and the Philippines with promísing materials being selected from CIAT /Colombia and Thai-CIAT introductions and local-CIAT crosses. In Thailand 7 clones with different adaptive niches have been or are in the process of beíng released over the period 1984-93. Varietal release. A total of 12 varietíes in five countries have been released. The number is expected to increase steadily in the future with selections from Thai-CIAT crosses and Thai-CIAT clonal introductions gaining in importance. Adoption. Rayong 3 in Thailand and Adira 4 in Indonesia are planted on more than 50,000 ha. The extent and factors affecting adoption of both ttlese varieties are currently being studied in close collaboration with the respective national research and extension programs. VC 2 and M Col 1684 (not officially released yet) in the Philippines and Nanzi 188 in China are planted on smaller hectareages. SoU conservation and fertlllty maintenance research in Asia Priaritv setting and improving research caoacity. Afler the establishment of the Cassava Agronomy Program in Asia in late 1986, a network of cassava agronomists and soil scientists warking in national programs in Asia was developed by conducting collaborative research on high priority topies. Through frequent visits to the national programs to see xvi the trials and to discuss the results, the organization of workshops and training courses and distribution of cassava literatura, the capacity and efficiency of the research was improved. Since1986 two Regional Cassava Workshops and a Symposíum of the International Society for Tropical Root Crops were organized in Asia, while a production training course for Asian cassava workers was held at CIAT headquarters in Colombia. Agronomy research results. Collaborative cassava agronomy research on cultural practices, on erosion control and soil fertility managements has been conducted in níne countries. The most promisíng economically viable erosion control practices were mínimum tillage, fertilizer application, contour ridging, close plant spacing, and intercropping. Ctissava responded positively to the application of N in short-term tríals, but required relatively high applications of K for sustained high productivity. The crop responded to P application sporadically, and soil aCidity, or lack of secondary or minor nutrients were seldom significant limítations in Asia. Practicas such as green manuring, intercropping, cover cropping and a1lay cropping generally reduced cassava yields, but they may be beneficial for improving the long-term productivity of the soil. Considerable additional research will be needed to integrata arosion control practices, judicious fertilizer use, crop rotations, green manuring etc. in order to manage the crop and the soil for sustained high yields while protecting the natural resource base. Africa Broadenlng the cassava germplasm base A collaborative project between CIAT and liTA is dedicated to the broadening of the cassava germplasm base of Atrica through introduction of germplasm from the Americas adapted to specific agroeconomical conditions. H commenced in 1990 with the introduction of nearly 90,000 botanical seeds representing 400 families. A new seed lot was introducad in 1991. A total 01 130,000 seeds (750 families) have now been transferred to Atrica being evaluated under humid, sub-humid, semi-arid and mid-aHitude conditions in Nigeria. Results obtained trom the material introduced in 1990 at Ibadan (sub-humid), Onne (humid) and Kano (semi-arid) show that progenies derived from crosses between Latín American germplasm and IITA's mosaic-resistant clones TMS 30001 and TMS 30572 are more resistant to the disease under conditions of high pressure observad at Ibadan. Progenies of crosses involving CIAT germplasm adapted to the acid soil savannas of South America and liTA sources also showed a better reactíon to an intense outbreak of C8B in the first 2 months after transplanting. Resistance to cassava green mite was observed in progenies obtained from crosses involving CIAT elite materials adapted to the dry areas of Latin America and resistant to that pest. . xvii The population evaluated under the semi-arid conditions of Northern Nigeria showed a remarkable performance in terms of root yields and growth, after enduring a 6 months dry season which started two months after transplanting. A recuperation period of 3 months after the dry season enabled the seedlings to recover and produce yields comparable to those obtained at the humid and sub-humid environments. Selection in the semi-arid location reflects the degree of adaptation of the genotypes to the harsh climatic conditions since no biotic constraints were observed in 1990-1991. Individual selected as promissory at Ibadan, Onne and Kano were cloned and are being evaluated in four location in Nigeria as part of the cassava breedin§! scheme of the Tuber and Root Crops Program of liTA. The results obtained so far from this large scale germplasm introduction support the feasibility of such a program and suggests that a preselection of parents based on their agroecological adaptation is a step forward in a germplasm-exchange programo Collaborative Study 01 Cassava in A1rica (COSCA) The Collaborative Study of Cassava in Africa (COSCA) is a joint project, managed by liTA, which aims to provide basic information about cassava in Atrica, to increase the relevance and impact of research related to the crop, and to help increase income and food security ter people in Africa. Implicit in the objectives of COSCA was a geographic characterization of how cassava production and utilization varied across the countries involved in the study. CIAT's Agroecological Studies Unit participated in the design of a spatial sampling frame fer the project. To do this it mapped cassava distribution in Africa and constructed a geographic database of climatic, demographic and infrastructural information. A member of the Unit assisted in training and data analysis for the project's first phase. This consisted of a village level questionnaire to elicit qualitative information. National teams were trained in sampling and mapping techniques. Later the Unit participated in analysis of some of the data, inciuding a description of the distribution of bitter and sweet varieties and relationships between varietal characteristics and environmental conditions. The Unit has used the information that it compiled fer the geographic sampling trame to construct a statistical model of the distribution of cassava in Africa. This uses population density, modified by climatic and edaphic factors, to predict cassava distribution in the year 2000, and to identlfy areas where cassava production is notably higher er lower than might be expected. This information has formed the basis for the preparation of an Atlas of Cassava in Africa which will be available in 1992. xviii CASSAVA PROGRAM STRATEGIC PLAN 1992·2002 In the 1990s the Program will continue to promote the íntegration and consolidation of national cassava research and development systems in tropical America and Asia, and 10 facilitate linkages between these systems and institutas undertaking advanced research on cassava through the Bíotechnology Network. Closer collaboration will be sought with liTA to help meet the needs of African programs. While maintaining a commodity system perspective, the program will emphasize germplasm resource development Crop management, utilization and market research will concentrate on strategic issues of global importance. Applied research in these areas will gradually be devolved to national organizations, with horizontal cooperation encouraged between countries at the regional level. The Pragram will focus primarily on technology development for the subhumid, semiarid and subtropical ecosystems of the Americas and Asia, interacting closely with CIATs new Resources Management Research Division on hillside, sevanna and forest margin ecosystems where en estimated 25%-30% of cassava is produced in tropical America. Overall core resources are projected to decline slightly over the period in terms of actual staff positions, and signíficantly in terms of positions approved by TAC for 1989-1993. ACKNOWLEDGEMENTS This report was written by the principal scientists of the Cassava Program and those of the Germplasm Resources, Virology Research, Biotechno!ogy Research, Seed and Agroecological Studies Units with which the Program closely interacts. We wish to acknowledge the very significant contribution made by our support staff. The achievements reported in the document would not have been possible without their hard work and dedication to the objectives of the Programo We are indebted to the Training and Communications Support Program for their assistance in the organization and execution of numerous activities which facilitated the exchange of information between the Program and our partners in nationa! institutions. We also acknowledge the invaluable support of Data Servíces Unit in data management and experimental design and analysis. We wish to express our appreciation for the generous collaboration of many people in both developing and developed country institutions with whom the Program interacts. Apologies are given should their contributions not be fully recognized in the text of the reporto xix Finally, we thank the secretaries who typed the individual sections. In particular, acknowledgement is due to Trudy Brekelbaum, Maruja Rubiano and Miguel Angel Chaux who were responsible tor assembling the component parts into what we hope is a coherent document. xx INTRODUCTION 1. THE CIAT CASSAVA PROGRAM 1 1. THE CIAT CASSAVA PROGRAM In the tropícs, cassava (Manihot escuJenta) is the most important root crop, ranking fourth after rice, sugarcane and maize as a source of cal ories for human consumption. It is a major carbohydrate food for 500 million people, and in tropical Atrica it is the single most important so urce of calories in the diet. Total world production has increased trom 70 million t in 1960 to an estimated 150 million t in 1990. Of this total, 43% is produced in Africa, 35% in Asia and 22% in Latin America. The crop ís principally used as a human toad, either fresh or in a processed form; however, it is of growing ímportanceboth as an animal feed and as a raw material for producing starch and starch-based derivatives. Cassava, which origínated in Latin America, has a number of attributes that have contributed to its reaching this level of worldwide importance and that have made it an attractive crop for small farmers with limited resources in marginal agricultural areas: ~ It is one of the most efficient carbohydrate-producing crops. .. It is tolerant of low soil fertility and drought and has the ability to recover trom the damage caused by, most pests and diseases. .. The roots can be left in the ground for long periods as a food reserve, providing an excellent insurance against famine. .. The crop is welJ adapted to traditional multispecies agricultural systems and subsistence cultivation in which farmers seek to minimize the risk of total crop failure. Cassava is one of the few major toad crops that is not grown to any significant extent in the developed world; thus until recently there had been a historie underinvestment in researeh both nationally and internationally. Over the last twenty years, however, there has been mounting recognition of the eontribution that cassava can make to íncreasing incomes and generating employment opportunities in the rural sector; and this has led to a greater flow of funds for research. The research programs at CIAT and the Internationallnstitute of Tropical Agriculture (liTA), established in the early 19705, have provided stability and continuity for the farmation of what can now be considered as a global eassava research and development (R&D) system, which links institutions ín both the developing and developed world. 1 1.1 History of CIAT's Cassava Program 1.1.2 The first ten years, 1973-1982 The Cassava Program at CIAT was formed at the time of the Green Revolution 01 rice and wheat. The wave of enthusiasm that 10llowed the great success with these two commodities not only led to the establishment 01 the CGIAR (Consuitative Group on International Agricultural Research )system but also defined its overall strategy: to apply modern science to the production of new varieties 01 broad adaptability that would greatly increase yields. The Green Revolution in wheat and rice was based on a large backlog of research, principally from Asia, Europe and the Americas. In the case 01 cassava, however, the scientific base for crop improvement was rudimentary. Nevertheless, the strategy of CIAT's Cassava Program evolved in the spirit 01 the times, revolving around increased understanding of the crop and using this knowledge to provide a genetic solution to the problems encountered in the field. By the end of the seventies, a great deal had been achieved in improving understanding of the plant; however, the goals of providing broadly adapted varieties to 1armers and increasing production were not achieved. The underlying reasons 10r this were: • Cassava was grown under such heterogeneous conditions (in contrast to irrigated rice and wheat) that an ecosystem approach to varietal improvement and concomitant agronomic practices was needed. • The principal constraint to increased production was often not in the area 01 production technology per se, but rather in the area 01 incentives 10r the 1armers to produce more. Strategies were altered accordingly, with emphasis on determining potential markets 10r cassava and on postharvest handling 01 the crop to ensure that marketable goods could be produced. Furthermere, a more holistic viewpoint was taken 01 the role 01 the crop in the overall context of rural development. The question became not ''what can CIAT's Program do for cassava?", but "what can cassava do for the poorer segments of the population in the developing world?" and ''what can CIAT as an Institution do to ensure that it effectively plays that role?" In the early eighties therefore, research on cassava utilization was initiated as a core- 1unded activity, the Program became involved in the first pilot integrated production, processing and marketing project, and a senior breeder was outposted to Asia to provide a link with national breeding programs and broaden the region's germplasm base. Concurrently with this period of internal reflection and realignment 01 strategies within the Cassava Program itself, questions were raised both by CIAT Board and Management and by the Technical Advisory Committee (TAC) 01 the CGIAR about the 1uture demand 1er cassava. The basis 01 this skepticism related to the observed decline in per cap ita 2 consumption of the crop in Latin America and the possibility that the experience on this continent might be a portent of the future for Asia and Africa. It wasargued that there would be little demand for improved production technology for a crop with a declining market. 1.1.3 The cassava demand studies During the 60s and 70s, economies of Latin American countries experienced a significant structural change, mainly caused by strong expansion of the industrial and service sectors, and rapid urban growth as a result of high rural-urban migration rates and an annual 2.8% population growth. Rising income and urbanization implied a rapid increase in demand for food products, with changing consumption patterns, especially in urban areas. Daspite the fact that agricultural production grew at 3.1% per year during this period, the ever-growing demand could not be satisfied. The resulting undersupply (or overdemand) led to an upward pressure on food prices and a negative trade balance. The majority of Latin American countries became even more dependant upon cereal imports. This issue stimulated government policies to focus on (and favor) primary products, among them cereals. As a consequence traditional energy providers like cassava (and other roots and tubers) had to compete increasingly with grains, at a substantial disadvantage and at a high social cost. This-together with the shifting consumption patterns 01 a predominantly urban society in which traditional staples such as cassava were being raplaced by more convenient cereal-based foods such as rice and wheat- based products--Ied to an overall decline in cassava production of 1.3% yearly during the 1970s. The assessment of this situation was translated by the 1984 External Program Review of the Cassava Program into a recommendation that 'studies be undertaken to assess the future demand of cassava and cassava products; that the future direction and scope of the Program should be reviewed after completion of these studies·. While supporting current and future cassava research activities as important, TAC endorsed the need for cassava demand studies as "the future demands for cassava were unclear," especially in Latin America. Between 1984 and 1986 the Economícs Seetíon of the Program, aided by several postdoctoral fellows and collaborating natianal social scíentists in several Asian countries, embarked upon a series of studies (known as "The Cassava Demand Studies") to assess the current importance and future potential of cassava in the most relevant areas of Asia (India, the Peoples' Republic of China, Indonesia, Malaysia, Philippines and Thailand) and Latín America (Colombia, Brazil, Paraguay, Peru, Venezuela, Dominican Republic, Mexico and Panama). At the same time a propasa! was developed for similar studies in Africa, 3 which became known as "The Collaborative Study of Cassava in Africa", COSCA. This project got under way in 1989, and the final results are expected by 1992. The preliminary results were presented to TAC in mid-1987. Overall the studies demonstrated a strong market potential for increased cassava production, depending on usage and region. There were six major conclusions: • The observed decline in fresh consumption in Latin America is due primarily to the urbanization process. High marketing costs have shifted relative prices of cassava and grain staples between rural and urban are as, resulting in lower consumption in urban Latin America. Fresh cassava has a positive income elasticity and consumption can be expected to grow modestly. New preservation technology is likely to reduce marketing costs and accelerate this growth. • Where human consumption of processed cassava has declined, this has largely been the result of government subsidies of competing cereals. These trends are already being reversed as these subsidies are removed, and demand for cassava in processed form can be expected to grow. These products will continue to serve as an important source of inexpensive calories to the very poor. • Starch, much of which is used in the production of various foods, is expected to provide a growing demand for cassava, especially in Asia. • A major potential for growth in demand for cassava is as a component of animal feed, chiefly for domestic use in Latin America and for both domestic use and export in Asia. • The growing market for cassava in Asia has already reached the point where production cannot keep up with demando • Cassava is an important tool for equitable development. Its characteristics are such that the benefits of new technology can be targeted to the very people who have normally be en left out of the development process--the poorer segments of the population. The demand studies clearly demonstrated that, firstly, there is sufficient potential demand for cassava production; secondly, that because of its unique qualities, cassava earns an important place in the CGIAR basket of commodities; and thirdly, that its different end uses under different socioeconomic and agroecological conditions imply a decentralized R&D strategy. The nature of consumption patterns, marketing requirements and multiple end uses imply that the traditional germplasm approach alone will neither be sufficient nor effective for cassava R&D to maximize social benefits. Production research must be integrated with activities on processing, utilization, marketing and policymaking. In the light of this assessment, the Program revised its research strategies for the 1987-1992 periodo 4 1.2 Cassava Program 1987-1991 1.2.1 Goal, objectives and areas ot actlvity The activities of the Cassava Program over the period 1987-1992 were set within the framework of the document titled "Global Cassava R&D: The Cassava Demand Studies and Implications tor the Strategias tor the CIAT Cassava Program" (1987). The goal of the Program is stated as follows: ''To contribute materially to increased income and tood supplies of small farmers and to improve tood availability in tropical developing countries." Furthermore, it is made clear that the Cassava Program is not in a posmon to reach this goal on its own; rather that it forms part of a global network dedicated to exploiting cassava as an important traditional rural and urban staple and to developing new forms of utilization tor changing economic circumstances. It Is within this context that the Program, in close collaboration with national and internatlonal agencies, set itself the objectives presented in Box 1.1. The Program's efforts to attain the foregoíng goal and objectives have been centered around the execution of a set of activities governed by the five operational principies that have guided all CIArs activities; namely, Relevance, Equity, Complementarity and Cooperation, Comparative Advantage and Sustainability (see Box 1.2). The specific activitíes undertaken can be grouped into ten broad areas of endeavor: • Assembly 01 a body of'basic knowledge on the crop. This has included the generation of knowledge on basic biology and growth processes; reactions to different environmental conditions; interactions with pests and diseases, their biology and epidemiology; and the physicochemical characteristics of the plant. • Genetic conservatíon and improvement. CIAT has continued to fulfill its mandate to collect, characterize and conserve the world's cassava germplasm. Efficient use of this germplasm has been achieved by concentrating the desirable characters·-for specific ecosystems or end uses--jn elite gene pools that are mada available to national breading programs. • Integrated pest management. The jmportance of individual pests or pest complexes in terms of losses caused and the potential area over which they can cause damage has been evaluated. For the most important pests and diseases, integrated pest management programs are being devised around alternative strategies including phytosanitary control, biological control, host plant resistance and appropriate cultural practices. 5 BOX 1.1 OBJECTIVES 01= ntE CASSAVA PROGRAM 1. Deverop components of production fechnolbgythat form the o~sis . for stable. cassava-based cropping systems with low costs per unit outpuF. .. 2. Develop technology that allows cassava tobegrownon present~Und~reXPIOitad lands. 3. Develbp pr'Dcassihg téchnológy thar makes cassava a low-c~sthigh.:quality, . conveníent food. . . . 4. Devalop bóth production and processing tachnology that Is cóst-cbmpetitlÍte, íncreasesfarmerinconie, and is sufficiently labor intensiVa té generare employment for landless labor. 5. Develop marketing strategíes for' Cásá~iVa that reduce the marketing margino 6. Stimulate thé develbpment of marketsfc)l'cas~ava that provida a stablepricafloor·· . for the rawmaterial, thus providing farmers the incentive to incraase productiOh, , thereby redllcing príce fll,lctuations forthe consumer. . (. Assrst in thtl development ofnoVel~$eSbf cassava that incteás¿ tf1eovét~1I demand fol'the crop. ..... . .. . .. .. 8. Develop w~ste-reducihgtechnologythat íncreases thé percentage of total . productíonthat iS finaliydonsuníed;.<~ . ... . . 9. Stimulate other agenciesto play an~~¡ve rolelh the cassava R&D process .. - "~,, "'" . . '" ;- 10. Increasethe capacityof natibhal~f6g~am~to carry out cass~V~Á&DPrOje¿ts. 6 • Sustainable ac;¡ricultural production systems. Measures to reduce and control soil erosion and to counteract soil depletion have been researched, together with the interaction of cassava with other species--both in multiple cropping systems and in mtation. • Improved root quality. Research has be en undertaken to elucidate genetic, agronomic, environmental and process variables that affect the qualíty of different end products. Simple, effective methods for screening germplasm for desired quality traits are being developed. • Improved preservation and processing technolog)'. The Program has taken a lead role in developing small-scale processing technology for dried cassava, flour and starch, and refining the technology for preserving fresh cassava. • New products and alternativa markets. Changing socioeconomic conditions leed to changes in people's consumption habits and preferences. The Program monitors these changes as a means of identifying new markets for cassava-based products or new ways of bringing traditional products more effectively into the changing market structure. • Development of an appropriate polic)' environment for cassava. Information is provided to national policymakers about the contribution that cassava can make to achieving economic development goals and about the effects that certain policy decisions can have on supporting or stifling that contribution. • Institutional r'nodels for development with eQuity. Through integrated R&D projects in selected countries, the Program has fostered the vertical integration of national cassava R&D systems and the horizontal integration of production, processing and marketing activities. These are considered vital elements for achieving the adoption of new cassava technology and provide the basis for bringing about equitable development in cassava-growing regions. • National proaram capacity. Training of national program personnel, the realizatíon of joínt projects and ínformation exchange--through a newsletter, workshops, regional networks and the provision of bibliographic material-have been the principal elements for strengthening national programs' capacity to implement relevant cassava R&D activities. 1.3 Program Organizatlon The Cassava Program has organized the activities contemplated within these ten areas of endeavor ¡nto three general groups (1) Strategic and applied research, (2) regional programs and (3) networks. 7 BOX 1.2 OPERATIONAL PRINCIPLES OF THE CASSAVAPROGRAM Relevance. Program efforts are directed toward increased foodpfodLlction that favors the poorer segments of the population, .rather than increasing the body of scientific knowledge per se. Research. isfocused on solving themost important problems in the major cassava-growing regions served by the Programo Eguitv. In seeking to contribute to increased food production, the Pr'ogram orients its efforts toward benefiting primarily thesmall, resource-poor farmer,as welf as the low-income urban and rural consumers. . Complementarity& cooperation ... Theprogram's activities représentonly one segment on the cassava R&D spectrlJl1{Consequently, allactiliitiesare designed to complemellt those of other organizatións. Of particular importanceare the efforts in support of<::assavaby national agrlcultural R&D agencies in coyntries that the Cassava Program seeks to serve. The Program maíntains strong linkagas with these agenciesandcoordinates its work program dosely with those ofeollaborating national programs. . Comparativeadvantage. 111e progra~.66~cenifateson resolvingJhÓseproble~s and delielopiúg those techniquesandmethodologies for whichit hasadaa!' comparative advantage in relaUon to othe(agencies, be they internationalorhatiohal, public or private .. The comparative advahtage arises from certain c)"iaracteristicsQf eor'nmodity r~~earch prograrns ininterh~.tionalbenters. including:@theabilitiJó bring acritiCa¡mass of sciéntists tobearohproblems; (b) thefacilitY tomOlie information,bl15logical controlagents;ahdgenétícf maferialsacróssbór'ders; (e)a high degréeófeontinüíty ofeffort; (dJlódation inthelropics~ (e) beihgin apósitioh to coordinateaetivities of difterent entitiéson a regionalor internatiol'jallevel; (f). the ability to capitalize on economies of scale; (g) and the ability to taka áfohg-term view . of the overaUR&.D process. .. ..... . . - -- - . Sustainability(The Program, in its endeavor to make a majar and lasting contribution to cassava pr¡)dlJction and consumption, emphasizes long-term cLlmulative gains . over dramatic>short-term impacts. Hencs it seeksto develop produetion technotogy that promisesenvironmental harmony;relatively broad adaptation to irnportant production constraints; and relatively low requirements for purchased inputs. 8 1.3.1 Strategic and applied research of global significance These activities have been carried out, principally at headquarters (HO), by a critical mass of cassava researchers supported by CIAT's Germplasm Resources Unit (GRU), Biotechnology Research Unit (BRU), Virology Research Unit (VRU), Agroecological Studies Unit (ASU) and Data Services Unit (DSU). Collaborative research is also carried out with research agencies in both developed and developing countries. This effort has been directed toward nonlocation-specific research in the following areas: • Physiology of the crop and identification of desired characteristics • Germplasm collection, characterization, development and distribution • Cassava diseases and arthropod pests • Cassava-based cropping systems • Socioeconomic studies of cassava production and marketing • Cassava quality and utilization 1.3.2 Regional programs Cassava as a commodity plays a different role at different stages of a country's economic development, and this has been reflected in the Program's approach with regard to the types of activity carried out in collaboration with national programs in Africa, Asia and Latin America. The Program receives support from the Training and Communications Support Program (TCSP) in the execution of these activities. 1.3.2.1 Africa. CIA T's support for cassava development in Africa is channeled exclusively through liTA, which has regional responsibility for the crop on that continent. The Program's strategy is therefore to (a) complement the IlTA research program in those are as where research based in Latin America has a comparative advantage over that undertaken in Africa itself and (b) support liTA in those areas of expertise generated from experience with problems similar to those encountered in Africa. The areas of collaborative activity have covered: • Broadening of the genetic base of cassava in Africa through the introduction and preliminary screening 01 materials 1rom the Americas • Identification, characterization, multiplication and shipment of biological agents for controlling the cassava mealybug and green spider mite • Participation in the organization and execution 01 COSCA • Development of varietal screening methodologies 10r tolerance to water stress 1.3.2.2 Asia. In most Asian countries a multiple market structure for cassava has already developed, and opportunities exist for absorbing increased cassava production--either in 9 existing or new markets--if prices are maintained competitive with alternative carbohydrate sources for tood, feed and industry. In addition, the majority of countríes havehuman resources dedicated to cassava R&D, both in the production and postharvest spheres. Consequently, the Program's efforts have been directed toward establishing an effective regional cassava research network, whích íncludes: • Generation, dístribution and evaluatíon of ímprovad germplasm .. Davelopment of ímproved cropping systems incorporating 5011 fertllity maintenance and eros ion control practicas .. Training of national program parsonnel in rasearch techniquas and methodologies .. Socioaconomic studias of cassava production and marketing, adoption and impact in selacted countries .. Assessment of needs for horizontal collaboratíon in postharvest research 1.3.2.3 Latín America. The demand for cassava in latin American countries still relies chiefly on tood markets and on-farm consumption. The demand studies showed that cassava can compete in alternative markets, the development of which has bean constrainad by the nature of price formation in the cassava tood market and government príce and subsidy poli cíes, which haya favored other carbohydrate sources, especially cereals. In addition, national cassava programs have on the whole been underfinanced and have suffered from a relatively high turnover in personnel. Furthermore, there is a very limited capacity at the nationallevel for undertaking activities in the areas of cassava processing and marketing research. To overcome thase constraints, CIAT has based ils activities on: .. Strengthening national cassava research and extensíon programs through training, lechnícal support and collaborative research .. Developíng projects wíth nalional programs that integrate production, processing and marketing activitíes in specific cassava-growing regions .. Consolídating the Panamerican Breeders' Network for the testing and evaluation of CIAT and local germplasm .. Creating a network for cassava R&D in the subtropics 1.3.3 Networks The continued progress in cassava technology development, the consolidation of national cassava programs, and an increased interest in the crop by advanced labs have laid the 10 groundwork for CIAT to catalyze and participate in the formation of a number of networks on cassava: • The International Cassava R&D Network. Members include international, regional and national R&D agencies, individuals in national agencies and advaneed research institutions in developed and devaloping countries that are either working or interested in the crop. Communieation among membars of the natwork is maintained through the Cassava Newsletter, tha information/documentation service on cassava, and contacts made during intarnational scientifie meetings. • Biotechnology Research Network. This raeently formad natwork will provida the mechanism for involving and integrating developed and developing country institutions in cassava-related biotech research through joint projects, scientifie meetings and a newsletter. • Regional and subregional cassava R&D networks. These include the germplasm exehange networks in Asia and Latín America, the agronomy network in Asia, the Southarn Cone (subtropics) cassava R&D network in Latin Arnerica and the Intagrated Projacts Network also in Latin Amarica. These networks are providing the basis for defining regional priorities tor research and opportunities tor horizontal collaboration among countries. 1.4 Human and Financia! Resources Tha activities undertaken by the Cassava Program have been finaneed by resources trom CIAT's cora budget and through complemantary funding 01 special projacts. The BRU, VRU, GRU, ASU, DSU Seed Unit and TCSP hava all contributad with human and financial resources for cassava-ralated activities. Table 1.1 show$ the projectad and actual care-fundad staffing pattern ot tha Program during the period 1987-1991. Although the staffing projactions were approvad by TAC, the positions 01 an additional HQ breeder, a second utilization specialist to work on quality and an economist in Asia have not been filled tor lack of funding. The activities that were to have been covered by the breeder / agronomist position in Brazil are now being largely undartaken through a collaborative project with the Centro Nacional de Pesquisa em Mandioca e Fruticultura (CNPMF) Cruz das Almas (Bahía). Previous and ongoing special projects ara listed in Table 1.2. 11 Tabl.l.L Cassaya Program: Proje01ed and actual oore4unded staffing pattern durlng the perlod 1987-1991. 1987 1988 1989 1990 1991 Projected Actual Projected Actual Projected Actual Projected AchJal Projected Actual Háadguar1&ra Le.d., 1 1 Physiologist 1 1 1 1 1 Palhologlst 1 1 1 1 1 Enlomologist 1 1 1 1 Br_r(s) 1 1 2 2 Economi$t 1 1 Agronomist 1 1 Utllization ..... Specialist(s) 1 2 1 2 1\) ~ Speciallst al liTA 1 1 Asia Bt •• der 1 1 1 Agronomist 1 F t F 1 1 Eoonomlst 1 LallD America AgronomlstfBteeder (Btazil) TOTAl 9 10 10 10 12 10 14 11 15 11 F = Fundlng from exlra.(;G ra.our ..... Table 1.2. Cassava Program: Previous and ongoing complementary activities financed through special projects, 1987·1992. Duration Activity Soil conservation research Green spider mite research Cassava quality research Cassava flour R&D Cassava stareh R&D Germplasm development, semiarid & subtropieal Callaborative study of cassava in Africa2 Human resouree development Geará cassava development, Brazil Manab f cassava development, Ecuador Atlantic Caast cassava development, Colombia Cassava development, Colombia 1 Extension sought. 2 Covering CIAT's contribution only. 3 DRI = Integrated Rural Oevelopment Fund of the Colombian Govt.; PNR = National Rehabilitation Plan of the Colombian Govt. 13 (yr) 3 7 3 2.5 4 5 2 4 3 2 10 2.5 Termination Date Donor Mar. 1993 Germany Mar. 1992 UNDP/IFAD Dee. 1989 UK Dee. 1991 1 IDRe Dee. 1992 Franee Ju!. 1995 IFAD Dee.l990 Rockefeller Dee. 1991 UNDP Feb.1992' Keliogg Mar. 19921 FUNDAGRO Mar. 19921 Fondo DRI3 Nov. 1991 PNR' HEAOQUARTERS-BASEO RESEARCH 2. GERMPLASM MANAGEMENT ANO IMPROVEMENT 3. PHYSIOLOGY 4. QUALITY 5. BIOTECHNOLOGY 6. PATHOLOGY 7. VIROLOGY 8. ENTOMOLOOY ANO ACAROLQGY 9. CROPPING SYSTEMS 10. PLANT NUTRITION ANO SOIL MANAGEMENT 11. SOIL CONSERVATION ANO PROOUCTIVITY 12. PROCESS ANO PROOUCT OEVELOPMENT 15 43 65 83 97 125 137 169 193 209 223 2. GERMPLASM MANAGEMENT ANO IMPROVEMENT Germplasrrrresource management and development has been one of the most important activities within the Cassava Program at CIAT. The Cassava Breeding Section has complemented national and international institutions in the development of components tor sustained improvement of cassava varieties and for promoting them to farmers. This has involved a multidisciplinary effort within the Cassava Program, with other units in CIAT, and close Iinkages with collaborating national and ínternational institutions. Germplasm resources represent a majar component 01 the natural resources base exploited in agriculture, justifying the Program's majar focus on them. New cassava varieties developed from improved gene pools have proved to be one of the most attractive basic components of production technology for stable, cassava-based cropping systems. Improved germplasm can be easily adopted by farmers at 3 low costand with a long-term contribution toward sustainable agricultural systems. The emphasis of the Breeding Section has been to improve the level and stability 01 cassava production and root quaJity, with efficient use of nutrient and water resources, and agrochemical inputs. Priorities and modus operandi at the continental level have be en determined by a combination of factars-sueh as the importance of cassava production, characteristics and strength 01 national programs, available genetic diversity, presence of liTA with responsibility in Africa and funding opportunities. Virtually all national cassava programs have identified varieties as a crucial part of the new production technology they want to develop--an area where CIAT has contributed significantly during the last four years. Improved germplasm generated over 15 years, together with the long-standing presence of C1AT and the readiness with which germplasm can be moved to countries in Latin America and Asia, made it possible to pursue the following objectives: help organize national breeding programs based on the most important constraints and production areas; channel improved germplasm to the levels of advanced experimental and/or on- farm mals, resulting in an increased number of released varieties; and stimulate the more developed national programs to take an active role in cassava breeding with a regional scope. In the case of Arrica, CIAT cooperation with IITA's ongoing cassava improvement efforts has led to an effective flbw of information between the centers and the introduction of broad genetic variability. The expansion of the germplasm base to backstop liTA will result in a more efficient breeding program for the most important agroecosystems where cassava is grown in Africa, as well as the possibility of developing cassava for semiarid regions. Within the worldwide responsibility far cassava germplasm management and development, the main objectives at CIAT have been to: 15 .. Maintain a representative reservoir of genetic variability for M. esculenta and the related wild species; characterize it and evaluate it across different agroecosystems of relevance fer the crop. .. Develop improved gene pools for specific edaphoclimatic zones (ECZs), taking into consideration the most important biotic and abiotic constraints, and market requirements. .. Study the physiological and agronomic bases of root yield (RY), stress adaptation, and root quality, and tha best way to incorporata relatad screaning techniques within cassava breeding programs. .. Develop a model for tarmer participatory research and the improved adoption of new varieties. .. Support national programs through the interchange of cassava germplasm and related information, training in breeding, and promotíon of networking activities among national and international institutions. 2.1 Breedlng Strategles and Technlcal Advances During tha 20 years of work in cassava breeding at CIAT, a broad genetíc base has bean assembled in a germplasm collection, which has also been evaluated across a range of agroclimatic conditions. A breeding scheme was established in the mid-70s, including genetic recombination among complementary parental material, salection within segregating progenies, and recurrent utifization of elita clones, resulting in gene pools with specific adaptation to biotic and physical constraints. 2.1.1 Edaphoclimatlc zone deflnltlon Given the broad ranga of conditions under which cassava is grown and tha difficulty of obtaining singla genotypes with adaptation to all the production situatíons, the Section has worked on developing broad-based gene pools directed to particular regional needs. The components used to define the ECZs were: world importance of the agroecosystem, climatic conditions, predominant soil type and related constraints, pest and disease problems, and use of end producto Seven major ECZs were identitied, sorne of which were further subdivided by utilization criteria. Table 2.1 presents a description of gene pools for defining cassava germplasm development, representativa regions within each subdivision, and the sites where major breeding activities are conducted. Two basic modifications were made to the original gene pool description. Based on \fL recognition 01 the growing importance 01 semiarid regions for expanding cassava production and the available genetic variability for enhanced water use efficiency (WUE), combined with findings in cropping systems that maximize the use of limited water 16 Table 2.1. Revlsed doscrlption ot gene pool. lo, definlng ca ...... germplasm d"""lopmenl. Maln Sites tor No. Oescrlpfion Representativo Courrtries/Regions Breedlng Work Subhumíd ttopics Mexloo (Yucatan Penlnsula); NE Brull; Media Luna NE Thalland; Oomin. Rep.; N. Venezuela (Colombia) lLC Subhumld troplca; Iow HCN Colombia (Attantic Co.SI & Santandero.); Media Luna Panem. (Coció); subhumld ben ot _; (Colombia) Ecuador (NW ooastl 2 Acld soü savannas Me.leo (Tabaaco); Plalns 01 Colombia & Vonezuela; Villa.icenclc Brazll (Carredo) (Colombia) 2LC Acid 8011 savannas; low HeN Cuba: W Alrlea savann •• ; Phlllpplnes; Wlavlcencío Panam. (Qcu) (Colombia) 3 Humid Iroplcallowlands Amazon basin (BrazU, Colombia, Peru); Villavlcanclo West Java & &matra; Malaysla; S. Vietnam (Colombia) 3LC Humid tropical lowIando; EquatorlalWest _ Villavicenclo lowHCN (Colombia) 4LC Mid.alliluda troplca; Iow Anda.n zona; contra! Bruman Palmlr. !-ICN hlghlands; mld.allilude area. ot Nigerla, (Colombia) Cameroon, East Afrlca 5LC Hlgh.aIIItuda troplca; Iow Andean zone; Rwanda; Surundi Popayán HCN (Colombia) 6 &btropics S Brazil; Atgonfina; China; N Vietnam Santa Catarlna (Brazll) 6Le Sublroplca; low HCN Cuba; Paraguay; S Mica Santa Calarln. (Bruil) 7 Semlarld NE BruU NEBrazil 7Le Samlarld; Iow I-ICN NE Colombia; (Guajira) semlatid bolt 01 NEBruil Wosl Alrica; Tanzania; Mozamblque; Ecuador (Coasl) Note: Low HCN gene poolo can Biso movo lo equivalen! ECls indlcated lo, 1ow/hISh HeN pools, but no! vice v.r .... resources, led to a major emphasis in gene pool development for those conditions. Research for this ECZ will have high priority within the Cassava Program as ít comprises sorne of the poorest and highest food-demanding regions of the world, 17 The second modification relates to the growing concern about the effects of hydrocyanic acid (cyanide) on human health. Most gene pools have been divided into a low-cyanide population and one for which cyanide is not considered for parental selection. Low cyanide can not be generalized as a selection criterion given, on the one hand, the restrícted avallable genetic variability wíthín this type of gene pool and, on the other, the preference for high-cyanide varieties for certain processed food. 2.1.2 Germplasm conservation and evaluation Germplasm management--which ¡nvolves the aetivities of colleetion, conservation, characterízation, documentatíon, dístribution and research aimed at improving the efficíency of germplasm conservation and use-has been one of the maln responsíbilities of the Breeding Seetíon and the GRU, with cooperation from the BRU and OSU. The Seetion has focused on Latín American germplasm, given that a large proportion of the total genetic variability exists in this hemisphere. Nevertheless, selection for local adaptation in both Atríca and Asia has resulted in new genetic combinations that should also be preserved. The Section has gathered a representative sample of the genetic variability of the species in its reservoir. Collecting germplasm had a lower priority during the last five-year period, with greater emphasis on characterization, documentation, identification of duplicates, definition 01 a core collection, and the initiation of a broad germplasm colleetion for wild Manihot relatives. 2.1.2.1 Status of germplasm colleetíon. Started in 1969, the cassava germplasm collection has grown gradually, with a total of 5035 accessions assembled In the field gene bank (Table 2.2). There will seon be a sígnificant increase with the forthcoming íntroduction of BOO accessíons trom Brazil, giving proper representation to one of the most ímportant centers of diversity for cassava. Other important contributions have come trom Asian and Nigerian introductions. Asian clones have represented a source of bread adaptation, as well as being one ot the bases for our cooperation with Asian national programs. Clones from liTA are already being used as parents tor the germplasm introduced to Atrica, providing a greater chance for immediate adaptation. There are several gaps to be filled in order to complete a world collection. For some countries of intermediate to high importance in cassava production, there are no accessions, or common landraces are missing among the few accessions avallable. The absence of a centralized African germplasm collection has restricted representation trom that continent. IITA's efforts to compile information on national colleetions and to consolidate an African cassava germplasm collection, together with recent advances in virus indexing techniques, will make it possible to have a better representation of the genetic variability available in Africa in the near future. 18 rabie 2.2. Descriptiva information available on the CIAT cassava germplasm colIeetion. '" w!th Passport Data 1 % with %wlth '" wlth No. 01 Agronomic lsozyme Morpholog. Origln Aoeessions BaSte Complet<> Evaluation2 Charaotertzation Charaeter. Argentina 15 O O 100 O 94 Bolivia 3 O O 100 O 07 Brazll 1085 25 O 73 72 50 China 2 O O 100 50 100 Colombia 2010 50 30 90 56 100 Costa Alea 147 O O 100 93 90 Cuba 74 O O 100 100 100 Dom. Rop. 5 O O 100 100 100 Ecuador 117 90 O 100 91 100 Fijl 6 O O 100 100 100 Guatemala 91 O O 100 as 90 tndonesia 51 O O 100 100 as Malaysia 68 O O 100 97 96 Mexico 100 70 O 100 81 56 Panama 42 50 O 100 95 95 Paraguay 192 10 90 95 68 83 Peru 405 40 O 95 88 88 Philipplnes 6 O O 100 100 O Puerto Rico 15 100 O 100 100 100 Thailand 8 100 O 100 100 100 Uni!<>d SI.tes 9 O O 100 89 89 Venezuéla. 240 SO O 100 98 100 CIATel.n •• 324 100 68 100 lITA oIon •• 19 100 84 11 Total S035 40" 17 98 89 90 1 Basic: origin; date of eollection; complete: orlgin; date o, oolfection; slte description; qualities and uses. 2 Evaluation in at 18uf ano majar ecosystem. 3 Excludlng clon ••. The present status of the wild Manihot species collection is given in Table 2.3. This represents the starting point of a long-term strategy for exploiting these potentially valuable genetic resources. Crossability studies, preliminary evaluation and biochemical characterization have begun, resulting in much greater polymorphism than that observed in cultivated cassava. Isozyme characterization and ONA fingerprinting will provide the basis for ciarifying some of the evolutionary patterns within the genus Manihot. 2.1.2.2 An integrated germplasm conservation strategy. The main responsibility of elAT, as curator of the world cassava germplasm, is to assure long-term conservation at low levels of risk of loss, contamination by pests or pathogens, or genetic modifications. The collection is maintained in the field, and 95% of the accessions haya been transferred to in vitro culture tor maintenance under slow-growth conditions. Not only is the in vitro 19 rabie 2.3. Wild Manihot speci •• maintalned In vitro al CIAT. Species M. aesculifolia M. a1utace8 M. anomaJa M. brachiloba M. CatlruJescflns M. carthagínensis M. ch/orosti<:ta M. cecropíaéfolia M. crassisepala M. eproinosa M. filamentosa M. fll!bellifolia M. frutic"losa M. glandulifolia M. glazlovil M. grnhami M. guarnnltíca M. hastatllaba M. JrwinH M. jlilCObinensio M. longlpetiolala M. mfchaeJis M. orbicular/$ M. pe/tata M. penfaphyi/a M. pilosa M. pseudoglazlovil M. purpUf8O 5 yr. 2 Excludes oon ... rvatícn acüvitie. in Afriea roletod lo COfII and 00_ coileotícns. " Ouplleatod at 2 O1her sitos (o.g •• liTA. CENARGEN. CTCRI. \/ISCA. CATIE). A Assuming .uccossful dovolopmon! of cryop«taOfYaIIn in !ho madium termo o Assuming cryopreservaflon is not dovolopad in !he madium term. Eslimatod No. 01 _. 5035 46Q2 5477 4900 600 600 4000 4000 2000 roo roo 500 roo roo 4500 4500 3000 3000 roo roo 600 roo roo roo 2.1.2.4 Core collection defjnition. A core collection representing the total diversity of the specíes Oncluding 13% of CIAT's total collection) was selected. It is anticípated that the core collection will be the first part of the collection to be duplicated at another institution, to be placad in cryopreservation, and to be evaluated intensively tor specífic new traits. 22 T able 2.5. evaluation 01 aS .... '.' .... bando for ""n ... tency acro .. repeoled sampl ... ' Ftequ.ncy No. of Groups No, Inconslstent in Sample Frequency In Frequency in with Inconsia. Bond Bond (No.) Sample (%) AlI Germplasm tena.s F/8quency Ranking 1 85 11.1 6.4 12 14.1 16 2 56 7.3 10.1 35 62.5 20 3 334 43.8 44.3 30 9.0 11 4 253 33.2 33.4 19 7.5 10 5 114 14.9 20.2 27 23.7 17 6 345 45.2 46.2 48 13.9 15 7 474 9.7 6,7 23 31.1 19 a 430 56.4 47.2 49 11.4 13 9 306 40.1 34.8 21 6,9 8 10 519 66.0 71.2 19 3.7 4 11 o 0.0 0.1 o 12 42 5.5 5.5 o 0.0 1 13 55 7.2 8.8 1 1.8 3 14 35 4.6 2.1 2 5.7 7 15 353 47.6 50.9 33 9.1 12 16 o 0.0 0.1 o 17 24 3.1 1.5 7 29.2 18 18 125 16.4 13.8 16 12.8 14 19 177 23.2 28.1 13 1.3 9 20 262 34.3 22.6 13 5.0 6 21 183 24.0 29.4 7 3.8 5 22 52B 69.2 66.4 5 Q.9 2 1 OOla from 783 sampl •• wilh 314 clon .. (avg 012.43 repo/clone). Criteria for selecting aeeessions for the eore eolleetion fell within four majar groups: (a) geagraphic origin (the most important); (b) diversity of morphological deseriptors; (e) diversity of uB-esterase banding pattern; and (d) a priori selection of aeeessions based on a predetermined criterion of speeifie interest. Tabla 2.6 presents a deseriptíon of parameters and weighting factors used in determining the no. of aecessions ehosen from eaeh eountry of origino . 2.1.2.5 Germplasm sereening. Agronomie evaluation of cassava germplasm aeross a wida range of growing eonditions provided informatíon tar detecting sources of resistanee to physical and biological stresses common to cassava-produeing areas worldwide, and for selecting high-potential and¡or widely adapted clones for crossing blocks and for national programs. The great majority of accesslons from Asia, all the clones from liTA, and small parts of the eollections received from Brazil, Colombia and Peru were evaluated over the last five years. Table 2.7 summarizes characteristíes of the mast distinguished clones. Asian clones excelled for their broad adaptation, high resistance to thrips and moderate resistanee to cassava bacterial blight (CBB). African clones were eharaeterized by their susceptíbility to mites. profusa branching, low dry mattar (DM) and intermediate 23 !n f~! J~} Í>" 1; ~~g it ~"I~ il'" § #. i¡~; J~U ~~lz w 24 to low RY potential. Accessions from Brazil with stable performance were selected as elite clones for ECZ 1. 2.1.3 Breedlng methodology development The breeding scheme implemented by the Section (Fig. 2.1) allows for selecting genotypes adapted to specific ECZs while offering the possibility of screening for genotypes with broad adaptation across the main ECZs. Several clones selected independently in ECZ 1 and ECZ 2 over the last four years constitute an important genetic base to be transferred to national programs working in transitional or highly variable conditions, and to be used in building gene pools for enhancing broad adaptation. Other strategies are being developed in order to improve the chances of selecting genotypes for other regions within an ECZ, outside CIAT's gene pools. An ideal breeding scheme for cassava would include the development of segregating progenies at CIAT and their subsequent evaluation within the same ECZ in different regions of the world. With the information on avg performance, parental material maintained at CIAT could be selected for recombination and families produced for the following cycle. The strategy will also allow breeding for constraints present on one continent but with the potential of being introduced into others (e.g., African cassava mosaic virus, ACMV). Broad-sense heritability and realized response to selection were estimated using information from the performance of sets of clones in consecutive years from 1981-90. Traits such as HCN content, % DM and reaction to superelongation disease (SED) can be used as primary selection criteria at early stages of a cassava breeding program, given their relatively high heritability (Table 2.8). Other traits, which are highly affected by the environment (RY), by inoculum pressure (CBB), or complex traits (foliage and root evaluatibn) need to be selected for under replicated and preterably multisite evaluation trials. Figure 2.2 A&B shows the genetic progress for RY at the main selection sites for the principal ECZs. The trend has been more consistent for ECZ 1 than for ECZ 2. The decline observed tor the later period in ECZ 2 is possibly related to the change in selection site from a high-disease pressure site (Carimagua) to a less stressful environment (Villavicencio). Emphasis has been on selecting elite, but not closely related parental clones for gene pool construction. A small number of clones with high RY potential and relatively broad adaptation have been used as parents; many of the elite clones selected by the Section have them in their pedigree (Table 2.9). Recombination among clones that have common ancestors may lead to considerable levels of inbreeding depression in subsequent generations. 25 Table 2.7. Avg performance DI sel.Clod gormplasm &CC8$$lons ovaluated in common tri.,. al differen! Eez sites. AY Commer. Clone (t!ha) HI _/pl 'loDM HCN TMp. seo a ea , (Media luna)' M Sta '91 13.4 0.50 3.0 34.6 6.0 1.0 M Sta 363 13.3 0.61 3.6 31.2 8.0 1.0 M Sta 390 '3.6 0.53 3.5 34.8 8.5 1.0 M Sta 589 22.1 0.52 5.0 3'.9 8.0 2.0 Mlnd 8 16.0 0.54 4.2 30.8 5.0 3.0 M Tal 5 '9.' 0.64 4.4 35.3 9.0 2.0 M Taí 8 20.0 0.64 4.9 30.8 9.0 2.0 Check. ,3.8 0.54 2.9 29.9 7.2 2.5 b. Eez 2 (la Ubertad)2 MArg 7 22.0 0.53 3.0 35.0 6.0 2 1 MArg 13 26.2 0.60 3.2 33.8 9.0 2 1 M Sta 97 24.9 0.60 3.2 34.3 9.0 3 1 Cheoks 16.1 o.s9 2.7 31.0 6.4 3 1.8 c. Eez 1 and i' M Ind 39 19.0 0.53 3.1 35.0 6.0 1.5 2 M Mai 37 22.4 0.60 3.3 33.8 9.0 1.0 2 1 M Mal 48 23.4 0.60 3.4 34.3 9.0 1.0 2 1 Check. 15.8 0.59 2.7 3t.0 6.S 2.5 3 1.6 I Mean. of trlal. GY8755. GY5510. GY5524. GY9024 • Means of trial. GYe715. GY5519, GY5910. GY9Q15 3 Means of trlal. GY5517, GY8924, GY9015, GY9024 2.1.4 Gene pool development for edaphoclimatlc zones Because of the wide range of conditions under which cassava is grown, the strategy of the CIAT Breeding Saetion has been to develop gene pools with adaptation to the main biological and physical environmental stresses for each of the ECZS. The strategy is based on the homology among environmental conditions found at particular sites in Colombia and those conditions in the predominant cassava-production areas of the world. Selected clones for a particular ECZ are used as parents for the next cye/e of selection, and they are also shipped in vitro to national programs in arder to be evaluated as potential new varieties. The link between cassava breeding at the national program level and population improvement at CIAT is the interchange of germplasm and information about its performance. 26 1\) .,,¡ Selectlon stage HybrldlZa!lon F, F, C, CO Descrlptlo n Cantralled croases Polyerosses From hybrld sead Flra! clonel generatlon Secondary SeJectlon Slte' Deslgn Principal Selecllon Slte CIAT HQ • o • ~0 "~4 • othsr Evaluatlon Sitas' EPR Observetlonel trlal' Prellmínary yleld trlal •••• •••• •••• •••• •••• ~Q ~~. • • ~ ER Yleld triai ••••• • •••• • •••• Year1 ••••• • •••• ••••• ••••• ••••• • •••• ••••• • •••• ••••• ••••• • •••• ..... 0 Yleldtrlal ••••• • •••• ••••• Vear 2 ••••• ••••• • •••• ••••• ••••• • •••• ••••• ••••• • •••• ER·· ••••• • •••• • •••• a Sama Eez as principalseleclion site b Eez lor diflerent pl10cipal selection site e Se/Odion at principal selec:tion site but stakes fO( ee from CIAT Ha d Accessions "cm germplasm collecoon amer al Ihis slake ••••• • •••• • •••• • •••• • •••• • •••• • •••• ••••• • •••• • •••• • ••• • ••• •••• •••• 0 . .... • •••• • •••• • •••• • •••• 0 ••••• • •••• • •••• • •••• • •••• ••••• ••••• ••••• ••••• ••••• ••••• ••••• ••••• ••••• • ••••••••• ••••• ••••• ••••• • •••• • • • • • {> 0 • ••• • ••• • ••• • ••• • ••• • ••• • ••• • ••• • ••• • ••• 0 0 • •••• • •••• • •••• • •••• ••••• • •••• • •••• • •••• • •••• • •••• ••••• • •••• • •••• • •••• • •••• • •••• • •••• • •••• • •••• • •••• Figure 2.1 5tages 01 selectloo lo CIAT Cassava Breedlog Program ••••• • •••• • •••• • •••• • •••• Table 2.8. amad·sensa herilability esllmates "" dIfIerent EOZs, 1881-90. HON FoIiage Root oae SED Enllironm"'l RY HI %DM Conlerll E.alu.flan Evaluation Resistance Re.lslancé EOZl 0.27 1 0.86 0.47 0.44 0.30 0.10 Sem.A (0.056) 2 ( 0.(68) (0.088 ) (0.048 ) ( 0.(68) (0.075 ) EOZ 1 0.51 0.69 0.35 0.86 0.56 0.33 Sem.e (0.141 ) (0.142 ) (0.167 ) (0.152) (0.116) (0.136 ) ECZ2 0.28 0.44 0.51 0.63 0.28 0.26 0.33 0.39 I\l Sem.A (0.086 ) (0.072 ) (0.062 ) (0.051 ) (0.066 ) (0.088 ) (0.100 ) (0.061 ) (Xl ECZ2 0.29 0.58 0.43 0.67 0.19 0.14 0.19 0.51 Som.S (0.121) (0.126) (0.118 ) (0.132) (0.092 ) (0.109 ) (0.073 ) (0.174 ) 1 Pooled estimal •. 2 SO (0.05). Progres. 1.5 ~-------------------------, 0.5 o B 0--0 Realízed 0-0 Expected .0.5 L_-L __ ...L __ L_-L __ ..L_---1 __ -L __ ..L_---1 __ ....I 1980 1981 1982 ~983 1984 1985 1986 1987 1988 1989 1990 Progress 0.3 A 0--0 Realized lO\" ~=g 0-0 Expected '-0_/ ~c 7a o 0.2 0.1 o ·0.1 ..().2 ° ..().3L __ -L __ -....IL __ -L ____ L __ -L __ -....IL __ -L __ -....IL __ -L __ -....I 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 v.ars Figure 2.2. Genetic Progress in root yield: (A) ECZ 1- Semester A and (B) ECZ 2- Semester A 29 Table 2,9, Frequenoy of uM 01 germpfasm accessions as patents ín elite clones ~ncfuding multlple erossea).1 Germplasm Accesslcn fCZ Frequency (%) M Col 22 ~lV 'ZT.7 M Col 1684 ~lV 20.9 M Sra 12 ~lV 7.2 SubtotaJ for fCZs ~ 55.8 M Col 647 I~III 16.3 M Col 1438 ¡¡·III 13.3 M Col 636 11·111 10.4 M Ven 77 I~III 7.4 SubtotaJ for fez. 1~ln 41.4 M Col 2060 V 22.9 M Col 1522 V 17.8 M Eou 169 V 14.7 M Col 113 V 13.1 Sublotal fot eez v 68.5 1 To1aJ 01 283 elite olcn8$ .nalyzed. 2.1.4.1 Lowland tropjcs with low to intermediate rainfa!! and long dry season (!;CZ 1). Media Luna (Magdalena) is the príncipal selection site for ECZ 1, with El Carmen (Sollvar) being a secondary location and the main site tor interactíon with ICA's cassava programo The ideotype sought tor ECZ 1 iocludes the combinatioo 01 good plant type, high RY, high DM and adequate levels of resistance to diseases (CSS and anthracnose) and pests (mites and thrips). Table 2.10 presents the elite clones selected from 1987-91 and their main characteristics. Local check M Col 2215 represents a high standard in terms of DM production and stability, and cooking quality. With a combination of superior RY poteotial and similar DM conteot, elite clones significantly outyielded the cheek in DM production¡ha. Greater emphasís is being given to cooking quality in advanced yield trials (ERS) io arder to prevent losíog valuable genetíe variability tor that eomplex traít when selecting tor just RY and DM production potential. 2.1.4.2 Lowland tropios wlth acid soil savannas and high rainrall (ECZ 2). Cassava breeding activities were moved from Carimagua (inner Eastern Plains) to ICA-La Ubertad (Piedmont) in 1987. The new site is characterízed by relatively high pressure from dry- season pests (green mites and mealybugs) and erratie incidenee 01 CSS and SED. The strategies implemented to enhance disease pressure involved susceptible spreader rows, artificial inoculation and crop rotation. Until reliable high disease pressure is aehieved, the Section risks selecting elite clones that will succumb under heavy disease pressure; or conversely, eliminating clones of avg productivity that could be a valuable source of disease resistanee. 30 Table 2.10. Adjusted perlonnance 01 ellle clones Ior ECZ 1. seie<:led frQm 1987-91. Reslstance 10:3 No. Paren! Clones RY Oommer. coo~ Clone Female x Male (tJha) HI Roots/pl 'lbOM HCN1 Quar Thlips M~.5 CM 3320- 4 M Br. 20 x CM 523- 7 19.4 0.60 3.8 32.7 9.0 4.0 1.0 1.0 CM 3&92- 9 CM 681-2 x M Col 1468 15.0 0.61 4.0 33.3 6.0 2.8 1.2 1.2 CM 3997·1 CM 681-2 x CM 849·1 12.9 0.41 2.0 30.4 8.5 3.0 1.5 12 CM 4013·1 CM 922-2 x CM 728· 2 16.5 0.55 3,7 32.0 8.0 ~.4 1.0 1.3 CM 4042-4 CM 1015·13. CM 180· 5 20.8 0.82 2.6 30.1 8.2 3.2 1.0 1.1 W CM 4063-6 CM 1015-42 x CM 849-1 20.6 0.62 3.2 29.3 8,8 4.2 1.1 1.3 .... CM 4365-1 CM 976-15 x M 0012207 17.0 0.58 2.2 35.9 4.8 1.3 1.5 12 CM 4733·2 M ara 12 x M Col 72 15.3 0.58 3.3 33.9 6.3 2.0 1.4 1.0 CM4m-4 M OoI22xM Mal 3 17.3 0.58 4.0 32.1 4.2 1.8 1.0 1.2 CM4m-2 M 001 72 x CM 523- 7 14.8 0.83 2.3 37.6 6.0 1.6 1.0 1.2 CM 4919·1 M 0012207. $M 301· 3 19.7 0.71 3.8 33.7 8.5 3,2 1.2 1.0 CM 5586-1 CM 681- 2 x M 0012215 16.0 0.55 4.5 35.2 5.5 1.3 1.0 1.3 M 0012215 10.2 0.47 3.0 34.2 5,2 1.8 1.6 1.3 1 1 "very low; 9 = very high 2 1 " excellenl; 5 = very poor 3 1 = highly resistant; 5 = highly susceptible Recently selected elite clones for ECZ 2 are presented in Table 2.11, along with their main characteristics. Most of the clones combine good RY with quality, representing an adequate genetic base for the developing cassava in the piedmont region. 2.1 A.3 High-altitude tropícs (ECZ 5). Breeding activities have been developed in the region of Popayán (alt. 1800 m). Due to the potential importance of CSS in this ecosystem, greater efforts are being made to incorporate resístance to that pathogen. Elites clones are also being evaluated in Carimagua for their resistance to the disease. The genotype X altitude interaction has been recognized as being very important when breeding cassava tar high-altitude trapics. In arder to incorporate that factor into the breeding scheme and broaden the range of adaptation of the developed gene pools, two strategies were developed: Elite clones for middle- (ECZ 4) and high-altitude (ECZ 5) zones are under evaluation at an intermediate site (alt 1400 m) in Mondomo (Cauca) to determine their adaptation range. The second strategy includes the simultaneous evafuation of segregating material at Popayén and at CIAT-HQ (alt. 1000 masl), selecting only genotypes that show good performance at Popayán and at least reasonable adaptation at CIAT HQ. Table 2.12 gives the main characteristics of the elite clones selected for ECZ 5. These clones represent the basis for Mure cooperation with cassava-production areas in high-altitude tropics, particularly the Eastern Africa highlands. 2.1.5 Varíetal evaluatlon with farmers Evaluation of cassava varieties with the participation of farmers during 1987-91 has provided information for defining a methodological model for participatory research and refining selection criteria for varietal improvement. Based on the results from ERs, CIAT breeders and ICA scientists sefected elite clones to be evaluated in regional and farmer- managed trials on the North Coast of Colombia. Open interviews provided relevant information at the farm level on production, market, teahnology demand and farmers' ariteria for choosing/rejecting different alternatives. CIAT's DSU combined common regression, principal component and stability analyses in an integrated approach to qualífy the relevance of different evaluated traits for cassava growers. Farmers gave major emphasis to criteria related te the fresh root market such as root no., slze and external color. Table 2.13 presents results for 26 clones evaluated by 379 farmers. Several experimental clones resulted with higher acceptability than the local varieties. In particular CG 1141-1 and CM 3306-4 always received hígh and stable qualifications at both early and late harvest. Many farmers in the region have been adopting these clones before release as a result of their outstanding trial performance. The farmer participatory research model and means for improving the adoption rate of new varieties were discussed in a three-day workshop held at CIAT in September 1991. It was concluded that new varieties should be selected according to farmers' preferences and that the release process must be accompanied by an organizad system far multiplying planting material and a proper followup of adoption rate. To extend the use 32 Table 2.11. AdJusted performance 01 elite clones lar ECZ2; seleoted from 1987·91. Reslstance to:3 No. Parent Clone. RY Commer. C~ Olone Fomel. xMaIe (lIha) HI Root./pl ""OM HON' Qu OBB SEO CM 3311· 3 M Col 1468 x CM 523· 1 31.5 0.59 3.5 34.1 1.5 22 2.7 1.1 CM3380-7 CM 596- 1 x CM 523- 7 16.6 0.55 2.3 31.0 6.8 3.0 2.7 1.6 OM 4157-34 CM596-1 xMBra12 23.0 0.52 2.5 33.1 7.0 2.2 2.5 1.0 CM 4402·4 CM 1335- 4 x M Ven n 25.8 0.59 2.3 32.0 4.0 1.8 22 1.3 ~ CM 4484-2 M 0011468 x M Cub 74 20.5 0.47 2.0 33.9 5.7 2.0 211 1.5 CM 4729- 4 M Br. 5 xM Cub 31 18.7 0.57 2.0 35.5 8.0 2.1 2.5 1.3 CM 4793·1 . M Col 72 xMVenn 18.6 0.60 2.3 32.1 8.0 2.2 2.7 1.0 CM 5253-1 CM 1223-11 x CM 523- 7 27.9 0.57 3.0 38.2 8.0 1.4 2.0 1.3 CM 5286-3 CM 1335- 4 x CM 1223--1 23.1 0.62 3.0 32.6 7.5 2.3 2.6 1.3 M Ven n 7.4 0.51 2.0 3Q.4 6.2 2.0 2.3 1.4 1 1 • very Iow; 9 • very hlgh 2 1 • excellent; 5 • very peor 3 1 • hlghly resistan!; 5 • hlghly susceptible Tablo 2.12. Elite clones selected under highland condltions (ECZ 5). Resistance to:3 Common Names or No. 01 Vi.1d Ht(N Cook. Branching Clone Parenta Trlals (l/ha) %OM Qual.2 HI Lev.1s Thrips CSS SEO Phoma cn 354- 2 M Col 309 x M Col 1468 4 21.7 35 4 0.51 3 3 4 5 3.3 cn 358- 3 M Col 335 x M Col 2060 5 18.0 35 3 4 0.35 4 3 3 5 2.7 en 401· 3 M Col 1522 x M Col 340 6 21.5 33 6 4 0.42 3 2: 4 4 2.9 en 402·11 M Col 1522 x M Col 647 1) 27.3 32 1) 3 0.63 3 2 2.9 CG 406- 6 M Col 1522 x M Ecu 169 11 23.5 35 4 1 0.49 4 1 4 5 2.9 CG 461· 3 M Col 2060. M Eou 169 6 27.0 35 4 2 0.57 4 4 3.0 en 501· 2: M Ecu 169. M Col 2060 11 26.9 35 4 2 0.58 3 1 3.1 CG 501· 16 M Ecu 169. M Col 2060 4 23.1 37 4 0.51 4 3.1 en 1.116- 32 M Col 2017 x M Col 2OIlO 3 21.5 35 5 0.39 4 2.7 cn 1118-118 M Col 2017 x M Col 2OIlO 4 30.3 37 3 2 0.49 4 .2 2.7 CG 1118-121 M Col 2017 x M Col 2OIlO 4 31.8 37 3 2 0.63 3 2.7 CG 1231· 3 M Col 1522 x M Col 2017 4 27.0 34 4 2 0.55 2 3.3 (.) CM4468-4 M Col 1522 x M Col 2OIlO 4 29.4 34 3 3 0.43 4 .2 2.9 ~ SG 350- 23 M Col 1522 4 23.6 34 4 3 0.39 .2 3.1 SG 350- 42 M Col 1522 4 26.9 35 5 0.56 3 3.1 SG 427· 87 M Col 2060 4 37.7 37 6 3 0.52 4 .2 3.0 SG 629- 4 M Col 1466 3 24.0 34 5 0.47 2 4 3.1 SG 638- 6 M Col 2061 4 32.1 35 5 4 0.56 3 3.0 SM 524- 1 M Col 1522 .2 22.7 32 5 0.53 3 2 3.7 M Col 1522 Algodona 19 22.9 32 4 3 0.42 4 4 5 2.5 M Col 2059 Setadovlo Setlo. 15 16.3 32 3 2 0.31 4 .2 5 5 2.7 M Col2OllO Regional Amarilla 17 16.9 33 3 4 0.32 4 4 4 2.5 M Col 2061 Regional Morada 15 15.0 31 3 2 0.47 3 .2 4 5 3.1 M Col 2351 PIlnarnoi\a 4 23.2 34 4 2 0.47 4 1 2.6 1) 1 • v''Y low; 9 = axtremely hlgh 2) 1 = e.eollonl; 5 = very peor 3) 1 = hlghly resistanl; 5 • hlghly 8usooptiblo Table 2.13. Proportion 01 total variation, explai!led by the first three principal components, 1987-91. Variables Root color: • External - Aesh Planl heighl Branching General evaluation RoOI no. ~ield) Evaluation by Researcher 0.646 0.743 0.4791 0.5405 0.497 Farmer Evaluation 0.932 0.552 0.457 0.4389 Components 2 3 Comparative Proportion 01 Explained Variatlon 17.54 34.58 45.91 of participatory research methods in the selection of improved varieties, training of personnel participating in integrated projects with a varietal component and of scientists who play key roles within regional networks will have high priority over the next few years. 2.1.6 True seed propagation The Cassava Program has developed a project to study the potential tor and to conduct research on the major constraints to developing a TCS alternative tor commercial cassava production. Comparisons between stake and seed progenies derivad from the sama parental clone showed comparable potential fer TCS in terms of root production, but disadvantages in terms of early crop development (fabla 2.14). Results from a factorial experiment involving different true seed populations and seed coating with two sources of P (rock and superphosphate) are presented in Table 2.15. The more soluble source of P seems to affect seed germination; whereas rock phosphate does not affect germination and increases seedling vigor, improving the crop's ability to compete with weeds. Seed germination under suboptimal conditions and early crop establishment are the most important constraints to be overcome in TCS research. 2.2 National Program Collaboration Breeding work at CIAT is oriented toward helping national programs better maat their objectives in varietal development. Effective collaboration results in more efficient and productive national programs. The interchange of knowledge (training and network development) and germplasm with national and international institutions is a major form of institutional support and strengthening undertaken by the Programo 35 Table 2.14. Comparison botw •• n clone CM 340-3Q and lis open-pollinated progenywhen seod propagaled. Propagation lA! al AY Tolal Ptanl Gtoup System 120 days ~/ha) '('",Id (t¡ha) HI CM340-3Q Vegetative 5.4 41 91 0.48 Progeny TCS 2.4 43 78 0.54 lSD (0.05) 0.9 8.7 9.2 0.07 TatNe 2.15. Effect of seed coat¡ng with two sources of P 00 two seeO popoJations, Populatioo Source of P % Germination Vigor1 SM1557 None 39.4 3.2 Rock phosphalo 48.7 4.3 Super phosphate 11.7 3.5 Mean 33.3 3.7 SM 1559 Nono 49.3 3.5 Rack phosphate 40.2 4.6 Supar phosphate 7.2 3.6 Mean 32.2 3.9 , 1: pool'. 5: 8xcellent. 2.2.1 Tralnlng Most of the training in cassava breeding has been provided to those participants in general production courses who, having certain interest in breeding, undertake a period of in-service training. This type of training has become a valuable tool not only for imparting knowledge but also for exchanging information and definíng common objectives and breeding methodologies. 2.2.2 Network development The Panamerican Cassava Breeders Network met at CIAT in 1987 and at CNPMF (Brazil) in 1990. The main purposes of the meetings were to exchange information on the status of the crop and research related to cassava varietal improvement in each country; and to analyze and propose specific breeding activities to be conducted in a network approach. The most relevant propased network activities are related to (a) characterization of the main environments for cassava in Latín America, using genotype by environment teehniques; (b) establishment of guarantees for cassava germplasm movement; (e) development of a basic set of traits tor charaeterization and agronomie 36 evaluation of breeding material; (d) conceptualizatíon 01 stake multiplicatíon schemes for different continental situations; and (e) improvement ofthe communication channels within the network. . Developments in the Asian Breeders Network are detailed in Chapo 20. 2.2.3 Germplasm exchange CIAT's cassava germplasm constitutes the main support for national program breeding activities and advanced research projects. Sexual seed from impreved gene pools represents a broader range of genetic variability for selection than the introduction of a set of clones vía in vitre culture or indexed stakes. Table 2.16 presents a summary of informanon about germplasm shípments from CIAT during 1987-91. There is renewed interest in introducing elite clones or basic germplasm with specific traits, either for evaluation as potential varieties or to be usad as parental clones in crossing blocks. At a globallevel the contiriuous progress made by the Gassava Breeding Section at CIAT is reflectad in an increasing list of ímproved varieties released for the most important cassava-production regions of the world (Table 2.17). 2.2.4 Project for semiarid and subtropical germplasm development CIAT has developed mechanisms to take advantage of special strengths of individual national programs in arder to benefrt other programs in a network manner. A prime example of this is a five-year project by CNPMF/EMBRAPA in Brazil to develop germplasm for semiarid and subtropical ECZs, especially in Africa but also for other continents. The project, started in 1991, is already generating promising results. Clones combining good yield, DM content and mite resistance have been identified and included in polycrosses. Germplasm screening at four selected sites for samiarid zones is expected to provida more information for selecting parents with good yield potential and stability. For the tirst time in 1991, there will be a simultaneous evaluation of the same progenies in Nigeria, NE Brazil and on the Colombian North Coast. The resulting informatian will help determine the homology among the sites chosen by the three institutians for this collaborative praJect. It will also pravide the opportunity for selecting broadly adapted genotypes as the basis for the nex! crossing stage. Activities for the subtropics at the Brazilian State Agricultural Research Agency EMPASC in Santa Catarina included tinding a breeder for the project. A germplasm collection has already been established with more than 700 accessians from neighboring states, a recent germplasm collection fram Santa Catarina, and introductions from Paraguay, Cuba and CIAT elite clones. This collection is the basis tar germplasm development tar the subtropics. The subtropical branch of the proJect will also support the Southern Cone Cassava Develapment Netwark in germplasm development and varietal ímprovement activities (seeChap.13, seco 13.1.1). 37 rabie 2.16. Summary of Internatlonal germplasm shlpments. 1981·91. No. No. No. No. Regían Form Shipments ero .... Seeds Oones South Amerioa Seads 10 681 36915 1089 In Vitro 36 65 Indexad stakes 7 10 Stak ... Mesoamerica Seeds 12 515 In Vitro 10 118 Indexad Slakes 6 65 North Amarica Seedo 9 InVitro 11 74 Indexad $lalce. 3 17 Caribbean Seed. 7 18242 In Vitro 11 75 Indexad slak •• 1 13 Alrica Se.do 13 1074 165290 Europe Seeds 15 126 40245 In Vitro 9 96 Inds;ced stakes 7 16 SI.ke. 13 AsIa S..d. 42 155462 In Vitr" 13 154 Mlddle east In Vitro 1 6 Soulh Pacllio In \litro 6 62 TOlal 226 5347 439413 1873 2.2.5 Varletal development: The Colombian case studies Close collaborative work between CIAT and ICA has resulted in the release of three varieties and the pre-release of another three experimental clones. In 1990 clones CM 523-7 and CM 2177-2 were released as new varieties under the names of ICA-Catumare and ICA-Cebuclln, resp .• for the Eastern Plajns region. Clone CM 2766-5 is in the pre- release stage for the same region. The superiority of the three clones in dry RY with respact to the check M Ven 77 can be sean in Table 2.18. These varieties, togethar with five other elite clones, have participated in on-farm trials, confirming their poten'¡al and farmers' preference. For the North Coast region, experimental clone ca 1141-1 was released as ICA-Costena in 1991; while two other elite genotypes, CM 3306-4 and CM 3555-6, have been pre- 38 Table 2.17. C ..... v. olon •• do.olopad w~h CIAT Input, ,éloasad as rooommendod vlItie~sln AsIa and Latln Amerioa. Vear of IAoatlon of IAoation Potential CateaorxlName Cou~ Rot ..... CIonoICoda HvbridlzaHon 01 Solectlon Adal!!!tion u...a Maln Fealur •• Land¡aoo and nation!!! DI'OO'am clone. er= excellent: 5 ... very poof released. Superior RY. DM content and cooking quality, comparable to the check M Col 2215. are the main characteristics of the three elite clones (Table 2.19). ICA-Costel'la and CM 3306-4 stood out in on-farm tríals for their superiority in nearly all traits and generalizad acceptance by farmers in the region. AlI the aforementioned clones were highly resistant to tha principal diseases and pests in the region of adaptation. The demand for good-quality stake material tor planting was further evidance of a high level of acceptability to tarmers. 40 2.3 Future Perspectives tor Germplasm Development The Breeding Section will continue to support national and international institutions by providing components for sustained improvement of cassava varieties and tor promoting these varieties to farmers. Within this mandate, the main objectives tor the period 1992·96 are to: • Playa leading role worldwide fer cassava and wild Manihot germplasm management activities, especially conservation and characterization • Build upon the knowledge base necessary tor sustained genetic improvement, in areas of genetics, physiology, pest and disease management, crop and soil management, and root quality • Develop improved germplasm through genetic recombination, in the form of broad- based gene pools, from which national programs can directly select superior new clones or, alternatively, use selected clones as parental material in further breeding • Provide training to national program personnel and mechanisms for networking among programs • Develop a production technology based on TCS These objectives necessarily involve a multidisciplinary eftort within the Program and with other Units at CIAT, as well as close linkages with collaborating national and international institutions. ' 41 3. PHYSIOLOGY The goal of the Physiology Seetíon is to conduct strategic and applied research in arder to improve yield potential and stability, and the adaptation of cassava to the different agroecosystems where it is grown. To accomplish this goal, a wide range of research prablems have been addressed in the past five years: • Evaluatíon of cassava germplasm under prolonged water stress and identificatíon of mechanisms that underly tolerance to stress • Basic research to characterize the photosynthetic potential of cassava in relation to crop productivlty • Plant nutrition and soil management research to evaluate cassava germplasm adaptability to poor soils in the tropícs (e.g., a collaborative research project with the Swiss Government facused on elucidating mechan/sms associated with adaptability to low-P solls) • Soil conservation in cassava-based cropping systems on hillsides continued through a collaborative research project with the University of Hohenheim, Germany, oriented toward collecting fundamental data on soil erosion to assess long-term effects on soil degradation and evaluating different crop management systems in relation to soil erosiono Aesults of these research efforts, which are highlighted here as well as elsewhere in this report (see Chapo 10 & 11), have important practical implications for cassava breeding strategies and for developing technologies to improve production systems and to conserve natural resources. 3.1 Yield Stability During Prolonged Midseason Water Stress Most cassava production occurs in the tropics where rainfall distribution is not uniform, with prolonged dry and wet cycles. Farmers usually plant cassava after the onset of a wet cycle, which allows the crop to get established with 3 to 4 mo of sufficient rain. The crop then has to endure a period of up to several months with no rain. Cassava adapts to water shortage by reducing its leaf canopy (CIAT Annual Reports 1986-89), thus limiting crop water use. Moreover, the plant reacts to soil and atmospheric water stress by partially closing leaf stomata (CIAT Annual Aeports 1984-89). Aithough the combined effect of reduced leaf canopy and stomatal closure would improve crop WUE through a slow depletion of the limited soil water, it would also lead to reductions in total biomass and RY (CIAT Annual Reports 1986-89). Reduction in AY would depend on variety, intensíty and duration of water stress. Varietal differences in response to water stress were found to be related to vegetative vigor and to leaf area duration during the growth 43 eyele (Fig. 3.1). Varieties su eh as CM 507-37, which maintain relatively higher than optimal LAI under wet conditíons, were found to produce well under both favorable and stressful environments. In this report some results are presented on the effect of prolonged water stress during the growth cycle on productivity and on certain physiological characteristics. This information is of paramount importance in understanding the mechanisms underlying drought toleranee in cassava. 100 • o~ o • • • ,. t '" so ~ 0 0 / 3 ! • • 1 so J ~[ " zo, " , , , 1 , , , , I '0,5 LO 1.5 2,0 ~5 ao 3.S 4.0 45 Gtow1h cyde averagol !..Al Figure 3.1. Relationshlp between yleld and avg LAI as affected by mldseason water stress. (1) and (2) represent varo M Col 22 and M Mex 59 (1980·81): (3) and (4) represen! var. M Col 1684 and CM 507-37 (1984-85): (5) and (6) represen! varo M Col 1684 and CM 507-37 (1986-87). Arrows Indlcale changes from well- waterad to stress conditions. 44 3.1.1 Canopy development, biomass production and RY Patterns of leaf canopy development, as shown by changes in LAI, are presented in Figure 3.2. It is apparent that in the first 3 to 4 mo clone CM 507-37 produced higher LAI than M Col 1684. At that stage, specific leaf area Qeaf area/unit lea1 dry wt) and leaf area ratio (Ieaf area/unit dry wt of leaf + petiole + stem) were also higher in CM 507-37 (186 ± 11 SO) cm2 / 9 for specific leaf area, and 98 ± 7 cm2! 9 for leaf area ratio) than in M Col 1684 (147 ± 8 cm2/g, and 73 ± 6 cm2!g). 80th var. reached peak LAI at 160 days after planting (OAP) in the unstressed crops, with the highest value in CM 507-37. Thereafter leaf canopies of both varo declined over time. Reduction in LAI was particularly pronounced in M Col 1684 between 160 and 222 OAP; whereas CM 507-37 maintained higher LAI and shed fewer leaves (Fig. 3.2), which may indicate better lea1 retention. Compared with the unstressed crops, water stress significantly reduced lea1 canopy in both varo Reduction in leaf canopy could not be attributed solely to leaf fall as the stressed crops shed fewer leaves than the unstressed ones (Fig. 3.2). After ralease from stress, CM 507-37 recuperated rapidly and formed more new leaves than M Col 1684. 7 S· 5 4· 'i 3 2 36 60 '60 32 28 " ~ 24 " o 210<)1> """"" " "'.6 -Q. "" .- -0.9 >" ,,~ • o ~i -1.2 -1.5' , , , , , , , , 20 ." eo 80 100 20 ." 60 80 100 Oays_ ..... Figure 3.3. Laaf water status and gas axchange as affacted by mldseason water stress. Comparad with the unstrassad leaves, lam COa uptaka rates were about 25% lower in the stressed leavas of both var. throughout the stress periodo The same trend was observed in stomatal conductance; however, the magnitude of reduction in stomatal conductance dua to stress was mora pronounced in CM 507-37 than in M Col 1684. Instantaneous leaf WUE (CO. uptakafH.O loss) was the sama in both strassed and unstressed leaves of both var. Thesa data (not shown) indicate that laavas remained activa in terms ot gas exchange under the prolonged strass. The absenca of a larga dacreasa in '1' l of stressed leaves (Fig. 3.3) might partialfy explain the relatively high and consistent CO. uptake rates. Another factor that might have enhanced CO, uptaka of stressed leaves is the notable decrease in their specifíc leat area. At 43 days of stress, specific leaf areas were 197 ± 11 cm'fg (mean ± SO) and 185 ± 5 cm2/g for CM 507-37 and M Col 1684, resp.; whereas the respective values for the unstressed crops were 233 ± 16 cm2fg and 206 el: 14 cm2fg. 3.1.3 Water uptake in stressed erops and crop WUE Patterns of water uptake of the stressed crops are shown in Figure 3.4. The two varieties extracted a major portian of the available soíl water during the first 35 days ot stress 47 ;).18; ~\ , \ 1.16r' • , \ , \ Q.14~ \ , \ \ ; ',,- \ 0"' \ \ ,,-. \ \ E \ , §. '--~ q, \ a, , \ .. \ \ l' --... ... ! \ \ I .... " ~ \ I :>- \ './ ~-~ \ \ ' \ ....... 0-_-...... ' ..... ~ 0, '-. '''\ ..... ---.-- , \ Soil de¡)th (m) Figure 3.4. Parterns ot water uptake during mldseason water stress. (A). ( ... ) a M Col 1684; (O), (.) a CM 507-37. Open symbols a 35 days; dosed symbols = 96 days. (about 109 mm for CM 507-37 and 93 mm for M Col 1684). At 96 days of stress, total water uptake was about 160 mm in both var., which is equivalent to 70-75% of the water available in 2 m of this soil. M Col 1684 extracted less water from the upper soillayer and more from the lower layer than CM 507-37. These patterns of water uptake suggest that rooting systems and their distribution in the soil profile differed between the two var.: CM 507-37 appears to hava more fibrous roots in the surface layer than M Col 1684. Total dry biomass production (excluding fibrous roots) between 117 and 160 DAP, when crop growth rate was at max., was used to estimate crop WUE (Fig. 3.2). At that stage of growth, LAI was near max. for both stressed and unstressed crops (Fig. 3.2); hence radiation interception was high. Total DM increments over 43 days were 9176 kgjha (213 kg/ha/day) and 8320 kg/ha (194 kg/ha/day) far the unstressed crops of CM 507-37 and M Col 1684. resp. As actual water uptake by the unstressed crops was not determined, a max. water use equivalent to pan evaporation of 190 mm (4.4 mm/day), recorded at a nearby weather station, was assumed. The resulting values of crop WUE in this case 48 were 0.0048 and 0.0044 kg biomass/kg water for CM 507-37 and M Col 1684, resp. For the stressed crops during the same period, increments in total DM were 6410 kg/ha (149 kg/ha/day) and 4574 kg/ha (106 kg/ ha/day) for CM 507-37 and M Col 1684, resp. The estimated water uptake from regular soil sampling in 43 days was 132 mm for CM 507-37 and 125 mm for M Col 1684. These estímates excluded evaporation from the soil, which was covered with plastic, so actual water use might have been underestimated; nevertheless, the crop WUE should be taken as a rough estímate in this case. Resulting values were 0.0049 kg biomass/kg water for CM 507-37 and 0.0037 kg biomass/kg water for M Col 1684. These estimates are similar to those for the unstressed crops. Max. growth rate of cassava with nonlimiting water approaches the lower limit of C. crops such as maize and sorghum, and its short-term WUE is equivalent to these crops. Because of its long growth cycle and low LAI, which prevaíls during a significant portian of the growing season, however, the avg growth rate and the seasonat WUE of cassava are reduced. The results presented herein with tINo genetically related clones (M Col 1684 is a parent of CM 507-37) subjected to prolonged midseason water stress reveal that certain physiotogical and morphological characteristics are apparently associated with tolerance to water stress and yield stability. Most notably, the ability to form rapidly and malntain leaf area to intercept enough solar radiation during the growth cycle Is of paramount importance. Second, genotypic characteristics related to better partitioning o, biomass between leaf and stem such as high specific leaf area (Ieat area/unit leat dry wt) and a high leaf area ratio (Ieat area/unit dry wt of leaves and stems) may lead to higher lavels ot LAI without an adverse effect on RY. Clone CM 507-37 maintained higher LAI and higher RY under both well-watered and water-stress conditions than the parent M Col 1684 (Fig. 3.2 & Table 3.1). It also shed fewer leaves under well-watered conditions (Rg. 3.2) despite high LAI, which might indicate better leaf retention. However, evaluation of a wider range ot genotypes under favorable and water-stress conditions (midseason as well as terminal stress) is warranted. Cassava appears to tolerate prolonged drought through some stress-avoidance mechanisms. In addition to reduction in leat canopy and top growth upon onset ot stress, the leaves clase their stomata partíally while maintaining reasonable CO. uptake rates, thereby reducing water loss and continuing to accumulate DM into storage roots. Moreover, cassava Is capable of extracting deep soil water slowly when avaílable. This is an advantageous characteristic under prolonged drought or with sporadic rainfall normally encountered in cassava-growing regions in the tropics such as Sub-Saharan Africa and NE Brazil. 3.2 Screening for Drought Tolerance and Low HCN Content Given the fact that most cassava varieties show increases in HCN under stress and become less suitable for human consumption in fresh form, research continued to ídentify genotypes with low levels of HCN under stress. Few clones were identified that tended 49 to keep their HCN at reasonably low levels (CIAT Annual Report 1989). These clones were grown at the Quilichao Station for two years to determine, besides yield performance, some physiological parameters related to photosynthesis efficiency as affected by prolonged drought. The clones were grown in the field; and commeneing 100 days after planting, half the experimental area ( .. 2000 m") was subjeeted to 3 mo water stress by eovering the soil surface with white plastie sheets. The other half ot the experimental area received natural preeipitation as well as supplementary irrigation in periods with rainfallless than the potential evapotranspiration ('" 4.4 mm/day). Data on leaf water potential (as an indieator of stress), canopy light intereeption (as an indicator of leat area), and leaf gas exehange (as an indicator of leaf photosynthetie capaeity) were eollected during the entire stress period (Ags. 3.5-3.8). Predawn leaf water potential (Fig. 3.5A) ramained around -5 bars throughout the stress period for all tour genotypes, with virtually no differenees between the stressed and unstressed crops. Midday leat water potential (Fig. 3.5B) was 1 to 2 bars less in the stressed erops, exeept tor CM 1335-4. Midday leat water potential tor both stressed and unstressed erops oscillated between -8 to -15 bars, depending upon measurement dates. These values are higher than those observed tor other field erops under stress, indicating that cassava conserves water. This eharacteristic is of a paramount importance for the crop's tolerance to prolonged drought. Sueh a 'stress-avoidance mechanism" underlies the plant's ability to endure several months of little or no rainfall in seasonally dry and semiarid regions. Coupled with this meehanism is the large reduction in light interception through reduction in leaf area under stress--a factor of great importance in water consumption (Ag. 3.6). Although reduction in leaf area would lead to water conservation, it would also lead to reductions in total biomass and RV (Fig. 3.7). Nevertheless, cassava can reeover rapidly, once released from stress, by forming new leaves, which increases Iight interception and compensates for yield losses during stress (Fig. 3.6). Cassava leaves also remain reasonably active during water stress (Fig. 3.8). The stressed leaves are capable of maintaining photosynthetie rates around 40 to 60% of that of the nonstressed leaves during the enUre 3-mo stress periodo Upon recovery from stress, the old leaves can approach the efficiency of the nonstressed leaves. Furthermore, the new formed leaves of the previously stressed crop showed even higher photosynthetic rates than those of the nonstressed crop (Fig. 3.8). There are apparent genotypie differenees in tolerance to stress as indicated by the magnitude of reduction in photosynthesis. Clone CM 489-1 appears to maintain higher photosynthesis during the entire stress period than other clones. Vield, top and total biomass, HI, HCN content and pereent stareh in storage root5 at final harvest are presented in Table 3.2. Across all varieties, reductions due to water stress were 9% in dry RV, 28% in top growth and 15% in total biomass. On the other hand, water stress increased HI by SOA>. These data suggest that prolonged midseason water deficit does not seriously Iimit cassava productivity and eonfirm that cassava is a highly productive erop where a long dry season oecurs. 50 ® o '" .0.5 • ~ .o,5 ;::;.;",..-,::;4 ~ ...... ·1,0 -LO ~1.5 CM 489-' CM 1335-4 OS 125 155 185 ~t.59S 155 185 • • ~ ~ .0.5 ..... -O" ". -1.0 ·1.0 .. __ STRESS "- CM 91/2·2 ~1.$ CM2:f31-2 !. ·111 95 125 155 ,8$ OS 125 155 18$ :; t. t ti) • • I .o .• .o . .o .• I .o.8 .1.2 .,.., .,.sJ.CM_' , , _t.eiCM 1335-.4 '25 155 18$ .. '25 155 185 O :1 -·-STRESS I -comAOL .o .. , .0.8 ·1.2 >Lar ..... -"" ·t CM 922-2 .1 \¡ CM 2,36-2 '''' ,8$ , 125 """'-- Figure 3.5. Leal water potentlalln water-stressed and well-watered cassava: (A) predawn and (B) mldday. 51 IOOr-------------------, l a' CM 489-1 , § 95 145 195 j ¡ 100,----------------, ~80~ 60L -V-.~ \ '\ ,.'t.. 40 . / • ¡ 20 \/ ' ........ RECOVERY o CM92H , 95 145 195 245 Oé:ys aftar planting Figure 3.6. Light interception in water-stressed and well-watered casS8\1a. -:-;------r=:=-;:-.. -.~ .. CM 489-1 CM 922·2 20 o SIressed • Unstre.ssed - --- Slress period {SO days) 10 CM 1335-4 CM 2136·2 20 - 10 2 6 6 11 Figure 3.7. RY as affected by midsE¡ason water stress. 52 30 CM4a9-1 20,~ i~-l'., 'JI. lO;... ~.~ .1 ' , , , . .R~RY 1 J CU~' • 10 , . .,.-- Newleave$ -- Figure 3.6. Leaf photosynthesls as alfectad by mldseason water stress. Note the higher mtes In the newly formad leaves of previously stressed plants. Genotypic differences in response to water stress also exisí. Clone CM 489-1 had high yields under both wet (19 t dry RY /ha) and stress (18 t dry RY /ha) conditions. This clone also maintained the highest leaf photosynthetic rates during the entire stress perlod (Fig. 3.8), suggesting that high photosynthetic capacity durlng stress could be used as a selection criterion for high yields. Previous findings (CIAT Annual Reports 1988, 1989) have shown significant positive corre/ations between leaf photosynthesis and both yield and total biomass among a wide range of cassava varieties grown under stress in the Patia Valley. cauca, Colombia. 53 Tablo 3.2. Yield, biomus (dry ¡/ha), st.roh IIfld HCN contonl ot final harvast (11 mol a. affeoted by 3 me midsaason water str ••• oommenoing 90-100 OAP; avg. 011987-89 seasons. Unstraosad ¡¡t,.osad HCN HCN ppm ppm Varioty Roo'" Tops Total HI %Starel! (dry M.) Roo'" Topa Total HI % Slarch (dry wI.1 CM 489-1 19.1 7.2 26.3 0.73 77 214 18.0 7.1 26.1 0.72 80 401 CM 922-2 14.8 7.6 22.4 0.66 83 142 15.0 5.9 20.9 0.72 81 190 CM 1335-4 18.1 7.8 25.9 0.70 83 107 16.5 5.1 21.6 0.78 82 123 (JI CM 2136-2 19.3 12.4 31.7 0.61 87 166 15.5 7.3 22.8 0.66 81 338 .¡:. Avg 17.8 8.8 26.6 0.66 82.5 157 18.2 6.4 22.6 0.72 81 263 % chango duo \ti str ... 41 -28 -15 +6 -2 +66 t.SOS% Varo (V) 1.25 1.51 2.44 0.024 3.1 59 Traolmenl (T) 0.88 1.07 1.73 0.017 42 VXT 1.76 2.1 3.44 0.034 84 The highest reduction in RY among this group of genotypes was 20% for CM 2136-2; whereas CM 922-2 showed no change in yield due to stress. Under stress, HCN content remained at low levels (123-190 ppm on dry basis) in two clones (CM 1335-4 and CM 922-2), but was much higher (ca. 340-400 ppm on dry basis) in two others (CM 2136-2 and CM 489-1). Maintaining low HCN under stress is of a paramount importance when fresh cassava is used for human consumption in drought-prone regions su eh as NE Srazil and Sub-Saharan Africa. 3.3 Potential Photosynthesis of Cassava and its Relation to Productivity 3.3.1 Effect of growth environment Environmental conditions under which leaves are developed are of paramount importance in determining potential photosynthesis of cassava. Studies were conducted to evaluate potential photosynthesis in a few cassava varieties from contrasting habitats. Three Colonibian var.--M Col 22 (collected in Córdoba, elevation 100 m, mean annual temp, 28°C; habitat: hot, dry), M Col 1684 (collected in Amazonas, elevation <150 m, mean annual temp 28°C; habitat: hot-humid), and M Col 1522 (collected in Cauca, elevation 1500 m, mean annual temp 20°C; habitat: cool-humid)--were used as test materials. Stem stakes (0.2 m) were planted in plastic pots, (0.33 m wide x 0.46 m deep), filled with a mixture of 40% (w /w) top soil, 33% compost and 27% sand. The pots, which were adequately fertilized with 15:15:15 NPK at the rate of 15 g/pot 15, 45 and 65 DAP, were kept well watered and left in the open at a high-altitude site near Cali, Colombia (elevation 2000 m, mean annual temp, 1TC). Two months after planting, the pots were brought to CIAT Ha in Palmira (elevation 965 m, mean annual temp 24°C), where gas exchange measurements were conducted on the same day. Prior to measurement, sorne leaves were removed at every other noda to balance source (Ieavas) with the confinad and limitad sin k (roots). Then the plants ware left outdoors for 7 days at this site to acclimata to tha higher temp. Gas exchange was measurad again on leavas having the sama aga as those praviously measurad. Four weeks latar the fully expanded and nawly daveloped leavas at the warmer site were usad for further measurements. Photosynthesis was substantially raduced in laaves of all varieties developed in cool climate (Fig. 3.9). This reduction was most pronounced in M Col 22, particularly at low leaf temp (Fig. 3.9A). The leaves that developed in cool climate and were then acclimated in warm climate for 7 days partially recovered their photosynthetic capacities, with the highest recovery (compared to rates before acclimation) in varo M Col 22; however, the ratas were much lower than for the newly developed leaves. The max. rates in newly developed leaves were highest in M Col 22 and lowest in M Col 1522. Another apparent difference was the notable upward shift in optimum temp in M Col 1522 from 25°C in leaves developed in cool climate to 35-40°C in acclimated and newly developed leaves (Fig. 3.9C). Varieties from hot climates had broad and stable optimum temp ranges between 30° to 4O"C in all sets of leaves (Fig. 3.9A,S). 55 The depression in photosynthesis at high leaf temp in leaves of M Col 1522 developed in cool climate was not due to stomatal closure as both leaf conductance and internal COz were increased at that temp range (Fig. 3.9D). The higher rates and the upward shift in optímum temp in the acclimated leaves might indicate the influence of biochemical factors. After 7 days of acclimation in warm clímate, changes in biochemical components of photosynthesis are more likely than changes in stomatal characteristics; moreover, the rates in cool-climate leaves of all varietíes were much lower at all light levels and had a lower saturation light than both the acclimated and newly developed leaves (Fig. 3.10, far var. M Col 1684). The differences in light saturation rates among thesa sets 01 leavas can be attributed mainly to differancas in CO. fixation capacity. The response of photosynthesis to intercetlular CO. in warm-climate grown plants of varo M Col 1684 is shown in Figura 3.11. tt is evidant that at the lowest CO. concentration usad (SO ¡.ILL" COJ, the rates of COz uptake wara positiva. The CO. compensation point, as estimated from the regression line, was about 20 ¡.ILL" CO. under the conditions of these experiments. This relativaly low CO. compensation point indicetes that cassava has low photorespiration. 3!i A • • 30 • 30 • '" • '" 20 20 ,. o o 15 ,. o .~ • '. '. § • • 1 ,. 20 "',. lO 20 .. 30 "" '" .. .1. L. ,. e o • { • ~ • 0-,. (1) ..... - ¡; . .. ~-;;;..; .. • l 310 -; : I 300'" -V 290§ • ~ ;/ . • . '/ :1 15 20 25 30 .. .~' .. ,j ... 15 20 25 ,. .. .. .. 50 !.taf ~AWtt (·el FIgure 3.9. Responses of leal photosynlhesls to leal temp as influencad by growth-temp: (A) M Col 22 (habilat: hot, dry); (e) M Col 1684 (habita!: hot-humid); (e) M Col 1522 (habita!: ¡;ooI-humid); (.) leaves developed at 2000 ro, da.y mean temp, 17"C; (o) acclimatad leal/es (7 days) at ClAT Ha, da.y mean temp. 24OC; ('1 leaves developed al CIAT Ha; (O) M Col 1522; (a) neI photosynthesis; (o) leal conductance to water vapor; (.) Intracellular CO .. Note the apparent upward shift in optimum temp in aecJlmatad leaves of varo M Col 1522. 56 35 30 l e 25 § 1 20 j e 15 1 10 ! 300 600 900 1200 1500 1800 -lO _flux density (pmoI m'21$) Figure 3.10. Responses of leaf photosynthesis (var. M Col 1684) lo Ilght as affected by growth-temp. SymboIs as in Figure 3.9. 20 1.. O.22x-4 12.0.92 .,: 15 8 • i 10 ~ .!! :1 J::: 1 5 ~ ·5 Figure 3.11. Response 01 leal photosynthesis to Intercellular CO.: plants (var. M Col 1684) grown outdoors al ClAT Ha. Measurements made w~I:lS leaves al leal lemp 01 28 ± O.SoC and pholon lIux density of 1500 11 mol m"/ s. 57 3.3.2 Photosynthesls o11leld-grown cassava Photosynthetie potential in relation to productivity was evaluated in 15 varo and breeding materíals from CIAT, grown under rainfed eonditions at the Santander de Quílíehao Experiment Station, Cauca, Colombia (alt. 990 m; lat. 3"30'; long. 76"31'W; mean annual temp 24°C) during the 1990-91 growing season. Measurements of leat gas exchange (CO. uptake and H,O loss) were made with a portable infrared gas analyzer (LCA-2) on several occasions from 23 July to 25 November 1990. Across all reps, 30 fully expanded upper eanopy leaves were .measured per varo AlI measurements were made from 08:00 to 11:00 h with a solar irradianes of 1200-2000 IImol m·2/s. Normal air (325 :1: 10 j.lLL CO,) was drawn from above eanopy usíng a vertieally mounted 4-m glass tiber probe conneeted to a pump. A small leaf chamber connected to the infrared gas analyzer was clamped over the middle portíon (6.25 x 10-4 m2 surtaes area) of the eentrallobe of the measured leaves and held toward the sun for 30-60 sec to obtain steady-state gas exchange. In tield-grown cassava, the photosynthetíc rates of upper eanopy leaves measured in wet soils were not light saturated up to 1800 11 mol m2/s (Ag. 3.12); moreover, the max. rates varied from 39 to 50 j.lmol CO. m-2/s (Table 3.3). These rates are higher than the max. rates observed in well-watered potted cassava grown outdoors under similar levels of solar radiation, photoperiod and ambient temp. The calculated ratio of the íntercellular to ambient CO, concentrations (CijCa) varied between 0.37 to 0.45 (Table 3.3). These values are similar to those in C. species ('" 0.4) but are lower than those in Ca speeies ( .. 0.7). These data íIIustrate that the photosynthetie potential of cassava is high but can be fully expressed only in near optimum environment. The seasonal avg net photosynthesis in upper eanopy leaves ranged from 26 to 36 11 mol CO. m'2/s (Table 3.3). These rates are relatively hígh, taking into consideration that most measurements were conducted during the dry season when water defidt was largely coupled with hot, dry air (Table 3.4).The known closure of stomata in reaetion to soil water stress or to low atmospheric humidity can lead to high leaf water status (CIAT Annual Reports, 1982-90). Maíntaining leaf water status near normallevels reflects positively on leaf photosynthesis. Studies at CIAT have shown that leaf photosynthesis under stress remained ca. 50 to 60% of the well-watered crop even with an extended stress of more than 3 mo (see also CIAT Annual Reports, 1988-90). Biomass and RVs were relatively high in this group of varieties (Table 3.3). Avg dry RV and biomass were 17 and 23 tjha, resp. This productivity under rainted conditions with prolonged water deficit commencing early in the growth cyele (Table 3.4) ¡lIustrates the capaeity of cassava to withstand drought. Under stress, cassava restricts its leaf canopy and extracts soil water slowly (see Seetíon 3.1). 58 Figure 3.12. Responses of leaf pholosynthesis lo Ughl In field-grown cassava. Meesurements made on upper canopy. fuUy expanded leaves during the ralny season. (A) CG 927-12; (B) eG 996-6; (e) CM 507-37; (O) CM 523-7; (E) CM 3456-3. These mechanisms allow the crop to maximiza WUE over a longar period whila maintaining good yield. MaJ<. photosynthesis was correlated with biomass and storage roots/pl (Table 3.5). The highly significant correlatíon with root no. suggests that expression of potential photosynthesis is partially dependent on sínk capacíty (CIAT Annual Report 1990). which might explain the much lower rates always observed in potted cassava. It must be notad that these high rates were obtained in 5- to 6-mo-old cassava when root bulking was in a linear phase and demand for assimílates was high. Moreover, at thís stage of growth, LAI was near maJ<. (see sectíon 3.1), and shoot demand far assimilates was mínimal. 59 T ab/e 3.3. leaf nel photoaynlheGia. yield and blomus of fleld-grown """ .... a. Male. Ne! SeasonaI I'holOllynlhesls el/Ca" Av; Net f'IloIosynIhesI. Total DIy $torageRoolllI Vatlely (n'6) (n-6) (n-30) DlyRY Bíom ... PI l' mol COa m",a ~ mol COa m"/a l/ha CG~ 49.7 0.37 33.8 18.00 23.64 14 M Bta 191 47.4 o.a7 35.S 16.50 24.00 14 CM 4664-1 45.1 0.39 34.0 19.22 25.55 13 CM 4145-4 43.9 0.40 31.7 17.42 22.44 7 CM 3456-3 43.7 0.43 31.9 17.97 23.12 11 CM 507·37 43.7 0.38 26.7 19.02 24.30 11 CM 4716-1 43.6 0.42 31.8 16.23 25.30 10 8 M Col 1684 43.0 0.42 30.9 11.11 22.20 7 CM457~1 42.8 0.39 33.2 17.95 21.90 7 CM 4617·1 42.8 0.46 31.4 17.20 22.87 8 CM 523-7 42.3 0.45 30.1 18.06 23.78 11 M Col 1468 42.3 0.44 30.3 14.63 21.67 9 CM 4701·1 42.2 0.45 30.9 18.08 24.00 9 CM 4111-2 41.3 0.45 30.9 17.21 24.61 8 CG927·12 39.3 0.43 26.2 12.15 15.74 8 Mean 01 all VIII. 43.5 0.42 31.4 17.12 23.00 lO lS05'!1. 1.7 0.08 1.8 1.74 2.38 2 " Cl/Ca = Inte!!?!!l!yl!!r CC!a: caloulaled from max. !!!af gas exchanga maasuremenlS. Amblenl co,. Tabl. 3.4. MeteorologJcaJ data al Santander do Ouilichao durlng !he 1990-91 ...... n. T em&:>eralu ... Aefallvo HumidJty Mol 601..,- Yr Aalnfall Evaporallon Aadla1lon Max. Mln. Mean Mean Mln --mm-- MJm'" "C --'1(,-- Mav/l990 134 109 465 28.7 18.8 23.6 80 54 Juno 36 120 515 29.8 18.5 24.3 . 75 47 July f17 131 50S 29.8 17.5 24.0 71 41 al AuguSI 55 149 li67 31.5 17.2 25.2 64 39 ..... Saplomber ¡g 151 618 31.4 18.2 24.4 74 36 Octobor 245 118 518 28.6 18.7 23.2 77 49 November lf17 127 555 29.2 18.7 24.0 78 50 Oeoembor 200 121 495 29.4 18.7 23.6 77 50 Januery/l991 33 144 585 30.9 19.0 24.7 70 39 February 72 134 648 31.3 19.4 25.1 70 40 Seasonal average photosynthetic rates were also correlated with yield, biomass and storaga roots\pl (Table 3.5). These findings again confirrn pravious reports (CIAT Annual Reports 1989, 1990) showing a direct ralationship between photosynthesis as measured in the field and productivity and suggasting that laaf photosynthasls, leaf area duration and storage roots/pl are useful eritería for parental selections. The two components of leaf gas exchange--that Is, stomatal and mesophyll conductances--were also correlated with yield and biomass. The negative correlations with intercellular COl (Table 3.5) suggest that nonstomatal factors were more important, especially as correlations with leaf WUE (CO. uptake/H.O loss) were positive. These data indicate that cassava has high photosynthetic potential, which Is assoclated with productivity when grown under natural and near optimum conditions. Cassava requires hlgh amblent temp and hlgh solar rediatían far optimallaaf development and for the expression 01 its photosynthetic potential; thus, when grown in environments--natural or artificial--that daviate from these fundamental climatic requirements, its photosynthetic potential would not be fully expressed. Compared to the efficient C. crops such as malze and sorghum and to warm-climate C. crops such as cotton and sunflower, cassava is hlghly afficient, considering its thin, hypostomatous leavas. Furthermore, cassava has low photorespiration and elaveted activities of the key C. enzyme, phosphoenolpyruvate carboxylase (Table 3.6), but lacks Kranz-anatomy, typical of C. species (CIAT Annual Report 1986). These physiological and biochemical attributes can be advantageous to cassava photosynthesis, particularly when the crop has to endure a long periad 01 drought coupled with dry, hot air in the regions where' it is grdwn. Moreover, these findings have practica! implications for the strategies adopted in cassava breading programs far different praduction ecosystems: • The high sensitivity of photosynthasis te amblent temp points to the need for genotypes more tolerant to low temp, which could be used in the highland tropics and in the subtropics where cassava is an important staple far resource-poor populations. • The positive association of photosynthesis with productivity suggests that selection for high photosynthesis in parental materíals may lead to higher yleld when combinad with other yield determinants. • The relatively hlgh photosynthasis under prolonged water deflcit indicates the potential of cassava as a toad erop in dry regions such as NE Brazil and Sub-Saharan Africa. In thase drought-prone areas, very few other foad erops would survive and produce reasonable yields. Breeding and selection far drought tolerance In thls ecosystem are already In prograss (see Chapo 2). 62 El T able 3,5, Corre!ationa among _a leal gas exchange charllCltlristlcs and AY. _ss and .torege roo! no, 01 fieldijIOWIl cuuva, DlyAY Total - l!! o j 'O <'l (,) ., c;¡ 'O ::t « Visual .t:: == ... .. ~ cfi )H( 11 Good Ouallty 11 Poor Ouality "a: - ~ ! .. ! cfi .. @ lO .. ~ !I (,) ~ )0< )0< Olfaclory Taste ~ .8 u: CASSAVA OUAUTY FACTOR S ~ lt .. ~ ! ! i :x: i:i'i .. 1 (,) )1(> Texture Alter- tasta Figure 4.3. Prolile of goOO- and poor-quality fresh cassa:va accordlng to results of expert taste panel. 73 Multivariate analysis of a no. 01 samples from different varieties helped identify which characteristics 01 the fresh root ware most relatad to like/dislike of the sama samples (Table 4.3): cansava taste and hard texture. The importance of a "cassava taste· factor in acceptability of fresh cassava has not praviously been recognized and merits further research to identify the volatile f1avor components ¡nvolved. Hard texture had been reported as a crucial characteristic affecting frash root quality in previous annual reports, both as a result of 19b studies and consumer surveys. Specifically. the occurrence of a hard. glassy texture in boiled roots as a universally perceived negative quality factor has been evidant for sorne time. 4.4.3 Characterizatlon of germplasm Numerous experiments have demonstrated that considerable variation for many of the important quality characteristics exists within cassava germplasm. Table 4.4 shows the range of values encountered for the principal chemical constituents of the fresh root Table 4,3. Mullivarlate regre .. 'on analyols of .. part t •• to panel r .... lIs: Delermlnalion of quality characleristios most signiflC3ntly asllOCialed wllh IIk. or distika of samplas from ti ca""""a V$l. auality Leve! 01 Slaniflcanca I:l!!I of Eacl1 Q!¡aracler Charactoristic HMCl CM6812 CM 1559-5 CM 489-1 CMC40 M Col 22 Color 55 4' 68 38 7' ao Starch 83 54 21 3' O' 17 Glasslness 16 13 l' O' O" 65 Moistur. 65 68 37 24 68 39 Freshneu 76 29 S' O' lO' 4' Smell 68 O' 7' O' 96 75 Delerioration amelJ 43 47 26 49 64 65 Swee! tasto 20 59 4' g' 7' 14 Bitlllr tasto lO' 64 93 11 60 53 Csssava tas'" 3' O' '1: O' 13 3' Fibrousness 78 53 91 64 79 7' Hatdness O' 7' O' 2' 5' 9' Consistency 68 94 51 72 O' 42 Dryness 48 92 27 ¡;¡ S' 69 Bitler aftertasto 38 87 lO' 30 31 64 Note: Level 01 slgniflosnco ('1(» takon from Ihe probabllity 01 !ha T -valu. Ior each cheraotorlstic, Valuas below lO lO¡ are significantly relaled lo likejdisllke _uaticn, 74 Tablo 4.4. Rango oncount.red In !he principal constituonto 01 cassava parenchyma and pool. percant lOry WI! Constituent Parenchyma Pool DM ('lI. lresh wt) 23-44 15-34 Storeh 70-91 -Total sugars 1~.3 5.2-7.1 Crudo flber 3.~.0 5.()'15.0 Ash 1.0-2.5 2.8-4.2 Proteln 1.0-6.0 7.0-14.0 Fat 0.3-1.5 1.5-2.8 Total eN (ppm) 30-1350 60-550 Knowledge of the vanability of starch charactenstics in cassava is especially weak, with only a few reports available in the literature. The Program intends to characterize the ccre germplasm collection germplasm for CN content and starch quality factors. 4.4.4 Characterlzatlon 01 wlld Manihot specles The few (9) species available thus far in the wild Manihot collection at CIAT have been characterized for total CN content in leaves and roots, and fer the presence of physiological deterioration-characteristics where it is known that the cassava germplasm evaluated to date does not hold sufficient variability to permit selection fer desired characteristics. AlI species had high levels of CN in leaves and roots (Table 4.5), and all showed susceptibility to physiological deterioration. As over 100 species exist in the Manihot genus, it may still be possible to find an acyanogenic species or one without physiological deterioration. 4.4.5 Effeet of plant age and ellmate on quallty The perennial cassava plant has no fixed maturity time; harvests are therefore fixed by a number of factors; for example, the farmer's need for land, pnce in the market, yield and root quality. Theplant responds dynamically to changes in the environment, especially as regards starch deposition and mobilization. These factors result in extremely variable root quality with plant age, and by season and location. Figure 4-4 shows the variabmty in reot DM and starch content and eating quality over a 4-mo penod for a variety harvested weekly at CIAT. This variability is typical of the crop, making it difficult to decide on the optimum harvest time for agronomic expenments. The effects on eating quality can be seen in one experiment in which the same variety was harvested on 7 different occasions over a 2-yr penad at CIAT, always at the same plant age (10 mol. 75 Tablo 4.5. OM and eN contents of 9 species of Manihot. Total eN (ppm, CM basi.) Manihot Species %CM Parenchyma Poel M. aesculifolla 40.2 582 2818 M. carthaglnensis 19.7 1443 2811 M. crassisepala 43.8 314 912 M. grahami 48.3 260 1564 M. michael/s 32.9 442 6106 M. p.oudog/aziovil 38.1 1082 6006 M. nabel/ifolla 42.2 733 1564 M. rubrica"". M. chlolOsticta 2636 Note: Mlssing data due to absence of storage roots in the planta sampled. Starch Ealing Quality % DM m 12 -r-..-.--,......,r--'-,.~--r~.---..-.--,......,,---.""""6~--r--.--, 40 Poor as 10 al 8 75 6 7D 4 65 2 Good 60 O O _._ .. + ...... ---&-- 2 4 Ealing quality DM canten! (%) 6 Starch cantent (%, DM basis) 39 38 \;./~, . , " '. \ 36 35 Lea' 636 1519 2123 3537 2943 3196 1622 3064 2969 Figure 4.4. Varlabllity In root DM and starch contents and eatlng quality (rapL ¡ evaluatlon method) of HMC-1, harvested at weekly Intervals from 6 to 11 mo at eIAT. 76 Ealing quality, as evaluated by Ihe expert panel, varied significantly in the 3 most important quality characteristics: hardness, glassiness and cassava taste (fable 4.6). 4.4.6 Crop management effects on root quallty The foregoing sectíon showed how variable Ihe quality of cassava can be, depending on plant age and certaín climatic factars. Experiments with ather preharvest variables--such as fertilizar applications. water stress regimes and stake selection--have shown significant effects on RY and are leading to the development of crap management strategies for cassava. However there has been relatively little infarmatian available an the effects of these variables on root quality characteristics. Advantage has therefore been taken of experiments conducted by other sections af the Cassava Pragram during the last four years. Collaboration has been especially fruitful wíth the Physiology Seetíon, with samples being taken from experiments in Santander and Pivijay, Colombia. Samples from 25 field experiments have been analyzed for root chemical composition and eating quality. These show that the effects of different treatments on root quality can be very significant. Tablo 4.6. Taste panol ""a1uallon 01 hardn ..... glasoin ••• and ea ...... laste lo, bolled _a sample variefl ••• grown al ClAT and harve_ al 10 mo of aso on 7 oecuions betwoen 1987 and 1989. Quality Variety Charaeterístíc Harvest HMCl CM 681-2 CM 1559-S CM 489-1 Haróness 39" 53" SS"" 57· 2' 77' 92- 47" 94" 3 60"" 92" 68" SO" 4 71" 102" 69"" 53· 5 52'" SS" 73" 91\" 6 108 • 74 • 70"" 68" 7 36' 46" 50""" 57" Glassiness 1 49 " 75' 56" 51"" 2 50" 109 • 57" 75" 3 33"" 109 • 71" 7S" 4 38'" 130 • 49 '" 44" S 26" 69" 33" 63"" 6 76" SO' 7S" 69' 7 44" 32" 36" 26" Cassava taste 1 65" 76" 71 ' 77"" 2 77" 53" 64<0 72 e 3 109" 66' 72 ' 76"' 4 69 00 49" 92"" 57" 5 115· SS" 98" 65' 6 49' 67 ' 71 • 7S" 7 99"' 97 " 102" 102 " Not .. : 1. V.lu •• wllh diff&rent ¡ott.r superscriplS we,. slgniflC8Jllly different (P • 0.905. OMRT); tilo m.ln offect 01 herves! dato was si-gnificant al P = 0.001 in aH cases. 2. EvaJuation sea.le: ~150 with O = ab$ence of character; 150 = intensa expresslon 01 charaeter. 77 4.4.6.1 son fertílity. Most results have been obtained from experiments in the area of soU fertility improvement. Comparisons of fertilized and unfertilized plats are not consistent; experiments at Pivijay and Patia showed no effeet of fertilization (NPK, 15-15-15) on DM, starch, total CN or eating quality. A similar experiment at Santander, however, did show a significant effeet, with fertilizer treatment increasing DM content of the roots and improving their eating quality. Response to fertílization is probably related to the nutríent status of the soil involved as another experiment showed that root quality (DM and starch contents) were significantly higher on plots after fallow than on plots with several successive cassava crops. . Similarly, annual applications 01 fertilizer resulted in better root quality (higher DM, lower HCN) than in plots reliant on the residual effects of previous fertilizer applications in two experiments in Santander. Nevertheless, a further experiment in Santander produced roots with higher DM but also higher total CN after annual fertilizar applications. Differences in the affeets of NPK root quality have also baen found. K seems to be especially important in determining root quality. In three saparata NPK trials at Santander, the zaro-K treatment producad roots with significantly less DM and with highar total CN contents. At Pivijay, two NPK trials showad no differencas batwaan diffarant treatrnent levels although a separate K experiment did give a significantly higher total CN content with zero K as at Santander (Table 4.7). Ongoing experiments in collaboratíon with the Physiology Seetíon should provide more information on the role of K in determining root quality in cassava. In contrast, the role of P appears to be negative: Two varietal screening trials at Santander have both demonstrated that higher levels of P incraase root total CN content significantly, while DM content and eating quality declined. Navarthaless, there are varietal differences: Although most varieties increase in total CN, a faw remain constant and sorne decrease in content (Table 4.8). Table 4.7. Effecl 01 K applica\ion on roo! DM and total CN Cóntent (mean 01 8 olones haNested al Pivljay). Total CN Treatment OM ('l<.) (ppm, DM basls) level 01 K ~/ha) O 34,5 SSO' 50 34.2 2B6" 100 35.0 277" 200 35.3 2B6" Signlficano. NS 0.01% Note: OIlferent lattar superscripla "'pres.nl significan! difieren ... in values, OMRT, 78 Tabl.4.8. Effect of P applieation on rool OM and lolal CN conten! 0133 clones harvested al Santander. P Applieallon (t/tla) O 75 O 75 Total CN Contenl Casaava ClonejVar. ""OM (ppm OllA basis) 00927-12 36.1· 33.8 187 206 00996-6 38.S' 34.3 131 183" CG 1374-2 40.0 39.3 128 154 CM 305-41 32.0 31.8 288 411" CM 2718-1 35.9" 35.1 288 360" CM 2774-11 37.3 36.7 96 119 CM 3285-7 37.S' 35.3 139 192' CM 3401-2 37.7' 35.3 172 208' CM 3456-3 36.4" 34.7 158 219" CM 3654-3 40.1" 37.7 138 154 CM 3667-1 37.1 37.2 251" 220 CM 3750-5 37.6 37.6 278" 249 CM 4145-4 37.1 36.9 124 177 CM 4575-1 37.1 36.9 234 285" CM 4617-1 35.8 35.5 424 477' CM 4701.1 38.5 38.2 102 132' CM 4711·2 37.6 37.2 41S' 262 CM 4711H 40.2 39.7 123 154" CM 4793-1 38.7· 38.7 257 354" CM 4630-3 35.0' 31.0 1114 343" CM 4664-1 38.3" 35.3 152 189" M Sta 191 35.6 35.9 166 227" M Sta 383 37.4 37.2 350' 305 M Col 2215 38.S" 34.9 291 382" M Pan 51 36.6" 33.7 146 293" SI> 104-264 37.0' 35.7 193 260" SI> 104-284 35.3 34.7 125 253" SI> 106-59 33.6 35.S" 166 150 SI> 250-3 37.5* 33.3 146 181· SG 302·1 36.1· 35.2 232 270" SO 458-1 39.S" 38.2 166 214" SG 545-7 33.9 33.S 207 198 SM 414-1 33.7 37.0' 132 204" LSO· o.s LSD·29 Nole, • Denotas slgnifiean! (5%) differen ... ""twHn mean •• placed noxl 10 Ihe highest valua 01 !he pairo Finally. two experiments at Pivijay have looked at the effect of increasing N leve!s on root quality. In both experiments OM and starch contents increased but total eN content decreased as N levels increased from o to 200 t/ha. Eating quality also improved. Future experiments will continue the attempt to clarify the roles of the different soU fertility factors on root quality. with more emphasis on identifying germplasm that provides good quality under poor conditions and also responds to fertilizer applications and other management strategies. 79 4.4.6.2 Effects of cover Q{ops. One recent experiment carried out at Pivijay has demonstrated the potential for improving root quality using a mulch cover. The use of mulch resulted in roots with a significantly lower total eN content while DM content remained unchanged. These results (fable 4.9) could be due to the lower soil temp under the mulch or to the extra nutrients provided by this treatment. Starch samples were also analyzed for their amylose content, using the new routine methodology; and the results showed the significant effect that the use of mulch can have on this important quality factor as well. Roots from plants treated with mulch had lower amylose percentages than the control plants (fable 4.9). This is the first instance of a significant effect of the preharvest environment on starch quality in cassava. 4.4.6.3 ¡;ffect 01 water stress. Three experiments have been carried out in which plants have been subjected to water stress during the growth cycle, for periods of up to 4 mo. Water stress significantly increased the total eN content of roots although differences in verietal response to the stress were found (fable 4.10). Root DM and starch contents were significantly reduced by water stress in some varieties, but not in others. The potential for developing verieties in which quality is not adversely affected by drought stress has been identified. This could be important for Africa, where drought stress has frequently been cited as a factor contributing to eN toxicity problems. Table 4.9. ElIec! al mulch apptleation on total CN and on star.h qualtty al roots 01 M 0011505 harve_ at PIvIJay, Colombia, Total CN Amylooe CM (ppm. CM) (% Starch) Mulch apptloation 33.6 139 16.7 Control 32.7 208 20.0 Significan! dilferenee al P< NS 0.001 0.001 Table 4.10, Effect 01 a 3-mo perlad al water StIé ... foIloWéd by 2 mo ot """,peratlon, on total CN and ataren oontants al 4 eassava varo CM 1335-4 CM 922·2 CM 2136-2 M Col 1684 so SO 87 88 Starch (%. OM buls) 78 77 80* 76* Note: * Denotes significant dífferenee between means a1 P<.O.Q5. 80 Total CN (ppm. OM bula) 123 148 162 723 108 155 340* 4.4.6.4 Other preharvest factors. This series of experiments has also shown that root quality was unaffected by association of cassava with maize or maize and yams. In a crop rotation experiment, the only treatment that resulted in poorer quality (Iower CM content) was planting cassava after two crops of maize. Planting cassava atter beans resulted in improved DM contento Two experiments comparing root quality in plants derived from selactad and traated stakas, comparad with unselectad and untreated stakes, were carried out at CIAT. Stake treatment had no effect on root quality; selection resulted in sígnificantly lower total CN content in one experiment. In one experíment where plants of 2 varo were subjected to differíng levels of mealybug infestation, root quality was severely affected. Root CM and starch contents and eating quality decreased as the intensity 01 the mealybug attack increased; there was no consistent effect on total CN contento 4.4.7 Relatlons between eatlng quallty and starch quallty In Section 4.4.2 the importance 01 the glassy texture problem was highlighted. Roots that become hard and glassy when boiled, rather than soft and floury jn texture, are negatively evaluated by consumers in all urban markets surveyed in Latin America. The impossibility of determining whether a root will be glassy or not befare boiling, makes purchasing cassava risky from the consumer's viewpoínt, even if the deterioranon problem has been solved. The body of field observations and experimental results also point lo the extreme variability 01 glassy texture phenomenon: Eating quality may vary greatly among plants of the same variety in one field or even among the roots of one plant. To resolve this problem, greater quality stability over time and over edaphociimatic variables is required. Research in this area is hindered by the lack of a simple, obJective evaluation method tar glassy texture: AlI root samples must be boiled and tested. Over the Iast four years, a collaborative project between the NRI and CIAT sought to identify the causes al this texture problem; i.e., to relate root texture to a chemical or physical characteristic al the root tissues. No simple relationship was found between eating quality and DM content (Table 4.11). It was hypothesized that changes in the starch component of the parenchyma (80-90% of total CM) was responsible for the changes in boiled root texture. To this end, 4 varo were selected for study, one (HMC 1) of goOO eating quality and 3 of variable quality (CM 681-2, CM 1559-5 and CM 489-1). In order lo determine variability in eating quality due to seasonal fectors, each clone was harvested on 7 separate occasions over a 2-yr timespen, always at 10 mo of age. AlI samples were evaluated for eating quality by the expert taste panel at CIAT; extensive variation was tound (Table 4.6). It Is noteworthy that even the variety selected as being of consistently good eating quality was in fact highly variable. AII samples were also subjected to a complete proximal analysis, and starch samples were exhaustively characterized by the NRI, together with Nottingham University. The results of the starch characterization work were surprising in that there was very little variation in any of the starch quality variables examined. There 81 Table 4.11. Cooklng time and •• ting quallty of cassava raot. _ dlffe'enl 01.1 """tenIS. % Aoot Cooldng Eating 0uali1v Clone 01.1 TIme (min) TUIe Texture M Pe, 245 26.1 20 good soft M Col 1522 25,4 36 poor hard CM 91-3 30.4 15 good 110ft CM84lJ.l 30.1 31 peor hard HCMl 35.8 20 good soft CM 922-2 35.1 40 poor hard was no relation between eating qualíty--specifically root texture--and any of the starch characteristics. Although the study did not therefore aclvance the specific objectives for which it was designed, it did provide an opportunity to carry out a comprehensive characterization of the properties 01 cassava starch. However, the problem of identifying !he causes of glassy texture in cassava remains unsolved. Further studies are required, concentrating on other root components that could affect texture (e.g., cel! wall pectins and hemicelluloses). lha lack of undarstanding of tha biochemical aspects of cassava eating quality is now a severa constraint to improving this quality !hrough breeding: currently only advanced lines can be screened for eating quality. 4.5 Future Prlorltles • Complete !he standardization of methods for starch quality evaluation, including Brabender viscoamylograph • Develop a rapid and raliable method for CN determination for use by breeders • Characterize!he core cassava germplasm collection and the available wild species of Manihot for all salient quality factors • Continue research on the physicochemical causes of variation in eating quality of fresh cassava • Continue research on !he effects of preharvest variables on root quality, concentrating on water stress and soil fertility, and on CN and starch quality 82 5. BIOTECHNOLOGY Early work on cassava biotechnology focused on tissue culture techniques for virus elimination, germplasm exchange and conservation. Following the organization of the Biotechnology Rasearch Unit (BRU) at CIAT in 1985, research was undertaken on genetic characterization, cryopreservation and plant regeneration, and later on ganetic mapping and transformation. BRU cooperation with the Cassava Program focuses on integrating biotechnology into the Program's o09oing research activities. Table 5.1 shows the research activities in cassava biotechnology at CIAT, the type of project, and the current state of technology development. Developments In cassava biotechnology over the last tour years will be highlighted, with emphasis on three interrelated areas of application: ~ Analysis and conservation of genetic diversity ~ Crop productivity research .. Institution-building activities Table 5.1, Ces8ava bi_nology lO""",eh _es al CIAT (1987·91). Type of Project Stalus 01 Tech, Aesearch ktlvltiu & Technologles Cera Speeial Oevelopmenl l. &¡aIvsis & !2!!nylV8tion !lt II!!n!!!isl dlve,!!x ,., l80zyme ftngerprfnting X X In use' , ,2 ONA ftngerprinting X Aavanoed '.3 In víIro activo gene bank X X In use1 1.4 Cryopreservation X Advanoed 2. Croo productiyitv research 2.1 Palllogen.tree dones X In "",,' 2.2 Immalure pollen culture X Inillating 2.3 Genetio ttansformation X Inillating 2.4 Molecular mapplng X lnitlatlng 2.5 _ quallly and co. assimllation X X Inillating 3. Institutlon building 3,1 Tralnlng/biotechnology Issu •• X X Underway 3,2 The cassava blotechnology network X Underway , Technologl •• shlfted f,om lIl. SRU 10 Ihe GRU and !he eassava Prog,em In 1989-90. 83 5.1 Analysis and Conservatlon of Genetlc Dlverslty In Manihot 5.1.1 Isozym~ fingerprinting The fingerprinting of cassava by a:,8-esterase (ESl) isozyme electrophoresis was developed through a collaborative project with the U. of Manitoba, Canada. In 1988, the BRU improved the methodology with the addition 01 4 systems: Diaphorase (OlA), acid phosphatase (AcP), peroxidase (PrX) and glutamate oxaloacetate (GOl). It was demonstrated that EST patterns discriminate morphologically similar varieties (Ag. 5.1 A), with practical applications for sorting mixtures and eliminating duplicates from the cassava collection. Studies on the genetics of isozymes can assist in determining linkages far constructing the cassava genetic map. Sorne 300 progenies from 11 crosses have been analyzed to determine the EST-1 locus, the fastest cationic bands of this isozyme. Results indicate that the ESTe 1 locus comprises 5 multipJe alleles, including one null allele (Aa, A 1, A2, A3, A4). There are 11 observable phenotypes and 5 nonobservable null alleles. This locus bahaves as a monomer and its inheritance pattern is compatible with a diploid model. 5.1.2 DNAflngerprlntlng In contrast to other methods, ONA fingerprinting covers the genome extensively, detects variation in coding and noncoding regions of the genome, and is insensitive to developmental and environmental variations. In a coIlaborative project with IBPGR, the BRU has been developing ONA-based techniques far analyzing the genetic diversity of Manihot. Techniques tor ONA extraction and far constructing a genomic library have been implemented using the varo M Col 22. ONA was digested with each one of the restriction enzymes (Pst 1, Eco Al. Bam HI, Xba I and Hind 111), then ligated into pUC 19 and the constructs transformed into Echerichia col; OH5a; ¡nsert size ranged from 0.2-7 Kb. ONA, ¡solated from varo M Col 22. M Col 1505 and CM 507-35, was digested with the enzymes Apa 1, Bam HI, Ora l. Eco RI. Eco AV, Hae 111, Hind 111, Hinf 1, Hpa 1, Mspl, Pst l. Taq I and Xba 1. It was found that methylation-sensitive enzymes such as Apa I and Pst I cut cassava ONA very ineffectively. Polymorphisms were d~ected with several clones; one Hind 111 clone was able to discriminate among the 3 varo Other probas have also been used far detecting polymorphism in cassava. The human mínísatellite probe (Jeffrey's probe) resultad ineffective in detecting polymorphism; however. the phage M13 probe and several random amplifíed polymorphic ONA markers (RAPO) yielded sufficiently variable band patterns to differentiate the varieties tested (Fig. 5.16 & C). These probes are being tested for morphological and isoenzymatic analysis of similar varieties and clones that have been in in vitro storage for over 10 yr. 84 a b e d e f 9 h j A B e Figure 5.1. Molecular flngerprlntlng of cassava var.: (A) EST isozyme fingerprints of 10 varieties (8i); (8) M13 proba ONA flngerprlnts of 15 varo (a-o); and (e) RAPO flngerprtnts of 16 varo (a-p). 85 Genotyping of cassava by isozyme and ONA fingerprinting techniques will be useful in assessing genetic variability, defining gene pools and contributing to the construction of the cassava molecular map. 5.1.3 The in vitro active gene bank (IVAG) The cassava lV AG is probably the largest and most complete in vitro colleetion tor any crop in the world. In the IVAG. plantlet cultures derived from meristem tip culture are available for micropropagation and distribution at any time. IVAG conditions are: 22- 24°C, 1000 Ix iIIumination and 12-h photoperiod; the subculture penod can be extended to 15-18 mo. Up to 1989 the IVAG comprised over 4300 dones in a 6x7 m room. The estimated operating cost ot the CIAT cassava lVAG Is US$30,ooo/yr, which covers the salaries of five technleians, reagents and other eonsumable items. Given the experience acquired on In vitro conservation at CIAT. a collaborative pilot project with IBPGR was undertaken from 1987-89. The objective of the project was to evaluate critical technlcal and logistical aspects ot establishlng and running an lVAG, using eassava as a model. The Pilot-IVAG project has provided valuable information on prooedures far: .. Entry ot material to storage trom the field, in vitro exchanges, including pathogen tests, no. of samplesjaccession .. Monitoring cultures (viability and stability) and facilities tor in vitro conservation .. Managing data including passport, fingerprinting, pathogen tests. subculturing, and distribution of culturas 5.1.4 Cryopreservatlon of ahoot tipa Cryopreservation providas a means to the long-term storage of cassava dones, seed and pollen. With the collaboration of the IBPGR, which provided support for a postdoctoral fellow and teehnical aid, research was conducted on cassava cryopreservation from 1988- 90. Given the impossibility of reproducing previous cryopraservation reports, it was decided to use seeds as specimens for freezing experiments. In 1989, 90-100% seed germination was obtained after rapid immersion of whole seeds !nto liquid N. Thawing had to be slow to prevent seed shattering. In 1988 very few shoot tips sUrvived cryopreservation, but in 1989 shoot tips consistently survived freezing to -25 to -SS·C. Only toward the end of 1990 was it possible to obtain an avg 01 70% shoot-tip survivai and 20% plant formation frem cryopreserved shoot tips in liquid N (-196°C). 86 The BRU's strategy was to develop first a protocol with var. M Col 22 and then test the system with other cassava genotypes. Every single step 01 the process was studied, paying special attention to conditions that can prevent or minimize growth 01 ice crystals within the tissue, such as composition of the culture medium and the kind and amount of cryoprotectant used prior to freezing. Use of sorbitol, sucrose and dimethyl sulfoxide (DMSO) in the preculture medíum, and the elimination of water from the surface of the explants contributed to increase tissue survival to 90% and shoot formation to 50% (Fig. 5.2A). The use of small shoot tips and purified cryoprotector increased shoot formation. lnterestingly, direct immersion into liquid N gave both a higher recovery rate and higher shoot formation than slow freezing (Table 5.2). The cryopreservation response 01 14 cassava genotypes, with M Col 22 as a check, was tested. Viability and shoot formation varied from 4-80% and 0-50%, resp. Work in 1991 focused on improving the preculture, cryoprotection and after-freezing culture conditions. The BRU has significantly increased the response of low-responding genotypes and obtained shoot formation from nonresponding ones. It seems that genotypic differences in response are not due to the actual freezing step, but to the culture conditions prior and after freezing. Recently, the Unit also succeeded in the cryopreservation of cassava pollen. This concerted effort should lead 10 the development of a cassava base gene bank. 5.2 Crop ProductlvHy Research 5.2.1 Tlssue culture for producing clean plantlng material Significant increases in yield have been reported as a consequence 01 eliminating viruses from cassava varieties. The technique consists in the culture of 0.1-0.2 mm meristem tips from terminal buds grown on stakes at 4O-42°C and 35°C day jnight for 3-4 wk. Indexing by cassava virologists has confirmed the elimination of FSD, CCMV, CsXV and latent viruses. International germplasm exchange has been facilitated by the use of pathogen-tested material. Up to 1989, nearly 2000 clones were introduced to CIAT from 13 countries in the form of in vitro cultures, and CIAT distributed many pathogen-tested clones in vitro to 35 countries. At Guangzhon, China, CIAT clone CM 321-188, selected and named "Nan-Zhi 188," has been propagated massively using in vitro techniques and the plantíng material distributed to 15 sites. Similarly, national agricultural programs in Panama, Venezuela, Mexico, Peru and the Philippines have advanced cassava micropropagation to the field-testing stage. 87 1 2 A 2 3 B Figure 5.2. (A) Cryopreservation of cassava shoot tlps: (1) shoot tlp retrleved from Ilquid N (-196'C), (2) growth afler 1 wk in post-freezlng medium. and (3) plantlets from trozan shoot tips after 1 mo; and (S) toward a haplold method In cassava: (1) lsoIated tetrad-stage microspores (arrow). (2) mitosis In tetrad microsporas (arrow), and (3) mlcrocallus from microspore. 88 Table 5.2. Effect 01 shoo!·tip si ... , ctyOprOIeetlon and free'ing ,al8 en viabillty and shoo! Io,mation alte, tre.,lng In Ilquld N (var. M Col 22). Treatment Sheet tip al, .. ,man (0.5-1 mm) la'ge (2-3 mm) CIyoproteetlon _le DMSO (ampul .. )' standard DMSd' wiIhou! cryoptoteetlon' cryoSHds Freezing tate" fut $Iow 1 Avg$ wi!h !he sarna letter _ no! slgnlfican1ly dlllerlln1 al 0.05 _. 2 Shoot tipa ""'" drMod prior lo freezlng, o Fas! _ dil'llct Immerslon In liquld N; SIow = o.S·C/mln. Av¡¡ Valu •• (%) 1 lliabilily 71A 57A 148 OS 90A 44A 83AB 33A 60S 118 67A8 OS 86A 57A !IIlB 29B 5.2.2 Toward a haplold method: culture 0' Immature pollen Shoots Haploids and double haploids will be of significant value to cassava as a research tool with various applications in genetics, evolution, expression of'important recessive genes and a braeding system for TeS. Initially, background research was conductad on tha reproductiva behavior of cassava, including in situ and in vitro pollen germination, pollen tube growth, and isolation and culture of young zygotic embryos. The microsprogenesis of cassava was a1so characterized, and macroscopic parameters haya been defined that correlate with critica! microspora developmental stages. The in vitro culture 01 whole anthers has led only to the formation 01 ca!lus; therefore in the last year, the emphasis has been on devising a technique for culturing isolated immature pollen. The technique utilizes some 50 fiowers; Le., 50,000 microspores in hanging droplets, per petri dish. Immature palien at the mitosis stage did not divide in culture as expected; rather it rapidly formed exine components that rendered it impermeable to medium components. Thus work has shifted to the culture of microspores at the tetrad stage (Fig. 5.28). Optimal culture density was 10" microspores/ml medium. Tetrad-stage microspores cultured in high osmotic concentration and then transferred to a medium with a lower osmotic 89 concentration were able to divide mitotically and gradually pushed their way out 01 the tetrad callose. Outside, the tetrad division continued, but at a slow rate. Mícrocalli then become visible (Fig. 5.28). Mlcrospores, Isolated from highly fertila varietles responded better than those from highly sterile ones. In summary, it has been possible to change the gametophytlc fate of cassava microspores to initiate sporophytic development. This is tha first necessary step toward androgenetic development. Future activities will focus on enhancing callus induction from tetrad microspores and monitoring this development cytologieaUy. 5.2.3 Genetlc transformatlon A basic requisite for genetie transformation is the ability lO regenerate plants 1rom cell culture. The regeneration of cassava plants by somatíc embryogenesis on immature leaves or apical meristems was demonstrated earlier (Fig. 5.38). Development of genatic transformation requires the avaílability 01: .. A genetie eonstruct containlng appropriate selectable and/or reporter markers, a housekeeping or tissue-specífie promoter .. A transfer vector or the teehnology for dlrect DNA transfer .. An effieient plant regeneratlon protocol .. Technology to monitor the transcri¡)tional and posttranslationat products 01 gene express ion and the Inheritance of the introdueed gene 5.2.3.1 Agrobacterium-mediated genetie transformation. • Because the efficlency of Agrobacterium infection of plant tissues is genotypic dependant, 4 cassava varo (M Col 22, M Col 1505, M Mex 55 & M Cub 74) were first screened with 25 A. tumefaciens stratns. M Col 1505, whieh atso regenerates well through somatie embryogenesis, resulted highly susceptible to strains 1182, 1183, C58C, B6S3 and EHA 101. • The plasmid construction pGV 1040, provided by Plant Genetles System (PGS) , Belgium was used. It contains two selectable markers (bar gene & NIPTII gene) and one reporter gene (gus A), which ara driven by strong promoters (NOS & CMVS 35). • Even at the lowest X-glu concentration (Jefferson's method), cassava somatic embryos produced endogenous GUS (8-glucuronidase) activity, whieh did not occur in leaves and stems. Using Kosugi's techniqua, somatic embryos were still GUS + with 2X and 90 3x X-glu concentrations; but endogenous GUS activity was not detectable at the lowest X-glu concentration. • leaves and stems were much more susceptible (8 mg/It) than somatic embryos (32 mg/It) to the product of the bar gene (herbidde BASTA = phosphinothricin). • The antibiotic Kanamycin resulted highly detrimentaJ to somatic embryogenesis. The condmons for efficientA. tumefaciens infection of M Col 1505 will now be defined in order to iomate work on transformation and regeneration. 80th single vector strains as well as the binary vector approach wifl be used. 5.2.3.2 Genetic transtormatjon by particle bombardment. The partide gun aveilable in the BRU accelerates ONA-coated metailic microproyectiles to high spead in such a way that they can penetrate cel! walls, thus being a vehicle for transformation with foreign ONA. • Transient GUS activity was demonstrated in cassava tissues 3 days after bombardment. • The vacuum for effective particle acceleration with minimum cell damage was < 550 mm Hg, and the distance of the specimen to the gun, 22 cm. • Freshly prepared plasmid ONA solution produced stronger GUS transient expression on cassave tissue. • Recently GUS expression was obtained on the tips 01 somatic embryos 3 days after bombarding embryogenic callus at the globular stage (Fig. 5.JA). Work continues on increasing the frequency and intensity of GUS expression on somatic embryos. This work should lead to experiments on stable genetic transformation of somatic embtyos using herbicide resistance as a selection factor. 5.2.4 Molecular mapping With the support of the Rockefeller Foundation (RF), a research project was begun to construct highly saturated molecular and physical maps of cassava using random genomic, cONA and YAC libraries. The project involves cooperation among the U. of Georgia, Washington U., lITA and CIAT. The maps wiJI be useful for analyzing the genomic structure of cassava and ita wild relatives and for tagging agronomically important traits, simply and quantitatively ¡nherited. Eventually the map witl be useful to ¡solate and clone cassava genes. 91 B _Rl _m _1 ;, ' "Be {) A B COA BC O AII cn AS cnA!I eo A B e o FIgure 5.3. (A) Genetlc transformation of cassava somatlc embryos: (1) somatlc embryos lnduced on meristem tips; (2) plantJets regenerated through somatlc embryogenesis; (3) transient GUS expressfon on globular somatic embryos (arrow); and (8) restriction fragment length poIymorphlsm (RFLP) wilh 7 restriction enzymes between A = M Thal 8, B = M Col 1505, e = M Nlg 5 and o = M. aescullfolfa. l:!Il:ú poIymorphlsms onIy wilh M. aesculifo/ia. 92 The projeot has built upon the experience gained at CIAT in the las! year. Thus protocols for DNA extraction, digestion and hybridization have been implemented. A Pst I random genomic library was generated using varo M Col 22. Some 200 single-copy clones of a total of 500 have been eharaeterized; Insert size ranges from 0.2 to 3 Kb. Recent developments in this projeot inelude: • Cassava DNA digested with Eeo RI, Eco RV, Xba I and Hind llllibraries provide probes with highest polymorphisms. • Polymorphism between var. M Col 22 and its wild relative M. aesculifolia was dramatically higher than the polymorphism found between two cultivated genotypes (Table 5.3, Fig. 5.38). Four cutter restriction enzymes dísplayed less frequency of polymorphísm then six cutter enzymes for cultivated genotypes. • POlymorphism displayed by Ora I was extremely low, indícatíng that regions rieh in adenine and thymine may not be spots for mutatíons in cassava. 5.2.5 Starch quality and COl asslmilatlon • Cassava sour starch, obtalned by natural fermentation, presents quality fluotuations that relate to the fermentation process itse/f. Previous research in Cassava Utilization described the termentation mechanisms and the role of!he microorganisms involved. Because the main substrate for these microorganisms Is starch, amylolitic aotivities must playa central role. A methodology is needed to determine and differentiate the enzymatic systems of the microbia¡ aotivity and to measure the produotion of excreted enzymes. Table 5.3. Ubrary Pst I Xbal Hlnd 111 BamHI !legre" 01 polymorphlsm de_ with al leeal oo. restrletion enzyme' .olng !!ve random genomic Ubtarie. ea _rce 01 probos. M Col 22 YO. M Col 1505 M. aesculílblla YO. M Col 1505 60 85 55 60 30 45 , Bast f8$ItiQtfon en2ymes, In ord., 01 polymorphlom, are: Eco AV, Xba l. Eco Al, Hlnd 111. 93 In collaboration with the Cassava Program, an electrophoretic methodology was developed to determine amylolitic activities. The method will be useful for selecting and characterizing amylolitic microorganisms by their respective electrophoretic patterns. These patterns in turn wiil rnake it possible to monitor the development of the amylolitic activities throughout the fermentation process. The causes for quality fluctuations will be identified and strategies tor improving the microorganism activities through genetic manipulation will be proposed . • Certain cassava photosynthetic parameters líe between ca (e.g., bean) and C4 (e.g., maize) plants. This suggests that cassava might have developed mechanisms for efficient photosynthesis, especislly under drought and high temp conditions. Early work at CIAT has shown enzymatic indications of this behavior, far which there is a wide range of genotypic variability. Now, it remains to be establíshed what are the actual mechanisms underlying these characteristics. The final objective is to use this knowledge for developing screening methodologies that can be incorporated into breeding programs. One approach that is being initiated is te develop in situ hybridizatien"techniques en leaf tissue using photosynthetic gene probas to assess their compartmentalization patterns. Homologous genes from cassava have to be used; therefore, these genes must first be isolated from cassava gene librarles using heterologous probas from maize (provided by Yale U.). The genomic library is being constructed. High homology has been observed between the maize genes and cassava total ONA, which indicates that these genes are highly conserved, thereby tacilitating the task of isolating the eassava genes. The genes being working are: RlIBISCO (rbes), malata dehydrogenase (rndh), malie enzyme (me) and PEP-carboxylase (Ppc). Another target would be the glycine carboxylase, an enzyme involved in CO. recycling-a mechanism that may be related to cassava's C3-C4 behavior. 5.3 Instltutlon Building 5.3.1 Trainlng/biotechnology issues Over the last four years 40 people from developing countries have participated in various training activities in cassava biotechnology. As a result of the training and follow-up activities, tissue culture facilities far cassava micropropagation have been implemented in China, the Philippines, Panama, Venezuela, Cuba, Peru, Brazil and Mexico. National agricultural research institutions such as CNPMF in Bahia (Brazil) have also requested and received training on cassava isozyme fingerprinting. In 1990-91 CIAT developed its Biosafety Guidelines and established the Institutional Biosafety Committee (IBC). The guidelines comprise peer-approved norms tor the 94 adequate and safe handling of R-DNA research in the lab and the testing of genetically engineered organisms in the greenhouse and the field. 5.3.2 The cassava blotechnology network (CBN) With the collaboration of liTA and many national and international organizations interested in cassava, the CBN was founded in a workshop held at CIAT in September 1988. The objective of the CBN is to facilitate the implementation of modern biological approaches and tools to contribute to the solution of priority problems in cassava production and utilization that are recalcitrant to traditional methodologies. The CBN will facilitate: .. The organization of meetings on cassava biotechnology and more basic research .. Effective mechanisms for developing country access te genetic resources, genetic constructs, technologies and information .. Participation 01 developing country scientists in biotech training activities .. Discussion and actions on critica! issues such as biosafety and intellectual property protection The network approach to cassava biotechnology has received wide acceptance by the scientific community and national and international funding agencies. Since íts foundation, , the number of proJects funded has steadily increased; currently 22 projects dealing with research constraints and technical bottlenecks in cassava are under way (Table 5.4). The CBN Steering Committee (SC), elected at the founding meeting at CIAT, hes met twice. Because of the interest demonstrated by the Dutch Govt. in biotech developments (particularly in cassava) for developing countries, the se recommended approaching the oolS, Ministry of Foreign Affairs, the Netherlands with a propasal tor funding the activities of the CBN for five years. The propasal seeks financial support far: .. A full-time coordinator /scientist .. Organization of scientific network meetings avery 2 yr and ad hoc SC meetings .. Training of developing country scientists .. Bridging funds tor critical research areas requiring attention .. Newsletter and brochure tor the network 95 T able 5.4. _arch projects underway In cassava biolechnology in _ped and developillll oountry Inotltution. (October 1991). Subleet Cyanogenesls Virus resistance: Inseet resistanee Photosynlhasls Plant regeneration Genetic transformation Radiatian induction ONA fln¡;¡orprinting Molecular mapping Cryopr ... rvatlon Qthe, actlllities at cIAr & lITA Instltutions U. 01 Nawcaotle Upon Tyne Royal Vot.AgrlC.U.o eopenhagon Mahldol U'o Bangkok Chio Stale U. Colombu. F"", U.o Amsterdam Washington U'o SI. Loui. WashinglOn Stata U., Pullman Australian Nat. U'o Canben'a U. 01 Georgia, Alhens U. París. Orsay U. 01 ElatI\ U. o121mbabwo, Hatar9 CIAT South China Inst. Bclllny, Guangzhou U. Nottinghem U. of Guelph, Canada CENARGEN, B Na-hypochloride solution. • Cassava debris from previous plantings should be bumed befare plenting new selected stakes to eliminate volunteer plants from affected stock. • Suspected diseased plants as well as volunteer plents from previous plantings should be rogued. To facilitate disease identification, weekly inspections of plantations are recommended. This production system end a description of disease symptoms were published as a pamphlet that has been distributed to farmers by extension agents and community leaders. Several demonstratíon plots have also been planted for field days. As a result disease incidence has been reduced and AY increases are startlng to be shown in several 101 locations of the area; for example, RY increases obtained by 4 growers who followad the recommendations and took selected material from a 25%-affectad plantation averaged 51% (range 23.5 to 87.1%) (Table 6.3). In addition to the above the following projects are in progress: • A program has been implemented to produce planting material on a commercial scele in areas where ecologicel condítions restrict pathogén multiplication and invasion. Material will be distributed to growers duríng the 1992 planting season (Jan.¡Feb.). • In cooperation with the Plant Science Institute (Microbiology and Plant Pathology Laboratory), the BRU and VRU at CIAT MLO-cessava probas will be usad to develop a program to provide MLO-free planting stocks, identify naturally infected hosts, insed vectors, and various epidemiological features relatad to this MLO. 6.4 SuperelongatJon Dlsease Studies on genetic control of SED of cassava (causad' by EIsintJe brasiliensis) covered the following areas: 6.4.1 F1 morphologlcal reslstance Histological studies on 3-mo-old shootS of resistant (CM 523-7 and M Ven n) and susceptible (M Col 22 and M Col 113) clones grown at different líght intensities showad that the thickness values of the external layers of the stems 01 resistant clones were Table 6.3. Gro_ No. 2 3 4 l1 RV incre_ oblalnad by 4 (lfOWér. who uaod ... _ planllng material taken from symplOml ... planl$ from • 25% WB-affoctad planl$don. ccmpared wiIh RVa of neIghboIIng piola where plandng material WlIS no! ... _'. l'tHh RV (t~l' RV I"",ease Selected Stak •• Unaetoctad Stak ... ('l!ol 16.3 13.2 23.5 15.6 10.0 58.0 16.5 12.0 37.5 17.4 9.3 87.1 16.5 11.0 51.0 _. and tooIs _ weradisinfested wilh Na./lypOd1I_ (2%); dabris from previous plandng wa. bum.d and suopocted plants wsre !CgUed. • Data tak.n from 15 pI!4 reps/gfOWér. 102 significantly (5%) greater than those 01 the susceptible clones when the plantlets were incubated at 20,600 to 24,000 Ix. This ratio decreases at Iow light ¡ntensities; e.g., at 7.200 Ix or lower. the thickness values decreasad dramatically, differences between resistant and susceptible clones disappearlng. Temp and RH had no effeet on these parameters. Similarly. measurements on otOOr stem tissues showed no significant differences between susceptible and resistant clones. Further evaluation of 30 clones whose susceptibility I resistance values were previously known showed the same results. This opens up the possibility of devising a rapid screening system for resistance to E. braS/7iensis. which can replace the current field screeníng evaluation system. 6.4.2 Reld evaluaUon To determine the most appropriate location for fie!d evaluation of genotypes tor resistance to SED, genotypes belonging to 17 familles were planted in three locations wOOre the disease Is endemic:Carimagua and Villavicenclo (Colombia), and Huimanguillo (Mexico). Diseasa scores were recorded at the height of the epidemic. and the probability for finding resistant clones among genotypes qf each family per iocation was calculated according to Grizzle. Starmer and Koch (1969)"". At Carimagua disease severity was higOOst. reducing the probability 01 finding plants with Iow disease damage in each family (avg disease scores were also highest). Consequently. this location appears to be the most appropriate for evaluating resistance to SED in the field. 6.4.3 Stablllty analysls Stability of resistance to SED and CBB was investigated by planting clones of three famiHes over a 5-yr periodo According 10 Digby's modified joínt regression method: • The general response of the clones to SED and CBB was unstable. • Despite this instability, differences in levels of resistance between families and among genotypes belonging to the same family were identified. • Disease pressure differed each year, justifying clonal evaluation for several growing cycles. • "The families that showed the most stability for resistance were CG 890 (CM 723-3 x M Col 638) and CM 3581 (CM 849-1 x CM 523-7). These parents have intermediate to high resistance to both SED and CBB. 0" Grlzzle. J,e.. Starmer. C.F., and Koch. G.G. 1969. Analysls of categorlcal data by linear models. Blometrics 25:489-504. 103 6.5 Cultural Practlces to Control csa 6.5.1 Mixed cropping The effect of mixed cropping cassava with maize was evaluated at Media Luna (North Coast of COlombia) by planting piots where two cassava rows alternatad with two rows of maize. Severity of two foliar diseases (CSS and brown leaf spot, SLS) was significantly (0.05 leveQ less in mixed cropping than monocropped plots. RYs were also significantly higher (Table 6.4), and there was a positive effect of mixed cropping in controlling root rots. Reduction of both CSS severity and incidence by mixed cropping cassava wíth maize was evident in commercial farmers' fields: 30% CSS incidence vs. 85% far monocropping (highest rating == 4.0). 6.5.2 Pruning This cultural practica (CIAT Annual Report 1972) was validated on several commercial cassava plantations on the North C9ast of Colombia (Table 6.5): Soth disease incidence and severity ware much lower on pruned plantations. On mixed cropped cassava-maize plantations, however, CSS incidence and severity were much lower. Consequently, pruning most of the aboveground portian of infectad plants and burning infectad debris in severely infected plantations are effective means of slowing down the spread of the pathogen and decreasing inoculum potential; however, this should be done only at the end of the rainy season to prevent new reinfection through rainfall splash. 6.5.3 Effect of Caa-Infectad stakes on RY Dissemination of the CSB pathogen by stakes and botanical seeds taken from affected plants has been demonstrated (CIAT Annual Reports 1972 & 1980). but quantification of the effect of using CBB-infected stakes far planting on RY was not investigated unti11987. When planting 25% CBB-infected stakes of a susceptible clone, losses in RY can be ca. 72%; from a 100%-affected plantation, 26% RY reduction in resistant clones, a1though much lower. can be of significant importance (Table 6.6). This shows the importanca of using CSS-free stakes for planting as an important cultural practica for controlling CBB. 6.6 Root Rot Problems Root rots are the most important pathological problems of cassava, especially in those areas where the crop ls planted during succassive cycles or in flooded areas, badly drained soils andjor during abnormally heavy rainy seasons. The most important root rot problems are induced by species of Diptodia (D. manihotis), Scytatidium (Scytalidium sp.), Fusarium, (F. oxysporum and F. solan/) and Phytophthora (P. drechsleri and P. nicotianae varo nicotianae). Each species requires eertain edaphoclimatic conditions, which restriet their incidence to specific ECZ. Consequently, research directed to 104 Table 6.4. I!!fect of CII_ 1M Ccl221$)-meID mIxad croppIng en _ AY (IJIIaland &IMIfIty of ces and BLS. _AY tI",al 16.7a 9.6b 13.7& 2.41> 3.7a 2.3b 3.5& , Avg data taken frcM 4 ptotsjtreatmenl wiIII 400 cassava p_ mixed cropped evel)' 2 rows _ malze and 800 ea_va plants ulld .. monoculture. Dala betwQn plantlng sy&Iema foIlowed by dlfferent 1_ are tignifieantly dIfIerent al 0.051eve1$ (DMRT). 2 0Is0au aeverity: ceB: 1 • leal 1If'CIS; 5 • plan! death or dloback. BlS: combinad data en _ leal ..... a and pareenl of ..- Iuves/pI. Table 6.5. llicldenca and &IMIfIty of ces on 5- 10 6-mo-oId comma""'" plantatIona _ caaaava monocropp!ng (ptuned or unpruned) and __ 2IH11ixed cropplng In Mandinga tBoIlvar, North Coaat of ColombIa). Cropplng 8y 1.0 1.0 2.0 lncIdence' fJ>1 3.6 100 50 1.1 10 , 01_ aeverity: 1 • angular leafllf'ClS; !I • pIant _ or death. 1ncIdence: % _ pi_in !he _ planlalien. • Avg data taken frcM pIanIaIIons rangIng frcM 0.$10 2.0 ha. o Ptunlng: moa! cauava pIants abOvagn:>und __ cuI, _ and ltumed. 4 Mixod cropplng: aI11Imato rows of cauava and ma/ze. TMIe 6.6. AY reductlon due 10 Ihe .... of CB8-lnIecled ablkaa lo, piantlng In a ceB-favo,ablelocallon. lI!!!I!l!!l!!i!2l1 Con! _.tantQone 'l!.CB8-InIecled AY' AY Reductlon AY AY Reductlon Stak •• IItha) ('l!.) (IJIIa) ('lf.) O 28.9 28.3 25 20.4 28.4 22.6 9.8 50 15.8 45.3 23.4 7.5 75 17.9 38.1 19.5 22.9 100 8.1 72.0 16.7 26.1 , Data taken ftom 6 repsjtreatmant, 30 planta eacl\. 105 controlling root rot problems ot cassava is location specific and tends to integrate cultural, chemical (only for stake treatments), bioIogícal and varietal resistance control measures. The following Isa summary of the research directed to control root rots in several regions of Latin Ameriea. 6.6.1 Media Luna (North Coast 01 Colombia) The Media Luna region is characterized by sandy soils with low OM (0.3%) and nearly neutral pH (6.8). The land has been used intensively for several deeades (mostly for cassava production) under a traditlonal system characterized by very limited inpUls. As a result, soils have been depleted of nutríents and have a high accumulation of inoculum potential of D. manihotis and F. oxysporum, which reduces crop establishment and plant vigor, and causes severe root rots (ea. 40% of total production). These two pathogens survive by affecting alternate host species (D. manlhoiis = Vigna ungulou/ata, Croia/arla espectabllis, Sesamum indlcum, Zea mays and Mangifera Indica; F. oxysporum .. C. espectabllis, V. unguícu/ata, S. Indicum), volunteer cassava plants, eassava debris or by infesting the soils. Initial RYs were relatively high (>20 tfha aooording to farmers' information) but have decreased sharply during the last deeade, (avg of only 3-4 tfha in 1988). Research developed by different sections of the Cassava Program led to the definitlon of the following cassava production systems for Media Luna: • Land should be left fallow if cassava has been cultivated for more than 4 consecutive oycles or when root rot is higher than 3%. Oebris from weeds must be incorporated , into the soil befare planting cassava by plowing once and disking twice. • Stakes should be taken from 10- te 12-mo-old plants of high-yielding, root rot-tolerant clones. Stakes must be 15-20 cm long (containing at least 5 nodes/stake), seleetad visually and treated for 20 min in a suspension of dimethoate (1 oo/It), benornyl and eaptan (3 g/It each) before planting. IItPlots should be sprayed after planting with diuron (1 kg/ha) and alachlor (1.5 It/ha) as a preemergent weed control practice. Aftar 2 to 3 mo weeds can be controlled manually during the growing oycle (11 mol as required. • Plots should be fertilized (15-15-15 NPK at 300 kg/ha) at arate of 100 kg/ha at 30,60 and 90 days after planting. This production package was vaJidated with great success during 1988-89 by 12 eassava growers and in 2 demonstration triaJs. They obtained a RY ¡ncrease of 208% when they planted a resistant clone and 300% when the clone was susceptible. More than 30 cassava growers were successfully growing cassava following this package in 1989-90. 106 6.6.2 Varzea region of the Amazon cassava is being planted in two different regions ofthe Amazon: (a) coastal or unflooded area (terra firme) with extremely acid (pH 4.0) and infertile soils; (b) varzea, or the flooded basin edges of rivers with a1luvial, fertíla and nearly neutral pH soils. On the former cassava AYs are ca. 6 t/ha/yr; in the latter > 20 t/ha after 6 mo. In 1983 asevere root rot outbreak was reportad in the varzea area, causing 60% Iossas on a1most 84,000 ha of cassava. A coIlaborative research proJect with CPM-Manaos (Centro de Pesquisa Agroforestal da Amazonia Occidental), the CNPMF and CIAT included research on control of the cassava root rot problem Qnduced by P. drechsJeri and F. so/aní) through varietal resistance, cultural practlces (crop rotation, mixed cropping, planting on ridges and selection of planting material) and chemical treatment of stakes. This led to the definition of the following cassava production package far the region: • Infested land (>3% root rot in previous harvest) should be rotated with maize or rice during the dry season. . • Soils must be plowed once, disked twice, and then ridges (about 0.3 m high) constructed; drainage canals should be constructed or improved. • Twanty-cm stakes from lignified stems of 10. lo 12-mo-old plants of tolerant clonas should be selected far absence of rot or mechanical injuries. Planting material should be produced in the coastaI area. • Selected stakes must be treatad by dippíng for 10 min in a suspension off fosetyl-A1 80% at a concentration of 2 gflt. Experimental results using this system and the cumulative effed on RY induced by each component of this system are presented in Tabla 6.7. This production systam was *a11y released during a field day that CPAA/CNPMF-EMBRAPA/CIAT organized in September, 1990, and is being evaluated by more than 200 cassava growers in 10 !T1Jnicipelities. Three resistant clones have been identified after screening more than 286 clones over a 5-yr periodo 6.6.3 CIAT HQ The Palmira Station is characterized by clayey solls with hlgh OM contento Planting cassava far more than 4 consecutive cycles decreases RYs sharply, and populatlons of F. oxysporum, F. soIani and P. nicotíanae varo nicotíanae ¡ncrease to levels of around 10" cfu/g of sol!. Crop rotation far one year with any crop comblnation of sorghum, maize, beans and even fallow recovered AY to original levels and were 3.2 times hlgher than controls. 107 Tabl.6.7. Cumuiatlv. affect on AY of On. toleran! and 15 stJsceptible clon •• of vario ... cullural practl .... for conlfolllng root rota Induced by P. drech.léri and F ... Ian/In tIle va",.a ragion 01 !he Amazon. Tolerant Susceptible Cultural Practi",. Clone CIones1 Traditlonal pfanting 6.5 0.1 Roiation wíIh malze Cf rI ... 16.0 5.0 Ora/AagO and planting on rldge. 22.5 8.0 Salactian 01 _es 24.5 5.5 FosetyI.A! 80% (2g/l!llfeatmenl 01 stak •• 29.3 6' .0 In!egrsted 'ystem 29.3 7.3 1 /Wg data Ior 15 _ptible clones. 6.6.4 North and Northeast Brazil The North and NE regions of Brazil produce 63.4% of the 21.6 million t of cassava roots that are produced on 1.76 million ha. Of this production 85% is directly or indirectly used as staple food for more than 45 million people. Asevere outbreak of cassava root rot in 1986 affected around 350,000 ha. Initial etiological studies revealed that the problem was induced by F. solani, F. oxysporum, P. drechsleri, D. manihotis and Scytalidium sp., which are located in different cassava-growing areas. A collaborative project is being developed with the CNPMF and the following state agencies: CPAA-Manaos, CPATU-Pará (Centro de Pesquisa Agroforestal del Tropico Umedo), IPA-Pemambuco (Instituto de Pesquisa Agropecuária), EMEPA-Paraiba (Empresa Estadual de Pesquisa Agropecuária), EPEAL-Alagoas (Empresa de Pesquisa Atricola), ENDAGRO-Sergipe (Empresa de Desenvolvimento Agrrcola, Rural e Operacional) and AMIDOGLUCOSE-Sergipe (a private enterprise). The primary objectives aII8 to develop technological packages for controlling the problem in each geographic area, promete their use, interchange resistan! genotypes, and build up collections of native varieties as sources of resistance. Experimental plots hava already been planted and project status ls as follows: • North Brazil .. CPATU-Pará. Appropriate cultural practices for the region are being defined. Twelve root rot-resistant clones have been serected, 2 of which have shown high RY potential and carbohydrate content, and Iow HCN contento These will be released to growers in 1993. 108 lO CPAA-Amazonas. A production package for the varzea region was defined and 2 root rot-rasistant clonas were released for tha first time in 1990 (see p. 6.11); a new resistant clone will be released in 1992. The technological package for tha litoral region is under definition, and a collection from the Amazon is building up, part of which is under evaluation in the litoral region . • NE Brazil lO CNPMF-Bahla. Etiological and epidemiological work is under way. F, hybrids from root rot resistant genotypes are being produced. lO Indlaroba Farm-Serglpe (a prlvate enterprise). Twenty-five salected clones are being evaluated for earllnass with the aim of avoidlng the root rot problem in the region, which normally attacks the swollen roots 10 mo after planting. lO ENDAGRO-Sergipe. Eighty-five clones are being evaluatad for root rot resistance. The technologlcal package is undar definition. ... EPEAL-Alagoas. F'lVe root rot-resistant clones, selected from among 50 clones evaluated in 1986, are in advanced fiald trlals. Two clones will be released in 1992. lO IPA-Pernambuco. Root rot-resistant clones from Mansos ware highly resistant in this state, but show susceptibility to Scytalldlum root rot and to adaptation. F, clones from root rot-resistant ganotypes are belng produced. ... EMEPA-Paralba. A asease survey for the state,! as well as evaluation trials, is underway. &:.7 Biologlcal Control of Cassava Dlseases Beneficial fluorescent pseudomonads and strains of Trichoderma harzianum (commonly found in mostcassava-growing areas ofthe Cauca Valley and neighboring hillsida areas) have been used In the biologlcal control of cassava diseases. Pseudomonads ware used because of their nutritional diversity, ability to grow under a wide range of envíronmental conditions, and ability to coloniza tha rhizosphere of many plant species. 6.7.1 Fluorescent pseudomonads 6.7.1.1 lsolation and surviyal of beneficial bacteria. Strains of Pseudomonas putlda and P. fluorescens were isolatad from soil or the plant rhizosphere of cflfferent cassava clones by using King's B (KB) medlum incubated at 27· C for 24-36 h. Bacterial isolates showing fluorescence on KB under ultravíolet light were purified from single colonies after serial dilutíons seeded on KB medium. High bacterial populations were found in the 109 rhizosphere of cassava plants or rooted shoot tips 2 mo after inoculation. There were variations in bacterial populations among clones and between strains; but a clone X strain interaction was not found, which indicates lack of host specificity with the strain tested. Survíval of fluorescant bacteria is probably dependant upon nutrient availability and/or spaca and callular migration to new sites. Similarty, quality of exudates (e.g., production and guantity of specific amino aelds or toxie eompounds) may influence their survival in the rhizosphere, in whieh case these characters could be genetically controlled. 6.7.1.2 Effect on plantlets under glasshouse conditions. Isolates of beneficial fluorescant pseudomonads have been characterized according to their inhibitory eftect in vitro on both bacterial and or fungal pathogens (1) and ability to promote roat system growth of plantlsts as related to pathogen inhibition in vitro (11). Wrth method I tour groups of isolates were identified: (a) those with no or very mild inhibition of pathogenic bacteria and fungi; (b) those that strongly inhibit pathogenlc bacteria, with no or very mild inhibition of fungi; (e) those causing very strong inhibition of fungi, with no or very mild inhibition of bacteria; and (d) those inducing strong inhibition of both fungi and bacteria. Generally, there were more isolates able to inhibit pathogenic bacteria than fungi. Among all isolates collected, those of P. fluorescens showed broader in vitro inhibition of the bacterial and fungal species tested than P. putida. Isolates characterizad by method 11 were also classified into tour groups: (a) those that did not inhibit cassava pathogens in vitro or increased root growth of cassava plantlets; (b) strains showing a high ínhíbitory etfeot in vitro of cassava pathogens (both fungi and bacteria; some strains of group d in méthod 1), but not promoting an increase in foliage or root system of inoculated clones (secondary metabolltas may haya phytotoxic or antibiotie activity); (e) strains that did not inhibit cassava pathogens but increased both,the foliage and root system of inoculated clones (antagonistic interactions wIth pathogens in the form of nutrient competition or antibiotic effects may have resulted in the exelusion of patliogens from the rhizosphere); and (d) straíns inducing a strong inhibitory effect on cassava pathogens and a significant increase in the aerial parts and roots of inoculated plantlets. It has been reportad that sorne strains of fluorescent pseudomonads produce growth regulators for other crop sJl8cies. 6.7.1.3 Practical agglications. The following applications can be envisaged at various steps in systems of cassava aimed at increasing production: • Biological control of foliar pathogens. Strains from method I group b, which induced the highest in vitro inhibition of xanthomonas campestrls pv. manihotis (causal agent of CBB), were usad to spray plots of susceptible and resistant clones plantad in an area where CBB is epidemic (fable 6.8). Both the no. of angular leaf spots/leat and too no. of blightad leaves/pl were significantly reducad by foliar applications of a strain of P. pulida. RY of the susceptible done (M Col 22), but not of the resistant done, increased 2.7 times. 110 Table a.a Clone M Col 22 lAVen n AY (t/ha). no. angular leal spolSfleal and no. 01 blighted leaves/pl 01 CBB-susceptible (M 00122) and resistanl (M Ven 77) elones afler foliar applications 01 Pseudomonas putida (stratn F-44). Scoring $ystem Sprayed 1'1018 Oontroll'lols No. angular leal spolSfleal' Mb 17.ea No. 01 blighted IO."'/pl 1.3b 5.5& AY" 6.Sa 2.5b No. angulat leal .poto/lea' 1.20 U5a No. of bllghled lo_/pl 1.7a 1.98 AY 9.&1 9.1. • i'lanlS received 6 follar applications of 0.5><10" ofu/ml of a su~on of P.putida 8115<1ay 1"IeNaI •. Control planto wero aprsyed wi1h distiHed ~O. Dala represen! avg no. of _ apo1sflea/ of 15 teaves/plot of 36 pi wilh 3 ,aps; and avg no. of bllghled leaves/pl of 20 _/piot. • AY n!!I::Otded ftom 3 tapo of 30 planta each (12 h_sIed pIanIS). Dala followed by lile sama let!er are not tignHlcanlly_ (0.05 QMRT). Control of other foliar pathogens of cassava by spray applications of specific strains of benefICial fluorescent pseudomonads has not been reportad. but it is possible. Practical and economic problems exist, such as base inoculum production, inoculum storage and preparation under aseptie conditions, and costs of spray applications. Further research is requirad to test the practical feasibility of controlling foliar pathogens of cassava . • Biological control of preharvest root rots. When soils infested with Pythium spp. or D. manihotis were drenchad with a bacterial suspension of P. putida (method 1, group d) before planting cassava plantlets, satisfactory control of root rot was obtained. Strains of P. fluorescens (method 1, group e) were also able to protect stakes against D. manihotís (fables 6.9 & 6.10). The protective effect was evident when stakes were treated with the bacterial suspension before or after fungal inoculations. Protection was relatad to sproutíng of buds and fungal establishment, as wel1 as invasion through the tissues of the stakes (Table 6.9). This type of protection was also evident for three clones taken either from farmers' fields or from meristem-culture derived plants (MCP). Bacterial protection was nearly as efficient as that obtained with the best fungicidal treatment (Table 6.10). Fresh RY increasad when plants were waterad with a 10-ml bacterial suspension of a beneficial strain of P. fluorescens (method 11, group d). AY tended to increase with an increased no. of waterings (Table 6.11). However, levels of increase variad according to clone, suggesting differentíal clonal responses to bacterial strains usad or differences in susceptibility of the pathogens to the beneficial bacteria. In another trial commercial plata of M Col 2215 (susceptible) and M Col 1505 (resistant) were plantad in a D. manihotís-endemic area (La Colorada, North Coast of Colombia), where CBB 111 Table6.9. T reatrnefÚ$ t 10mln In PI. CSa 20min in Om. 10 mln In Qtb2 20mln in Qm. lah In Dm. 20 mln In DH"O 20mln InDm. Controls" Effect 01 PSeudomonaslluo,...cens (_n PI. CSa) on aprouting 01 _es (M Coi 1684) alter Inocula!íon wilh DlpIodia manihol!. 10m.) Sptouting' Stake Infec!Ion (%01_) (%of~n .... os) 100a3 16b IOC 97a 4()b 85a OC llJOa llJOa OC , Oip Ir_ont In suspenslons 01 1.1xl0· cfu¡ml 01 P. /lucre.""". (8Iraln PI.CSajin dlsllllod water (DH¡O), followed by a dlp ueatmenl in a suapenslen 01 5.&10' pycniospore.jml 01 Dm. or viceversa. 2 Date tal 11.31> 12.3ab 13.3& M 0011468 38.1b 40.2b 44.5ab 49.3& M Col 12 42.1b 46.5ab 46.8ab 47.0. 1 IlacteriaJ lrealmanl$: 10 mi ola _01 susponalon (I.lxlo" oIu/ml) _. pou'adatllle baee of """" plan! at 1 = 1 mo; 2 • 1 and 2 mo; 3 • " 2 and 3 mo alter plantlng. • Data lakan from 3 rep. of 30 pl/location/clone anó trealmant. Sordar planl$ __ eflminated. AY data foIlowad by lile sama -,(s), comparad across bacteria treatments, ..... no! slgnilloantly dilferent al lile 0.015 level (DMfIT). Table 6.12. Effect of Pseudornonas f/uorescens (Slraln P.f. C5a) on commerclal AY 01 1,,,,,,, rool. of 2 clone. (M 0012215 = susceptible; and M Col 15015 = resietanl) planlad in a Dip/odla man_ .. ndemlc at .... wilh moderate CSS and anlhracnose inlection during 111. growlng cyde. Seuroo 01 SIak •• Oommerclel plsntalions-becterized 1 MCP-bectsrlzed' Commerclal plantatlons--untreated MCP-untreated M Col 2215 2ge 16ab 13 b AY ~/h8)/CIon. M 0011505 20a 20a 17a ISa 1 Bectarlzation: aIake. dip treated in a bacteria! suspension o! Ixl o" oIujml; and plata splayed monlhly willl a bacteria! .uspension of 5x1O" oIu/ml. • Plota harvesled ID me alter plantlng; resulto lollowed by lile samele!!., not statlotloally ditferent (DMRT). 113 shown on untreated plants from stakes taken from MCP. indícating the efficiency of the protective effect asserted by the native microflora living on stakes taken from commercial plantations. Finally. the bacterial treatments did not significantly íncrease RY 01 the resistant clone because the intrinsic genetic resistance in this clone overcame pathogen-induced stress. Controlling root rots of cassava with beneficial bacteria, replacing fungicidal treatments of stakes. is teasible; however. further research ia needed to define practical systema. Table 6.11. Location Carlmagua Media Luna CIAT.HQ Iwg E!'( 01 fraah roots 01 ~ oIon .. pIIanted al Carlmllg .... Media Luna and ClAT in I$Iaticn la fraatmanlll wiIh Pseudomonas _. (PI-6IlI. RY (t/ha¡ in RoIation la Bacterial Treatmenta' Clone O 2 3 M Col 1914 12.11>" 14.5b 16.5& 17.6& M Col 1916 11.2b . 14.5b 15.7a 17.7a M Pan 19 10.7. 10.1a 12.0. 12.3& M 'len n 14.5b 14.5b 18.1a 21.5& CM 342·170 9.011 l1.1ab 11.3ab 12.5& M Col 72 12.5b 16.0. 16.3& 18.3& M 0012215 10.3b 11.3b 12.3ab 13.3& M Col 1468 38.1b 4O.2b 44.1Sab 49.3& MCoIn 42.1b 48.6ab 48.6ab 47.0. , Bacterial _ 10 mi 01. _ IIIspenaion (1.1xlo" cIu¡ml) ""re poured al !he baae 01_ pIant at 1 s 1 mo; 2 • 1 and 2 me; 3 • 1,2 and" me aftet planting. 2 Data _ ftom 3 repa 0130 pI¡'Iooa!Ion/cIone and _en!. Borde< planta ""re ellminatad. RY data Iollowed by !he same 1eIIer!s). ccrnpared _ ~ fraatrnanta, ... no! s1gnlticantiy dIffererrt al !he 0.05_ (OMRT). 1_6.12. EIfect of Pseudomonas """"""",no ($Itain P.t. esa) on commercial RY of fresh roots 012 oIon .. 1M 0012215 s III~; and M Col 1505 - -.mI plan1ed In a DIpIodía manillotfl>oandemlo area, wlth modera. CBB and _ Infection dutlng !he growlng ~. Source 01 SIakee Commercial pIMtati~' MCP_' Commerolal pIanIa1Ian ..... nfraal<10' Cassava-cassava..eassava 7.2a 70.4 5.6 22.8 1.1.10" ea-va-mal •• ¡sesame_va '6.38 b 57.8 5.2 18.8 5.8xlo' Cassava-malla¡cowpo_ 4.8b 60.3 5.1 19,9 2.4xlo' , Dala 1aken ffom 4 repajlreatrnent with 30 pI_. Date on ffash RY followod by dlflerant 10118". are slgnificanUy difieren! al 0.01 (OMRT). 1988); and (d) the long growing oyele of cassava and its vegetative propagation allow infection and dissemination of these parasites. Rndings were as follows: • Several fungal species were isOlated from intemal tissues of the epidermis, eolenchyma and parenchyma of stems of symptomless Iow-yielding plants of varlous clones. The fungal speeies isolated and their frequenoy were: Septoria nodurum, 52.1; Fusarium oxysporum, 7.2; CoIletotrlchum gloeosporíoídes, 5.8; C. graminicola, 4.3; Alternaría termissima, 2.9; Trichoderma sp., 2.9; Botrytis sp., 1.4; Toru/a sp., 1.4; Nlgrospora sp., 1.4 and others 20.4. Most of these specíes have been reported as endophytes in literature. "'noculations on plantlets and callus tissues by spraying, puncruring or immersion in a fungal suspension of some of the ¡solated fungi did not induce visible symptoms, but inoculated fungal specíes were reisolated from internal tissues near the inoculated points 30 days afterward. • Total root dry wt (tfha) of plots planted with stakes taken from commercíal fields and treated with a systemie fungicide were higher than for similar prots treated with a protectant fungicide or untreated controls (Table 6.15). However, differences were recorded among prots with plants derived from rooted shoots; this may be due to the existence of a higher probability of endophyte infections of stakes than shoots. • Histological studies observing fungal invasion ¡nto the host tissues have shown hyphal growth in coIenehyma and parechyma tissues of affected stems. 117 Table 6.15. Clone M Elr. 191 M Col 1468 M Col 2215 Aoo1 dry wt (l/ha) ot a plol planted wllh _as laken fJcm commaroial fialds and wllh pIanls óarivad from rootad shoots; bolh waro traated wllh oyslomic or prot_t fungicldas every 2 wk, 3 mo alter planling and for a 6-mo periodo Troatrnent Source 01 Benomyl Chlorolhalonil Control Planting Material (oyotomlc, 1500 ppm) (praleClanI, 1500 ppm) (llistilled HzO) Rooted sheet. 24.4' 19.7 17.1 St.ka. 25.5 15.0 14,7 Acolad shooto 20,8 7.6 6.7 Slakes 22.1 14.0 11.8 Acotad shoots 2U 15.6 12.4 Slake. 17.7 12.8 13.2 I Avg dala lakan fJcm 9 pI/plo!, 3 ropo; planto haNaslod al 10 me. The eff~ of 9 of the isolated endophytes was studied on 3 cassava dones (M Col 2215, M Bra 191 and M Col 1488) inoculated by spray, immersíon and puneturíng methods. Results were as follows (Fig. 6.1): • Both detrimental and beneficial endophytes affeet cassava, but most of the ¡solated spedes were detrímental. • The detrímental/beneficial effeet of sorne endophytes depended upon the inoculation method: on M Col 2215, Curvularía sp. was detrimental when spray-inoculated, but beneficial when inoculated by immersion or puncturing. ->Sorne fungal species behaved as endophytes in a given plant tissue but as pathogens in others: Rhlzoctonía sp. did not induce symptoms in leaf and stem tissues, but induced necrosis when the roots were mechanically wounded, similar to any root pathogen. • There were varíeta! differences in relation to the behavior of the endophytes on cassava, which cou!d be 01 importance for genotype selection. These findings stress the importance of selecting planting material from high-yielding plants of commercial plots and the need for genotype evaluations through several growing cydes under field conditions in order to eliminate susceptible clones showing low RY stabilíty. On the other hand, the existence of detrímental endophytes in cassava may partly explain the gradual degradation of cassava clones over continuous growing cydes. Further research Is needed to elucidate other ínteraetions between cassava endophytes and genotypes of the crop, as well as epidemiological features related to this group 01 parasites before control/preventíon systems can be devised. Special attention will be 118 100 80 60 40 ~ ~ o C!l 55 20 Ctt = Chaetomlum sp CI " CIIIdosporium sp cu • CurvulMIs sp a j o ~--lijlllllllllllll i~ IIIIIIIIII~ 11 u ~ 20 C!l E "C: ~ 40~ Illm~~IIIIIIIIIIIIInIT.HfIIIIIIII.;1 60 80 M COL 2215 Fu " FusIJlfum sp He " Hen/IersQnuIa sp Hu " Humlcola sp MBRA 191 Inoculatlon Method/Clone No. RII = Rhlzoctonla sp P1 = Pan/cl1/lum sp P2 = Pen/cll//um sp I li lñ ;.\. ~lttR t M COL 1468 Figure 6.1 Effect of 9 cassava endophytes(lnoculated by 8=spray. e=immersion. and .... =puncturing) on 3 cassava clones in relatlon to the percent of root welght 01 uninoculated controls. given to the beneficial endophytes in arder to increase biomass production or induce plant protection against detrimental parasites. 6.9 New Method for Interchanging Indexed Vegetative Planting Material 01 Cassava The exchange of vegetative planting material of cassava Is being done as MCP in test tubas. This material Is sublect to few quarantine restrictions; however, as the plantlets are weak, Iosses are common (up to 90%) and there are delays in obtaining adult plants (1 1/2 to 2 yr). To overcome these problems, the following aystem was developed: 3O-day- old shoots from indexed mother plants (MCP previously tested for sanitary status) ar.e cut into 1Q-cm sections and placed in a container with deionized H20 to prevent dehydration. Shoot stakes are then removed from H.O and dusted with Thiuram (mercuric mercurous chloride); each end (1-2 cm) ia wrapped with parafin. Shoot stakes are then organized in bundles of S, and their bottom half is wrapped with paper towels previously sprayed with a benomyl-captan (3 g/It each) suspension, covered with a plastic bag and sealed with rubber bands. Bundles are then arranged in cardboard boxes, with several holes to a1low air exchange during shipping. Shoot stakes have a shelf Jife of more than 15 days. At the recipient location, the stakes are plantad in pots with sterUe soil and covered with plastic bags with 4 to 6 holes to maintain a high RH and permit air exchange. Soil is kept moist but not saturated. When shoots hava moted and buds germinated (10-12 days), the plastic bags are removed. Two wk latar the plantlets are planted In the fleld. Using this syatem it is possible to obtain 90 to 1ClCl% establishment 20 days after packing and adult plants 7 to 10 mo later. If mother plants ara virus-indexed and maintained free from insect vectors of cassava viruses, quarantine risks can be minimized. This system was included in the FAO/IBPGR Technicai Guidelines for tha Safe Movement of Cassava Germplasm and several countries have accepted vegetativa planting material using this system. 6.10 Storage of Vegetative Planting Material Research on storage of vegetative planting material of cassava has identified the following factors that affect quality during storage: 6.10.1 Dehydration This factor can decrease establishment when stakes are stored at RHs below 80%. Varietal differences in resistance have been found, but there do not appear to be sharp differences in the speed of H.O 105S from 5take5 stored at low RH. The highest percent of H.O IOS5, which occurs during the tirst days, is related to the ambient RH at the storage site. 120 6.10.2 Ught Ught intensity showed significant effects on establishment aftar long periods (4 mol of storage of 1-m long stakes: at high Iight intensitles (13,400 FC) establishment was asfor the controls. $tarro content In stekes decreases over time, Independent of Iight intensity. Results a1so showed the existence of varietal differences in sterro Ioss, but the relationship between sterro content and establishment was not clearly shown. data suggesting a varietal X starch level interaction. 6.10.3 Chemlcal treatment Chemical treatment of stakes prior to storing is essential for eliminating both insects and pathogens and preventing reinfestations during storage. 6.10.4 Effect of atorage perlod and IIght Intensity on RV Generally RY ls affected when uslng stored stakes lor planting, RY decreasing as the storage perlod Increases. Consequently, stakes should be stored only when necessary. A storage perlod of more than 2 mo can induce severe RY decreases in highly sensitive clones. Hlghly resistant clones exist, and this character appears to be genetically controlled. The following storage system is advisable lor tropical environments: • Starns should be selected from visually healthy plants of l1nbranched clones showing satisfactory Iavels of resistance to storage (>80% establishment after 60 days); mother pIants should be those that yielded most. . • Stems should be 1.2Om long, taken from the first half of 8- to 11 mo-old mother plants. Stakes should be arranged in bundles (approx. 10 stakes) and treated with a fungicide- pesticide mixture (benomyl, 3 g/lt; maneb. 3 g/lt; and malathlon, 1-2 ce/lt). • Bundles should be stored vertically in the open field by planting the first 5-10 cm of the stakes In the ground and watering for the first 2 wk . • 'fhe first 10 cm from eaeh end, as well as shoots produced durlng the storage perlod, should be removed. Stakes should be 15-20 cm long and treated agaln with the fungicide-pesticide mixture before planting. Fertilization (based on soU or plant tissue analyses) should be implemented at 45 days. This system was implemented in plots planted with stakes stored lor 4 mo under very dry conditions (PiviJay. North Coast, 28°C avg; 60% AH). RYs were similar to those obtained on plots planted with unstored controls (Table 6.16). This system is now being validated 121 Tablo 6.16. RY. of piola planted witll stakes 1 01 2 clone. from Media !.Una followlng a storage oystem 1 ler tropical environmonts. Trea1ment Unsforod control SIoraga In opon fIeld M Col 2215 Unlettillz.d 16.3 15.2 Fr.oh RY (t/haIlClone fertilizad 20.2 17.9 Unfertllized 14.5 15.5 M CoI2ll15 Fertilizad 20.8 20.8 1 Slakes tsken Imm firsl hall of .tem. 01 ll-mo-old plants; long {1.2D m} stem stak .. 8IOtad 4 me {Jan . ..f.Iar •• tIle driost poriod of tIle ragion}. . in 4 different locations on the North Coast and has been adopted by several progressive cassava farmers at Media Luna. 6.11 Geographlc Distributlon and PotentiaJ Risk 01 Six Cass8va Diseases Extrapolating data on epidemiological studies, surveys on diseasa saverity and climatology (with tha cooperation of the ASU), the geographic distribution and potential risks of CSS, SED, WS, Fusarium and Phytophthora root rots, and Diplodla stem and root rot are presentad for Latin American (Fig. 6.2). Tha data show the following: . • The highest potential risk for CSS and SED exists In areas with moderate tamp during the wat season (20-26.5°C), mora than about 1200 mm .rainfall¡yr, and prolongad periods of high RH (for SED), or where day ¡night temp fluctuate more than 8° C during the wet months (for CSS). -The highast risk for Fusarium stem and root rot pathogens exists in areas with high temp (>25°C) and frequent haavy rainfalls during short periods of the rainy season. For Diplodia stem and root rot pathogens temp >26°C, annual rainfall < 1000 mm and dry saason >4 mo. H tha temp is more than 20°C and the ¡and Is periodically floodad during the rainy season or badly drained, Phytophthora spp. can also be of great importanca . • The mycoplasm-induced WS is characteristically found In areas where the temp rangas from 20 to 24°C tor at least 3 mo¡yr. Symptoms are modarate to mi/d, and may disappaar as temp increases. Consequently, WB can be a thraat in areas where cool temp occur during the year or for mora than a 3-mo periodo This information can be usad to selact appropriate clones in the different geographic areas of Latin Amarica, as well as guide breeding programs for different ecological areas and defining strategies for controlling these diseases. 122 Figure 6.2. Geographic dlstribution and potential risk of slx cassava dlseases In latin America. 123 7. VIROLOGY Tha VRU is responsible for investigating the diseases of cassava causad by virusas or viruslika agents. The development of a methodology to assure the movement of virus-free cassava germplasm is a research priority. This is especially importantbecause there are diseasas of cassava of unknown etiology. The control of viral diseases raquires eithar the identification of resistant germplasm or the implementation 01 cultural practicas that mitigate disease lossas. Most viral disaasas are controllable with the currant technology; and continued developmant of rapid diagnostic techniquas, together with the deployment of resistant germplasm, should further reduce the losses caused by viruses. 7.1 Frogskin and Carlbbean Masale Dlseasas Frogskin diseasa (FSO) is a viruslika diseasa of unknown etiology that was tirst reportad in 1971 from southern Colombia. Thare áre no leaf symptoms produead in most cassava clones affected by FSO, but there can be very severe root symptoms. The malady was named frogskin disease because the root periderm and corky layers anlarga to form characteristic raisad lip-shapad tissures on tha roots. Tha roots often show a zone of constriction where these fissures ara prominent. In saverely affected plants. the roots do not fill with starclí, and there are yiald losses of more than 90% in areas where the disease is endemic. A few cassava clones devalop mosaie symptoms on the lsavas and growth Is stunted. C8ribbean mosaJc disease (CMO) Is a viruslike diseasa of unknown etlology, reported in 1981 from the northern coastal regions of Colombia. Susceptible cassava varietíes affected with eMO show foliar mosaie symptoms and significant yield loss. Yield losses in susceptible cassava clones can be severe, but tolerant clones produce normal yields. These two diseases have been reported as different because the root symptoms associated with FSO are either absent or mild in plants affected with CMO. While the root symptoms are apparently different, these diseases share many similarities. 80th disease agents are transmitted by grafting and tha cassava varo Sacundina can be used in indexing programs te detect both CMO and FSO. The mosaic symptoms on the Ieavas of Secundina are expressed most prominently when the plants are kept in an area where the max. temp is kept below 30°C. Constant temp (> 28°C) suppresses the foliar symptoms of both diseases. Neither FSO nor CMO can be inoculated mechanically, and the only known hast for both diseases is cassava. Insect vectors have been suspected for both diseases as they spread rapidly in the tield. 125 7.1.1 Identlflcatlon of a phytoreovlrusllke agent associated wlth FSD and CMD Isometric viruslike partieles (70-80 nm in diameter) are found in thin sections of leaves, petiolas, stems and roots of CMD-affected plants. Similar virusllke particles are also found in the same tissues of plants affected with FSD. These partielas have been found in planta affected with all the FSD and CMD isolatas that have been tasted (Rg. 7.1A). Viroplasmlike bodies have also been found in FSD- and CMO-affected planta; they are often assooiated with the chloroplasts of the cells (Rg. 7.18). Double-stranded RNAs were purified from cassava plants infectad with either CMO or FSD and run on both agarose and polyacrylamide gels. On!he former, there appear te be 3 or 4 bands; bol on the latter, there are 9 bands, which are consistently prasent in plants affeoted with FSO o( CMO (Fig. 7.2). The ds-RNA segments are estimated to be 4000, 3800, 3400, 2600, 1900, 1800, 1700, 1100 and 1000 bases in length (FIQ. 7.3). Radioisotopic-Iabeled cONA probes were prepared from the isolated ds-RNA for use in hybridization analyses to determine the relatedness of the vanous FSD and CMO isolates. Based on the limited studias to date, there appears to be a fairly high degree of similarity between the isolates. For example, FSO isolate 14 hybridizes with CMO isolate 80 (FIg. 7.2). This is further evidence that ds-RNAs associated with FSD and CMO are either identical or closely related. The partial purification of the phytoreoviruslike particles has been attempted. but they are very labile. The particles have been banded on a cesium sulfate gradient, and then visualized with a transmission electron microscope (TEM). Soma particles are complete, bol tha maJority of them apPear to have degraded into a 5O-nm structure typioal of the cora vinons of phytoreoviruses. To determina whether the Isolated particIes contained the ds-RNAs associated with CMO and FSO, a dot blot assay was performed uslng a radioisotopically labaled cONA probe prepared from the ds-RNA. The dot blots were positiva, indicating that tha fraction containing the viruslike partielas also containad the ds-RNA. This Is evidence that thase phytoreoviruslike particlas contain a ds-RNA genome. The viruslike partielas and the ds-RNA bands present in cassava affected with FSD or CMD are similar to those reported for in phytoreovirusas. Neither FSD nor CMO has been mechanically transmitted despite many attempts; this is conslstent with the causal agent being a phytoreovirus as these viruses are not mechanically transmitted. 7.1.2 Vector transmlssion experiments Whiteflies have long been suspected as the vectors of FSO. For most of the transmíssíon experiments to date, the indícator done Secundina has been usad. The mosaie symptoms in the leaves were used as the marker to determine successful transmission. Table 7.1 is a list of the FSD and CMO isolates tor which transmission by 8emisia 126 Figure 7.1. A: B: A group of vlrusllke particles found in a thln sectlon of a leaf of the cassava clone Secundina affected with the mosalc symptoms associated wIIh FSD. Viroplasmlike bodíes found in a similar seclion as lhe viruslike particles; V indicates lhe vlroplasmllke bodies. 127 Figure 7.2. Figure 7.3. (1) An agarose gel showlng the dsRNA segments extractad from plants infectad with FSD aOO CMD; (2) blot hybridization using a flrst-straOO cONA proba preparad from FSD-dsRNA (lsoIate 14). A = dsONA markers; B = FSD isotate 14; C = FSD IsoIate 24; O ~ CMD IsoIate 80. A poIyacrylamide gel showlng the dsRNA segments lsoIatad from pIants infectad with FSD or CMD. A: FSD ¡solate 24, B: CMO isoIate 5. C: FSD isolate ToIima, o: FSD isolate 29. M: dsONA markers. 128 Table 7.1 USI 01 isoI_ or FSO and CMO lila! have been trenomllled \O oassava by 111. whltefly B. tuberoulata. Source Receptto Date Di ........ isolate Plan! Plan! Roaulls WF-mo ... lc 3 WF from fleld' Secundloa 1986 FSD ;¡g Secundlna Secundlna 1990 CMD 80 Secundloa Secundlna 1990 CMD 86 Secundlna Secundlna 1990 FSO 24. M Col 72 Secundloa 1991 FSD ToRma Secundlna Secundlna 1991 , Thl. Iso/ate OfIglnated from whlleflles coUOChId In 1ha flaJd; lata' ""porlmenls ahowed R waa _llIad by B. tubemulata. tuberoulata has been shown. A one-day acqulsitlon penod gives the highest rates of transmission. There ls a latent perlod during which the virus ls not transmitted by the vector. The mlnimum perlod of lnoculation that has resulted in transmission is 3 days. Additlonal experlments must be done to characterlze more accurately the transmission of the FSD agent, but preliminary results indicate that the agent is transmitted persistently, presumably in a circulative manner. There ls no evidence at this time that the agent multiplies within the vector. , Plants that devaloped mosalcsymptoms wera transplanted in soll ¡nsida a scraanhouse. The plants were grown for 6 to 8 mo, and then the roots were harvested and inspected for symptoms of FSD. Plants infected with FSD ¡solate 29 and CMD isolates 80 and 86 all showed mild but dlstlnct FSD symptoms on the roots (Fig. 7.4). The plants that developed the mosaie symptoms in the transmission tests were analyzed for the presence of ds-RNA species. In most cases, both the mother plants used as the souree of inocula, and the plants infected in the transmisSion experiments had similar ds-RNA patterns. The exception has been FSD isolate 29, which loses some of its ds-RNA bands during the transmission experiments. Only the 9 bands consistently associated with FSD and CMD are present in the plants infected during transmission experiments either with lsolate 29 or the others. Thls suggests that FSD isolate 29 may be co-existing with a second virus. 129 Figure 7.4. Roots showlng mild symptoms typlcaI of FSD. These roots were haIvesIed from plants infected with the phytorOOlllruslike agent using B. tubercu/ata. as the vector. Plants were grown In a screenhouse to prevent other SOUIC8S cf infection. 130 Viruslike partícles 80 nm in diameter, similar to those found in FSD- and CMO-infected plants, have been found in the newly infected plants from all isolates tested. Viruslike particles have been found directly in B. tuberculata individuals fed on infected plants; none was found in B. tuberculata individuals fed on healthy plants. 7.1.3 Control strategies The damage caused by FSD can be limited by using clean stakes. In heavily infested fields, the old crop should be removed for at least a month and then the field should be planted with clean stake material. Use of the indicator clone Sacundina can facilitate the selection of clean stakes. As the disease is not mechanically transmitted, no special care is needed when handling stakes. The key for sustainable yields is the continued selection of clean planting material. There are no leat symptoms in most cassava clones; therefore, the selection of the planting material must be made at the time of harvest. The roots must be ínspected carefully for symptoms of FSD. As many of the root symptoms are mild, care must be taken to recognize them and elimínate stakes from these plants in order to mitigate the losses causad by FSD. 7.1.4 Quarantlne Impllcatlons and diagnastlc tests FSD ís the most serious viral disease that is endemic in Colombia. While progress has been made on idantifying a phytoreovirusliké agent associated wíth the disease, there are still some unanswered questions concerning its etiology. Therefore, extra care must be takento assure that the germplasm exported from CIAT is free of FSD agents. Currently, two tests are being usad to detect FSD. The first method. in use for many years. (s to graft ttle test plant to the indicator clone Secundina. Tha second methoo is to extract ds-RNA from the plants and confirm the presence of FSD either in a polyacrylamide gel or with a dot blot assay and hybridization probe. If either test is positive, the clone is assumed to be infected. Only clones that have no history of FSD will be considered for exporto 7.2 Cassava Common Masale Virus Cassava common mosaic virus (CCMV) is a member of the potexvirus group found throughout Tropical America from Mexica 10 Paraguay. The partícle morphology is a semiflexuous roo approx. 15 X 495 nm. The viral partlcles cantal n a single coat protein with a relative molecular wt (Mr) of 25,000, and a ss-ANA genome of approx. 6400 bases. Nuclear inclusions typical of the potexvirus group can be found in cassava and Nicotiana benthamiana. Cassava plants infected with the virus have mosale symptoms on the leaves and may suffer yield losses of more than 20%. In terms of total yield loss, CCMV is considered to be the most destructive viral disease in cassava in the Americas. 131 7.2.1 Molecular characterlzation TIle sequeneing of CCMV genome is nearly complete. TIle virus (6400 bases in length) is most closely related to potato virus X (PVX). CCMV has a genomie organízation typical of other members of the potexviruses, whose genomie sequence,is known. Most of tha sequeneing of CCMV was done at the VRU at CIAT.-perhaps tha first plant virus to be sequenced in Latin Amariea. TIle cONA eloníng and sequencing of CCMV coat proteín was an essential element for the Cassava-Trans projact, based atWashington U. TIle coat protein gene ofCCMVwas successfully introduced into Nicotiana benthamiana, and thase plants show almost complete immunity to CCMV. Coat protein-madiated cross protection will be one method fer developing resistant germplasm to CCMV. TIle major technical limitatíon is the transformation of eassava. 7.2.2 Quarantine implications and control. Antisera against CCMV are available, and the virus ís readiiy detected by ELISA. AII germplasm from the Amerieas is routinely screened for the presence of CCMV befere exportation from CIAT. TIlere are no reports of a vector far CCMV, and the leaf symptoms eausad by CCMV are very distinct. The disease is controlled by selecting plants without foliar symptoms and using these stakes for propagation. If the area is heavily infested, cuttíng tools should be disinfected between plants. 7.3 Cassava Veln MosaJc Virus Cassava vein mosaic virus (CVMV) is a member of the caulimovirus group, which has isometrie virions approximately 50 nm in diameter and a ds-ONA genome of approx. 8000 bases. The only known host for CVMV is eassava, and the vector is unknown. Symptoms inelude chlorosis of the veins, which ean either appear as a chevron pattern or coalesee to form a ring-spot pattern. Some leaves also show a mosaie pattern over the entire leaf. There is often leaf distortion and young leaves somatimes show epinasty. Symptoms are variable, being exprassed batter at high temp. The virus is reportad in many states of Brazif, being most prevalent in the NE, espacially in the hot semiarid zones where it is not unusual to find more than 50% of the pfants infected. CVMV is not known to oceur outside Brazil, and there is very little information on yield losses. 7.3.2 Quarantine implicatlons and dlagnostic tests Initial eharacterization of CVMV was done by Ors. E. Kitajima and A. Costa (Brazif). An antiserum to the virus was mada by Dr. M. Un around 1980, but it is not known if it is still viable. Infected pfants have distinctive symptoms. that ean be confirmed by visualízation 132 ot SO-55 mM isometric particles using TEM. Symptoms may not always be present; and as the virus can intect in vitro cultured plantlets, a rapid and reliable diagnostic test is needed. DNA clones to CVMV have been prepared at the U. 01 Kentucky, and one has been provided to fue VRU. A polymerase chain reaction (PCR) test for rapid detection of CVMV should soon be available at CIAT. There is a need to prepare a new antiserum or transfer a PCR detection method to EMBRAPA/CNPMF to facUitate screening of cassava germplasm for CVMV in Brazil. 7.4 Afrlean and Indlan Cassava Mosale Viruses African cassava mosaic geminivirus (ACMV), found throughout tropical Africa, and a similar virus, Indian cassava mosaic geminivirus (ICMV), found throughout India and Sri Lanka, cause the most destructive viral diseases of cassava. Symptoms inelude mosaic, yellowing, distorted leaves and stunted growth. The vector 01 these viruses is the whitefly, B. tabeei" Resistant varieties exist, and in these plants symptom expression is erratic. It is presumed that distribution of the virus is restricted in these varieties. 7.4.1 Quarantlne implicatlons Recently ACMV was introduced to the island Praia, Cape Verde. Losses were nearly 100% as the germplasm (of Brazilian origin) was not resistant to the virus .. Similar losses can be expected if ACMV and an active vector are introduced into tropical America. In Asia these geminiviruses have not been reported outside India and Sri Lanka; therefore, much of Asia faces a similar threat. In vitro cultures 01 cassava received at CIAT from lITA (Nigeria) have all been tested lor the presence 01 these geminiviruses at fue Scottish Crop Research Institute (SCRI). Monoclonal antisera to these viruses have been produced at the SCR!. There are antisera and cDNA probes to ACMV available at CIAT that can be used to detect these geminiviruses. While B. tabae; is common throughout the Americas, it does not normally colonize cassava. Recently a biotype of B. tabaei, which has beco me dominant in Florida, Puerto Rico and the Dominican Republie, has colonized cassava. As this biotype is a possible vector of ACMV, there is a need to monitor its distribution. 7.4.2 Control strategles The use ot resistant clones is the most effective method ot mitigating losses caused by these geminiviruses. Resistant clones exist in Atrica and in India. liTA has shipped clones with resistance to ACMV to CIAT, but these have severe agronomic and disease problems when grown in Colombia. The clones have been used in a breeding program as one source of resistance. Another means of identifying resistant germplasm is through the joint CIAT-IITA program (see Chap. 23). A third possible source of resistance could be obtained through the Cassava-Trans project, which is working on coat protein- mediated cross protection for both CCMV and ACMV. With the spread of the new 133 biotype of B. tabaei, the threat introducing ACMV into tropical America has increased; therefore, additional emphasis will be placed on identifying ACMV-resistant germplasm adaptee! to tropical America. 7.5 Latent Vlruses of Cassava in Latln Amerlea There are several latent viruses known to infect cassava in Latin Amenea. These latent viruses were discovered while working on other virus or viruslike pathogens of cassava. None of these viruses is known to cause disease in cassava or to have any effect on RY. There is only limited information on the distribution of these viruses. CIAT is committed to sending germplasm that Is frae of pathogens; therefora, the primary concem of the Investigations on these viruses is to find methods of datection that make it possible to certify that germplasm daes not contain these latent virus es. 7.5.1 Cassava X and cassava Colombian symptomless virus es Cassava X (CsXV) and cassava Colombian symptomless (CCSpV) viruses are both potexviruses discovered during attempts to identify the causal agent of FSD or CMD. These viruses were discovered because they are mechanically transmitted to diagnostic hosts. Subsequent tests have shown that they are not present in most of the plants affected with FSD or CMD. Neither virus Is known to cause symptoms or disease in cassava. either alone or in combination with other viruses. With respect to quarantine implications and diagnostic tests, there are antisera to both viruses. Allin vitro germplasm that originates from Colombia is checked for tila presence of these viruses by ElISA before shipping. 7.5.2 Cassava American latent virus Cassava American latent virus (CAL V). a member of the nepovirus g~oup. was discovered by Dr. B. Walters. who isolatad it from cassava also infected with CCMV. The virus was isolated from samples from Guyana and Brazil (Manaus). While the vector of the virus is unknown. it is probably a nematode. It Is a1so possible that this virus is seed' transmitted. As no cassava plants infected with this virus have been found in tha field. it is difficult to test whether the virus is sead and/or nematode transmitted. As for quarantine implications and diagnostic tests. Dr. Walters has made an antiserum for CALV available. This antiserum. which has been used to test cassava in Colombia and Brazil. has been useful in certifying seed lots as frea of CALVo Thousands of seeds and hundreds of plants grown at CIAT HQ, were assayed tor CALV, aII of which were negativa. CALV does not appear to be present in the Cauca Valley; therefore, the further testing of materíals grown at CIAT is not warranted. Of approx. 200 plants (18%) of the germplasm collection at EM BRAPA/CNPMF at Cruz das Almas tested for CALV. a11 testee! negativa. CALV does not appear to be present at the CNPMF farm in Cruz das Almas. 134 Some additional testing may be done on germplasm being developed for the semiarid and subtropical regions (see Chapo 2. seco 2.4). The distribution of this virus is not known. and the only reported source of the virus is from the humid tropical regions of the Amazon. The virus has not been found at the maln sites where true cassava seeds are produced for export; thus the risk that this virus is contaminating these cassava seed lots is extremely low. 7.6 Concluslons and Future Research Objectlves There are 4 main diseases causad by virus or viruslike agents. These are FSD, CVMV. CCMV and ACMV (ICMV). While the losses causad by these diseases can be mitigated through cultural practicas. there is a need to continue efforts to identify and distribute cassava clones that are tolerant or resistant to these diseases. Researcn on FSD continues. focusing on the confirmation of the association of the phytoreoviruslike agents and the entire complex of disease symptoms. Hybridization . assays have been developed tp detect the ds-RNAs associated with the disease, making it possible to sereen cassava germplasm more effectively. Now that the assay is available. more emphasis will be placed on looking far resistance. Additional research is also needed in NE Brazil to determine the losses caused by CVMV. The vector of this virus needs to be determined to understand Its epidemiology. Control of this virus should be a part of the integrated pest management strategies developed tor this area of Brazil. There are excallent datection and control measures for CCMV. Sorne additional research Is needed lO identify resistant germplasm. but directly in those areas where the disease Is causing losses. Screening CCMV-resistant material from the Cassava-Trans project can be done at CIAT. The grealest need for ACMV Is to identify additional resistant germplasm. As resistant germplasm Is identífied in the CIAT /IITA collaboratlve project. it needs to be transferred back to CIAT so that there are adequate sourcas of ACMV resistance in germplasm adapted to tropical America. A similar project is needed for those areas of tropical Asia in which the virus Is not presento Beside !he major diseases. there are saven other known viruses that Infect cassava. Although most of these viruses do not appear to cause disease, !hey are of quarantine significance because of their limited distribution. Diagnostic methods haya been developed for both the viruses that cause diseases and the symptomless ones. These dlagnostic methods help assura the safe movement of cassava germplasm. and a research objective will be to develop even more sensitive detection mathods. 135 8. ENTOMOLOGY ANO ACAROLOGY Early research in cassava entomology and acarology defined the arthropod complax in cassava and identified key and potential pests. A large complex of pests was found, aspecially in seasonally dry lowland areas of the Neotropics. Extensiva efforts far developing control strategies were made for species associatad with significant yiald Iosses. Pasts that attack the crop over a long period of time (3-6 mol cause the greatest losses. These inelude mites, thrips, mealybugs, lacebugs, whiteflies and the burrowing bug. Some species such as shootflies, fruitflies, scale ¡nseets, gall midges, termites, leafcutter ants and stemborers occur sporadically, causing IMle or no yield reduction . . Cassava is produced primarily by small-scale, resource-poor farmers, and pesticide use is limited or negllgible. In areas where cassava production is under intensífication (e.g., Colombian North Coast, Coastal Ecuador, NE Brazil). however. increased agrochemical input as a means of Increasing yields may eventually result, particularly in seasonally dry areas where arthropod pest problems are severest. In crops such as cotton, potatoes and rice, dependence on pesticide usa with prophylactic applications has aften led to the development of resistance to pesticides. negative environmental effects, and aven to abandonment of the crop when the cost. of pest control exceeded retums trom production. Integrated pest management (IPM) is often implemented at this paint in order to reverse this process. The obJective of the Cassava Entomology and Acarology Saetion is to implement ecologically sound crop proteetion practices while cassava produetion is in the intensification stege, before the use of pasticldes becomes significant. Accordingly, research has facused on host plant resistance (HPR), biological control and cultural control practicas. Casseva Entomology and Acarology is a dynamic program that responds to changing needs in crop proteetion. Considerable basic ¡nformation and improved technology components have been generated for control of pests such as mites, mealybugs and hornworms. Although substantial research is still needad, technology based on HPR and biological control far mites and mealybugs, and augmentative biological control of the homworm are in the implementation stage. A major proJeet for testing and refining this tachnology with farmars, and for training farmers and crop protactionists in eco!ogicafly sound pest and diseasa management has been developed for NE Brazil in collaboration with the Pathology saetion. As part of the same projeet, lITA willlead a similar effort in West Atrica. UNOP funding is being sought far this undertaking. Research on pests such as whiteflias, burrowing bugs and lacebugs is more recent; and considerable basic study is needed before sound technology can be recommended. 137 Some sources of HPR to whiteflies and lacebugs have been identified; however, these need to be evaluated against the various species found in the different cassava ECZs of the Neotropics. Research on pests of dried storad cassava, true seed and wild Manihot species has been initiated recently to meet the future demands set out in the Cassava Program strategy for the 19905. In addition the potential 01 wild Manihot species as sources of HPR to such pests es hornworm, lacebugs, burrowing bugs and certain whitefly species (Bemlsia tabacl) will be evaluated. 8.1 Mltes The Cassava Green Mlte (CGM, Mononychellus úmajoa), a major pest in some areas 01 the Neotropics, wes accidentally introduced to Africa from the Americas in the 1970s. Control 01 CGM in Africa ia one of the most Important challenges facing crop protectíon today. CGM is a serious pest in the most important cassava-growing area 01 the continent--NE Brazll. A sibling species, M. caribbeanae, is a pest in subhumid areas of Venezuela, Colombia, Ecuador and the Caribbean basin. This report summarizes CIATs contribution to IITA's classical biological control effort far Amca, describes the agroecological basis of the CGM problem in NE Brazll, and outlines a strategy for its solution. 8.1.1 Exploratlon for natural enemles of CGM In the Neotropics In order to implement classical biological control of CGM in Amca, a collaborative, multinstitutional effort was organized by lITA. CIAT and EMBRAPA were engagad to explore for predatory phytoseiid mites, considered to be the most important natural enemies 01 CGM in South Americe; while liTA implementad the release and follow-up campaign in Africe. Other institutions including universities In developed countries and several Atrican national programs were also involved in the effort. Explorations for phytoseiid natural enemies 01 CGM were conducted by CIATfrom 1983- 90 in most cassava-growing countries 01 the Neotropics. CIAT conducted extensive surveys in Colombia, Venezuela and Ecuador, and smaller scale surveys in NE Brazll, Trinidad & Tobago, Guyana, Peru, Paraguay, Mexico, Cuba, Panama and Nicaragua, visiting 1261 cassava tields. In NE Brazll, a parallel exploration effort was conducted by EMBRAPA from 1988-90 covering 427 cassava fields. Explorations by both institutions involved qualitative and quantitative avaluation 01 CGM, other tetranychid mites, and natural enemy populations on cassava, and in neighboring vegetation. Agroecological criteria were usad to prioritize the exploration efforts. High priority was· given to seasonally dry and semiarid lowlands; humid lowland sites were also includad in the surveys. 138 8.1.1.1 GeoaraQhic patterns of distributioo. host plaot raoge aod abuodaoce of CGM aod related species io the Neotropics. Four species of Mononychel/us were fouod in cassava. The most geographically widespread was M. caribbeanae, which occurred throughout Central aod South Ameries in all countries surveyed except Brazil, Peru and Paraguay. M. mcgregorl wes found in the inter-Andean valleys 01 Colombia, Ecuador and Peru aod in the Colombian and Peruvian Arnazon Basio. M.tanajoa was found io Panama, Colombia, Venezuela, Guyana, Trinidad, Ecuador, Paraguay and Brazil. M. planki was detected in 5 fields in Colombia and in one of 52 fields surveyed by CIAT in Brazil. In na samples taken from vegetation adJacent to cassava fields, M. mcgregorí, M. caribbeanae and M. plankl were found io 11, 8 and 1 plant species other than cassava (M. esculenta) resp. Although reported from wild Manihot spp. in Brazil, M. tanajoa was found only once on a host plaot other than Manihot spp. On Paoama). Avg M. tanajoa 00./lea1 were significantly higher in NE Brazil than in Colombia. Wlthin NE Brazil significantly higher deosities (P=.10) were found in areas with 385-700 mm rainfall/yr than io other raiofall zonas; whereas in Colombia M. tanajoa was found in ooly 1 of the 15 sites sampled with 365-700 mm rainfall/yr. The sibling species M. carlbbeanae was found in 4 of these sitas. Higher population densities of M. caribbeanse (X = 127/leaf; n = 34) than M. taosJoa (X = 88/leaf; 0=34) were found in surveys of semiarid and seasonally dry areas of Venezuela; and M. caribbeanae was the ooly species found in seasonally dry and semiarid areas of Ecuador, suggesting this species is well-adapted to subhumid conditions. 8.1.1.2 Geographic distribution of ghytoseiid natura! anemies of CGM in the NeotroOics. The phytoseiid complex on cassava in the Neotropics attained max. diversity in Colombia, where 40 species were identifiad. The composition 01 the complax in Colombia varied regionally (CIAT Annual Report, 1990). Amblyseíus limonicus sensu lato was the dominant species in all areas surveyed, except in the seasonally dry to semiarid Guajira, whera M. tanaJoa was replaced by M. caribbeanae as the dominant species of Mononychellus andA idaeus was the dominant pbytoseiid. In other seasonally dry areas of Colombia, M. tanajoa was the predominant species 01 Mononychellus. Although considerable regional vanation in the phytoseiid complex was observed within the seasonally dry zone, 3 species (A. Ilmonlcus s.I., A rapax, A dentills) occurred consistently (CIAT Annual Report, 1990). In Colombia, Venezuela and Ecuador up to 12 phytoseiid species were found per cassava field, and 29% of the fields contained 3 or more species. In Brazil, 22 phytoseiíd species were reportad on cassava (Source: EMBRAPA); however, only 2 ofthem were common. Only 3% of the fields contaíned 3 or more species, 56% contained only one species, and 28% were devoid of phytoseiids (CIAT Annual Report, 1990). The no. of dry mo/yr was a useful agroecological criterion for separating Neotropical phytoseiid species ecologically. Amblyseius chiapensis, A peregrlnus and Euseius 139 Species C~.P[NS COflCOROI P(RI:CRIH H(LvtOLlJ 8(LlOTTl ANONi'MUS UWQNtCIJ A~II'O A[RI"'US ZIJ"LUAC.4.1 CANNA(HS O[NllUS AN1i(CTEN ¡,;..cROPIL AAfAlt lOAruS ftNUISCU "O o I----{Ol---..... I----{O)---..... >-----O---i I------ H +' e 11 e o Q. E o O Figure 8.4. 1.' S ¡ 0.0 '.1 4A lA - •• 8- B -1..3 f a .• l.? -. t •• ... §J. -, Componant 11 •• 'Te 1.' Componant 1 Posltlon of 5 A /imonicus s./. raees on 3 principal component axes. Componen! 111 Is no! representad. (1 ~ BrazH. Trinidad; 2A ~ BetulIa. Baranoa. La Paz. Ciénaga de Oro; 2B.. Alfona. eerete. PMJay, Las Flores; 3= Villavlcenclo; 4A = Santander. Calcedonia. CIAT; 4B = Vijes; 5 = Venezuela). backcrosses, tha % of eggs inviable was consistently Iower, and egg-to-adult survival was consistently higher whan tha male rather tIlan the female parent was an E, hybrid. The differences io raproductive parameters observad batwaao tila homogamic and hybrid crosses corroborata the existence of races deducad from tIle electrophoretic study. The status of A Ilmonlcus s./. as a true species distioct from tIlat of A IImanicuB, originally dascribad from avocado and citrus in California, was coofirmad through crosslng experiments aod elactrophoretic analysis. 8.1.3.2 Euogal pathogeos. Naozygites sp. ls pathogenie to all mobila stages of CGM. No avidaoee was found for traosovarial transmission of the disease. lo tests involving exposure of A limonlcus s./. to infected CGM, 00 evidence was fouod for patllogeolcity of tIlis fungus to phytoseiids. The disease cyc!e of tIle fungus was determined in CGM. Conidiogenesis does not occur at RHs below 65%; and tha formation of anadhesive conidia, which are rasponsible for propagating tIle iofection, is inhibited at RHs above 65%. The time batween inoculation and appearance of symptoms was lnversely related 145 to temp in tha range of 2O-32"C; howevar, the infection rata of exposed individuals peakad between 24 and 28"C. Efforts to culture Neozygites sp. on artificial madia involva extracting and transferring the hamocoel of live infected individuals to culture media. Thus far thasa haya been unsuccessful; howavar, a mathod for provoking conidiogenesis directly onto culture media is under development. Cultures based on conidia obviate the extraction and transfer of hyphal bodies from the hemocoel-the steps associated with contamination. 8.1.4 Estimation of field impact 01 natural enemies 8.1.4.1 Phytoseiids. Several species of tetranychids may be present simultaneously on cassava in many areas of the Neotropics (CIAT Annual Report, 1990). Differences in development and fecundity have been demonstrated as a consequence of providing phytoseiids with diets composed of different tetranychid or nonacarine prey; however, ability to utilize a given prey type under lab conditions does not guarantee it will be consumed under field conditions. Electrophoretic analysis of gut contents of field- collected A. limanicus s./. confirmed that this specíes preys on M. tanajoa, M. caríbbeanae and M. mcgregorl. Oídium man/botís, a fungus hypothesized as an alternative food source for A. IImanicus s.l. duringperiods of lowecarine prey densities, was not identified in the gut contents of 143 A. limonicus s.l. ana!yzed from 8 field sites in Colombia and Venezuela; neither were tetranychid species commonly associatad with cassava (Ollgonychus gossyplf, O. peruvlanus and T. tumldus). Analysls of 64 A. idaeus suggests that M. car/bbeanae is the principal tetranychid prey consumed by this species in Colombia and Venezuela: 70% had consumad M. caribbeanae; 10%, M. tanajoa; and in the remaining 20%, no prey were detectable. In the Guajira (Fonseca), M. caribbeanae, M. taneJoa, and O. peruvianus comprised 60, 31 and 9% of the tetranychid populatión resp. A. ¡deeus, Typhlodromus annectens and Euseius alatus comprised 83, 8 and 5% of the phytoseíid population, resp. The protective effect of phytoseiids on yield was estimated by chemical elimination. Yields 8 mo after planting averaged 8.5 t/ha in plots where phytoseiids were eliminated compared to 18.3 tfha in plots where they were presento Infestation intensity 01 Monanychellus spp. in plots with and without predators averaged 48,841 and 74,851 mite-days{leaf, resp. Phytoseiid densities accumulated to an avg of 456 phytoseiid-days{leaf in plots where they were excluded and 1342 phytoseiid-days{leaf in untreated plots. Yield deciined byan avg of 1 tfha of cassava per 2691 mite-days{leaf; conversely, 90 phytoseiid-days{leaf provided a protective effect equivalent to 1 t/ha, indicating that A ídaeus is an important biological control agent of M. caríbbeanae. In a second predator excJusion site, Ciénaga de Oro, where M. tanajoa was the predominant species of tetranychid (97%), avg yields of 25 and 21 tfha (12 mo after planting) were obtained in plots with and without predators, resp. Although the yield difference was not significant, mite infestation intensity was 8,897 and 17,398 mita- days/leaf in plots with and without predators, resp.; and an avg of 1 tfha was lost par 2125 mite-days{leaf. Conversely, 4.8 phytoseiid-days{leaf provided a protective effect 146 equivalent to 1 tjha. In thls seasonally dry area, A Ilmonicus s./. and A rapax comprised 69 and 27% of tha phytoseiid population, raspo 8.1.4.2 Eunga! pathogens. Controllad studies of the impact of Neozygítes sp. are not yet ava/lable; however, indirect evidenca trom field observations suggests that under cartain conditions Neozygltes may play an important role in biological control of COMo In an epizootic of the fungus detectad in PMjay (Magdalena), 78% of M. tanajoa indlviduals were infected with Neozygites sp. on the sampling date immediately before the M. tanajoa population crashed (Fig. 8.5). Mortaiity dua to Neozygites sp. and to ralntall . (280 mm), which fall during the period precading the fifth sampling date, may have been the key factors rasponsible for this. 8.1.5 Characterlzatlon of CGM In NE Brazll Ouring exploration in NE Brazil for natura! enemies as part of the CGM biological control effort for Africa. CIAT and EMBRAPA observad that the severest CGM attacks occurred in semiarid areas. In surveys conducted by lITA in !úrica, CGM attacks in semiarid areas were not observad; nor hava they been frequently observad within the Neotropics (excapt In Brazil), possibly because M. caribbeanae rather than M. tanajoa predominates in such areas, excapt in Brazil, where M. caribbeanae doas not occur. figure 8.5. Sampling Oat" Dynamlcs ot Neozygltes sp. Infectlon ('lb mites Infeeted) In a fIeId popuIallon ot CGM; infectlon rate not avaUable for sampling dates 12·16. 147 In an extensive biotaxonomic analysis of M. tanajoa, polymorphism was found with respee!: to the lengths of the dorsal setae, a character usad in species identification. Polymorphism was found in all Neotropical regions where CGM oecurs, except for NE Brazil, where 100% of 352 mites examined trom 35 collection sites were classified into a single morph (very short) by cluster analysís, corroborating earlier taxonomic work done in Brazil. Similar work in Africa (Rogo et al. 1988)&.3 demonstrated that the Africen CGM is polymorphic. Mites of the genus Mononychellus are rarely found on hosts ather than cassava and its wild relatives, and the specíes of !he Mononychellus complex appear to have adapted to different cassava agroecosystems. M. caríbbeanae occurs primarily in hot, dry areas in the Caribbean besin and along !he Ecuadorian coast. M. mcgregorl occurs in humid inter-Andean vaileys and in the Amazon Basin of Colombia, Peru and Ecuador. M. planki is an extremely rare species. M. tanajoa has a broadar agroecological range than its sibling species; however, it daes not occur north of Panama or south of Colombia in the Andean Region. M. tanajoa is !he only species of the complex present in Paraguay and Brazil. Although M. carlbbeanaa is wall adaptad to subhumid araas, it does not occur in NE Brazil. ' These biotaxonomic and agroecological distribution data suggest that the trophic association between Mononychallus and cassava may have originated in the northem Neotropics. Outside Africa, the only regions where M. tanajoa occurs in the absenca of other Mononychellus specíes are Brazil and Paraguay. In Africa, M. tanajoa has not invaded semiarid areas although the data from Brazil demonstrate that this Is possible. The monomorphic character and semiarid adaptation of CGM in NE Brazil suggest that a race or subspecies 01 CGM has developed there. A possible explanatlon 10r this is that CGM was introducad to Brazil from the northem Neotropics; and it was able to colonize semiarid areas in the absence of competition from M. caríbbeanae and predation by species such as A tenuiscutus and A rapax and straíns of A IImonícus s.l. adapted to semiarid areas. 8.1.6 Estlmated impact The CGM has been the object 01 a pest management effort by EMBRAPA in NE Brazil sinca the 1970s. Losses have been estimated at 10-50%, depending on agroecological zone, variety, plarning date, planting system and length of the crop cyele CSQurca: EMBRAPA). ... Rago, LM., 0100, W. Nokoe. $. and Magallt, H. 1988. A study of the Mononychellus (Acari: Tetranychiclae) species complex from selected cassava growing amas of AfrIca usíng pñncipal component analysis. Insect Se!. AppIIc. 9, 593-599. 148 8.1.7. Strategy for managing CGM In NE Brazil Several phytoseiid specíes and strains not detected in Brazil have been found in homologous seasonally dry and semiarid cassava-growing areas in northem South America, suggesting a potential far increasíng the effectiveness of local natural enemies in NE Brazil through augmentation and conservation practices,and for improving the level of biologicaJ control through the introduction of exotic specíes. The CGM management strategy should also inelude the use of resistant clones. goad-quality planting material, as well as agronomíc practices that conserve soil fertility. The deployment of Neozygltes sp. as a biological control agent will depend on selecting virutant strains adapted to the agroecological conditions of target areas and developing simple technology for multiplying and applying the fungus. Further investigation of culture methods, specfficíty and an analysis of possible risks to human health are necessary. This should be accomplished as part of an integrated effort involving farmers in the testing and adaptation of crop protection technology far CGM and other pests and diseases associated with targeted agroecological areas. Sites for crop protection pilot projects should be chosen in areas where farmers have access to markets far the additionaJ production accruing trom . reductions In pesl impacto 8.2 The Cassava Hornworm Erinnyis ello (L) The cassava hornworm is one of the most serious pests of cassava in the Neotropics. E. elfo has a broad geographic ranga, extending trom southern Brazil, Argentina and Paraguay to the Caribbean basin and the southern USo E. ello is polyphagous with at least 35 recorded foad plants, including 21 specíes of Euphorbiaceae. Larvae feed on cassava leaves of all ages, stems and leaf buds. Severe attacks result in complete plant defoliation, bulk root loss and poor root quality. In simulated damage studies, yield Iosses in fartile soils ranged from O to 25% for one attack and up to 47% after two consecutive attacks. On less fertile soils, losses were batween 15 and 45% for one attack and up to 64% after two attacks. Losses in farmers' fields after one attack were 18%. Repeated attacks are mast common when iII-timed pesticide applications do not destroy 5th instar larvae or pupae, but eliminate the natural enemies that build up during the initial hornworm outbreak. Control measures basad on sound ecological principies are needed to avoid pesticide use. 8.2.1 Blology and ecology The migratory flight capacity of E. ello is weU documented. Adults migrating en masse will oviposit in cassava fields. Under conditions of high leaf area, up to 600 eggs/pl may be found; larval populations may exceed i00/pl. A 5-yr study 01 E. ello populations using light traps at CIAT showed that hornworms are present throughout the year and that peak activity coincides with the rainy season, during which the abundant 10llage required for 149 developing larga hornworm populatians Is available. These data are supported by observatlons and data fram ather areas in Brazil, Colombia, Mexico and Cuba. Current hornworm mass-rearing methads (CIAT Annual Report, 1989) yleld 4000 to 5000 eggs/day and provide larvae of known ages for experimental purposes. Larval duration at 15, 20,25 and 30°C averages 105, 52, 29 and 23 days, resp. Development Is 5 times taster at 30·C than at 15°C. The minimum temp threshold (MTT) for development is 11.2° C and 398 degree-days (" O) are required for development to adulthood (CIAT Annual Report, 1987). These data suggest that peak hornworm activity should occur in lowland to middle altitudes (800-1200 m) in the troplcs and during the summer penods in the subtropics. 8.2.2 Blologlcal control . Research on the cassava hornworm natural enemy compJex was initiated more than 15 years ago. Approx. 30 species of parasites, predatars and pathogens have been studied, leading to the hypothesis that migratory capacity 01 hornworm adults is a mechanism for escaping from natural enemies, which greatly reduces the effectiveness of natural biological control. Because their rate of reproduction is limited, predators and parasites cannat usually compensate quickly enough to suppress dramatic hornwarm outbreaks. Although well-timed pesticide applicatians can bring hornworm irruptians under control, they are costly, toxic to natural enemies and to the enviranment, and may cause outbreaks 01 ather pests such as mites. For biologícal control to be effective under these conditions, native natural enemy populations must be augmented when adult invasians occur. An easily manageable natural enemy must be available for introduction dunng the early stage of larval feeding. A granulosis virus of the family Baculoviridae infects hornworm larvae and is effective in managing hornworm populations. To determine pathogenicity, infected larvae were collected from the field, Iiquified in a blender and filtered through cheesecloth. The resulting liquid was diluted with water and applied to cassava plants. Within 72 h after feeding, larval mortality reaoMed 100%. In a field trial in El Patla, Colombia, fresh virus preparation was applied to hornworm-infested fields. Hornworm numbers were monitored on 50 plants in treated and nontraated plots before and 48 h after application. Mortality at 48 h was 98% (CIAT Annual Report, 1988). The effects of virus conen. and larval instar on mortality were evaluated after 72, 96, 120 and 144 h. At 0.9 mi virus/lt water (1.5x10· inclusian bodies/It), 90% mortality was obtained. A sigmoidal relationship between concn. and martality was found for the 1st, 2nd and 4th instars. In the 3rd instar the relationship was asymptotic, with higher 150 mortality at a lower concn. than in other instars (Sellotti et al., 1991)"""'. Most 5th instar larvae reached the prepupal stage. At the lowest conen. (185x10' inclusion bodiesjlt), 30% mortality occurred; 53% mortality was obtained at the highest concn. (4.5x10· inelusion bodies¡lt) (Table 8.1). At the lowest canen., 60% of the prepupae reached the pupal stage; whereas only 28% pupatad at the highest conen. Considerable pupal deformity was observad. Adult emergence from the pupa! stage was 45% at the lowest conen. and 15% at the highest. Wing deformity was common in adults. Very few female adults emerged, dying without producing progeny. These results indicate that virus applications to 5th instar larvae at high canen. can effectively reduce poPulations of subsequent hornworm populations. The lowest LC.., (0.06 mi virus¡1t water) was found far first instar larvae. LC .. ¡ncrsased to 0.45 mI of vlrusjlt (Fig. 8.6) far 4th instar larvae, indicating that progressively higher conen. are needad far adequate control of each succeeding hornworm larval instar. This reinfarces the importance of early detection of hornworm attacks so that virus preparations can be applied while larval populations are in the early instars, which are susceptible te low canco. of the virus. TabI. 8.1. Effect of concn. of 8aculovlrua en ... _ of 11th ¡nalar _ of E • • tllo. " SuMváI (No. lncIue!on BodI .. jIt H,O) Prepupa Pupa AduR O 85 80 57.l$ 1.85 x 10' 70 60 4!i 7.4. 10' 65 4!i 37.5 2.2. 10' 65 40 30 8.8. 10' 65 40 30 1.9. 10" 152.5 32.5 25 3.30 10" 47.5 21.5 17.5 4.5 x 10" 47.5 21.S 15 n .4O/stage ... Bellottl, A.C.; Arias. B.; Guzmán, O.L 1991. BioIoglcaJ control of lhe cassava homworm Erlnnyis ello (l.). Florida Entomologlst. (In press). 151 leso (mi vlruallt water) ::~ . I 0.3 ~ I ! i i 0.2 j I 0.1 o .. ~.~._-.., 11 111 IV lnatar Figure as. SusceptibHity of larval instars of E. ello 10 BaculOllirus. Duration of pathogenicity was determined after applying a 20% virus solution to cassava tields. At O, 1, 5,9, 13 and 19 days after application, leavas were removed and placed in petri dishes in the lab and fed to tirst instar larvae. Mortality was measured 96 h later. At 24 h after application, mortality was > 96%, declining to 11% after 19 days. At 9.4 days after application, 50% mortality was attained (Fig. 8.7), indicating that in the case of prolonged attacks, the virus should be applied every 9 days to obtain optimal control. There was no significant difference in mortality between virus preparations applied with or without an adjuvant. Hornworm attacks tend to increase in frequency in areas of intensiva cassava production, such as coastal Ecuador and Colombia, and NE and southern Brazil. Natural blological control is inaffective in prevanting outbreaks due to adult migration. The hornworm virus provides an attractíve management option given its easa of manipulation and storage, and low costo Research on improved storaga and application of the virus is still needad. The virus is being usad by farmers in some parts of Brazil. 152 - Virus + Adjuvant -e- VirU3 - Adjuvent -e- Control .. Mortailty 100 o 90 80 ! - 70 80 x 50 o 40 o es 30 20 O "'-! I 10 8 : 9 9 oi O 5 10 15 20 Thrt_ 'Caye) Figure 8.7. Perslstence of Baculovlrus. applled wlth and without adJlJvant, to fle!d poPulatlons of E. ello. 8.3 cassava Mealybug Numerous species of mealybugs attack cassava; however, only Phenacoccus herreni and P. man/hotí are important economically. Both are of Neotropical origin; but P. man/holi, a major pest of cassava introduced te Africa, is confinad to Paraguay, certaln areas of Bolivia and the Matto Grosso area of Brazil. P. man/holi, which has caused heavy yield Iosses in cassava in Africa, has been the object of a majar classical biological control program. CIAT collaborated extensively with liTA in Jdentifying tha geographic origin of this pest and in studying its natural enemies. The hymenopteran parasíte, Epidinocarsls lopezi, discovered in Paraguay, has become 153 established in Africa and Is bringing the mealybug under control. At present CIAT research is focused on P. harreni. 8.3.2 Damage yleld losses P. herreni, reported only from South America, causes damage similar to that of P. manihoti. It has been detected in certain areas of Colombia, Venezuela and the Guayanas, and Is spreading through NE Brazil. Yield losses up to 80% have been estimated in farmers' fields. Yield losses in experimental plots have reached 88%, depending upon clone, plant age and duration of attack (CIAT Annual Reports, 1984, 1987, 1988 and 1990). Mealybugs extract Ca trom cassava teaves during feeding. Loss of Ca may result in weakened, less rigid cel! walls, which may be responsible for the leaf curling characteristic of mealybug damage. Reduction in photosynthetic rate, transpiration, mesophyll efficiency and moderate increases in water pressure deficit, internal CO. and leaf temp were found in infested plants. A positive correlation was found between low photosynthetic rate and lower leaf Ca content, suggesting that Ca-rich clones may be more tolerant of P. harreni attack than Ca-poor clones (CIAT Annual Report, 1988). 8.3.2 Biology and ecology P. herreni populations peak during dry seasons (van Oriesche et al. 1990)""·. Rains reduce pest populations and permit plant recovery. The development of P. herreni was studied under controlled conditions at 20, 22, 25, 30 and 35·C. Female development time was 90 days at 20·C, 38 days at 25·C, 39days at 30'C and 59 days at 35·C. Optímal female development occurs between 25 and 30· C. The MTT for females was 17.S·C, and 331'0 were required to complete development (Herrera et al., 1989)"". 8.3.3 Mealybug control A combination of HPR and biological control can ofter stable control of mealybug populations. Partíal resistance (Iow to moderate levels) may reduce P. herren; to levels that make biological control more effective. Where adequate numbers 01 natural enemies are present in cessava fields, high levels 01 HPR may not be required to maintain mealybug populations below economic injury levels. •.• Van Driesche, R.G., A.C. Bellottl, J.A. Castilo and C.J. Herrera. 1990. Estimatlng total losses from parasitolds for a field populatlon of a contlnuously breedlng insect, the cassava mealybug Phenacoccus herren! (Homoptera: Pseudococclclae) in Colombia. S.A. Florida Entomologist. 73(1 ):133-143 . •.• Herrera, C.J., R.G. van Driesche, and A.C. BeIIotti. 1989. TemperaturlHlependent growth rates for the cassava mealybug, Phenacoccus herren!, and two 01 its encyrtld parasitolds, Epidinocarsis diverslcomis and Acerophagous coccols in Colombia. fnt., Exp. Appl. 50:21-27. 154 8.3.4 81010glcal control Approx. 70 species of parasites, predators and pathogens of cassava mealybugs have been identified in the Neotropics. Current research efforts are primarily directed toward the natural enemy ccmplex associated with P. herreni; however, collection, evaluation and provision of natural enemies to lITA for evaluation and possible release in Afrícan continues. 8.3.5 ParasUes Parasites of P. harreni identified in COlombia, includeAcerophagus coceo/s, Epidinocarsis diversicomis, Anagyrus putonophilis, A insolítus and Apoanagyrus e/gari. Recent explorations in Venezuela led to the identification of Aenasius sp. (near vexans) as an important parasite. E. divers/cornis preters 3rd instar nymphs; whereas A coccois parasitizes male ccccons, adult females and 2nd instar nymphs with aqual frequency. Ovipositor penetration by E. diversieomis causad 13.2% mortality of first nymphal instars (van Dríesche et al., 1990). In-field studíes, using trap plants with mealybug hosts set out in cassava fields, 54.9% mortality Was estimated far the combined action of the two parasitoid species present (van Driesche et al., 1988, 1990)8.7. A n. vexans, introducad to Colombia from Venezuela, is in culture and under evaluation in the lab and field. A n. vexans is more SpeciflC for P. herreni than othar parasitoid species evaluated. Ina choice test offering the preferred stage of each host, A n. vexans parasitized 36% of P. herreni vs. 2% of P. madeirensis. A coccois ccllected from Venezuela parasitized P. herreni and P. madeirensis with equal frequency, whereas A eoccois collected In Colombia is highly'specific to P. madeirensis (Table 8.2). Distinct biotypes of A coceo/s may be involved; however, they were not distinguished through electrophoresis of esterases. In studies of host stage preference of A. n. vexans, A coccois and E. diversicornis, all three species parasítized all host stages; however, within- species preferencas for specific instars were ídentified. A. coccois (Ven.) prefers 2nd instar nymphs; E. dlversicornis prefers srd instar and adult females; and A n. vexans prefers 2nd and 3rd instar nymphs and adult females (Table 8.3). When A n. vexans parasitizes 2nd instar nymphs, nearly all offspring are males. Parasitism of 3rd instar nymphs also results in a Iower sex ratio; however, parasitism of adult female P. herreni resulted in 86% female offspring. E. diversicornis is parthenogenetic and no males are produced; however, reproduction was maximized when 3rd instar nymphs were the host. The fact that A n. vexans females are larger than E. diversicorn;s and A coccois females (0.47 mm, 0.43 mm and 0.23 mm, resp.) may ínfluence progeny sex ratio and host stage preference. Although both species (A. n. vexans and E. diversicomis) will parasitize first instar nymphs, no offspring result (Table 8.4) . ... Van DrIesche. R.G., J.A. Castillo and A. C. BellOItI. 1988. Field placement of mealybug·lnfested potted cassava plants for the study of parasltlsm of Phenacoccus herrenl. Entomo!. Exp. AppI. 46:117·124. 155 Table 8.2. Preference 01 4 parasitoids (Encyrtidae) In a choice test for P. herreni and P. macls/tens;s. Parasitoid Acerophagus coccois (Venezuela) A coccois (Colombia) Asnasius n. vexans (Venezuela) E. diversicom/. (Colombia) 1 Data taken from several experiments. P. llenen/ 32 o 36 32 % Parasitism 1 P. madeirensis Z1 Z1 2 16 Tablo 8.3. Proferenco 01 lhreo parasitold speclo. (Encyrtidao) in a choice test Ior lile stages of P. llenen/. Stage Instar I Instar 11 Instar 111 Mult temale , Data taken from severat experiments. (P = 0.05, DMRT) A coceol. (Von.) S.3e 63.3 a 26.7b 1.7 d 'l(, Parasitism ' ParaaRe Speeies E. dlverslcom/s 10.3 e 19.0 b 34.9 a 35.8 a A n.wxans 3.5 e 35.1 a 35.1 a 26.3b Tablo 8.4. Comparison of parasitism ratos A n. \I8xans and E. diversicomis compotlng for dilferon! lifo stages of P. llenen/. Mealybug Slago' N' Nymph 1 250 Nymph 2 250 Nymph 3 250 Mull 250 Total 1 Five reps of 50 individuals from eaeh mealybug stage. 2 Five paJrs of A n. vexans and five E. dlvers/comis. Females O d 1 e 17 b 37. 55 156 No. Mealybugs Parasitized Species A n vexans E. diversJcom/s Males Female. Od Od 17b 19 e 24a 32a Se 21 b 47 72 Egg-to-adult developmental time is 16 days for A n. vexans and 18 days for E. diversicornis at 28°C and 60:1:10% RH. P. herreni develops in 39 days, indicating that there are at least 2 generations of the parasitoids per host generation, a favorable ratio in a biological control effort. Fiald releases of A n. vexans were made in cassava fields at ICA-ViIlaviceneio, Meta, Colombia. Terminal buds and leaves were artificially infested with P. herreni. Parasites were released and % parasitism was compared in (a) terminals completely enclosed in sleeve eages, (b) enclosed terminals with eages opan at one end, (e) uncaged terminals artificially infested with P. herreni, and (d) uncagad, uninfested terminals. Parasitism lavels were measured 10 and 45 days after release of A n. vexans. Considerable parasitism was obtained in uncaged terminals and on terminals in open cages, resulting in a lower rate of mealybug population inerease than in the eaged terminals. A n. vexans persisted in natural mealybug populations in cassava plantings adjacent to the experimental field throughout the subsequent rainy season, indicating that establishment of this parasitoid may have oeeurred under field eonditions. . 8.3.6 Predators A large complex of predators, mostly of the family Coccinellidae (Coleoptera) have been identified associated with P. herreni. Several of these have been evaluated to determine their potential for biological control (CIAT Annual Repcirts, 1987, 1989-90; Sullivan et al., 1991)···. Previous studíes determinad tha davalopment cycles and eonsumption rates of C/eothera onerata and C. notata. Prasant research has concentrated on Hyperaspis sp., a new species discovarad during axplorations in tha Colombian Uanos. Oevelopment tima of Hyparaspis sp. was maasured under controlled conditions at 20, 25, 28,30 and 33°C. Female development time was 88 days at 2Oc C, 29.6 days at 25°C and 20 days at 33°C (Fig. 8.8): The MTT for female development was 15SC, and 251.4° O were required to complete development. This compares favorably to the development rate of the prey, P. harreni (17.5°C MTT and 331°0). Hyperaspis sp. mortality was highest at the lowest temp; 33% mortality occurred at 2Q°C, 6% at 28°C and no mortality occurred at SS·C, indícating adaptation to high temp conditions, whieh also favor P. herreni population growth. Hyperaspis sp. prefers to feed on e9gs, followed by the 1 st, 2nd and 3rd instars and adult stages. Consumption capaeity is greatest during its 4th instar, when it can consume 185 eggs/day (Fig. 8.9). These results are similar to those obtained for other cassava mealybug predator spacies. e.e Sullivan. O.J .. JA Castillo and A.C. Bellonl. 1991. Comparative biology 01 six species of coccinellld beetJes (CoIeoptera: Coccinellidae) predaceous on the mealybug Phenacoccus herreni (Homoptera: Pseucoccldae). a pest 01 cassava In Colombia. Soulh America. Envlron. Entomol. 20(2):685- 689. 157 Figure 8.9. I11III Famal. _ Mal" Figure. 8.8. Effect of temp O" development time of Hyperaspls sp. Prey Lile 810.0.' _ Mull F.m.I. _13 1llIlIllIl12 911 m Egg Ptedator Lita Sta;. Consumption capacity 01 Hyperaspls sp. Ilfe stages lor P. hartení eggs. Immature stages and adult !amales. 158 Survey data from South America indicate that the P. harren; parasitoid and predator complex is large in Colombia and Venezuela, but smaller in NE Brazil, where P. harren; populatíons have caused severe crop damage. P. herreni may be of northern South American origín and may have been disseminated along the eastern caast of South America to NE Brazil, where it appears to be sn introduced pesto Introduction of natural enemies from northern South America has potential for stabilizing P. harTenl below economic damage [evels in NE Brazil. 8.3.7 Host plant reslstance Approx. 3000 lines from the CIAT germplasm bank have been screened for P. harrenl resistance at two selection sites over a 5-yr periodo Artificial infestation was used to ensure uniform population density. Damage evaluations based on s scale of 1 (no damage) to 9 (necrosis of leaves and branches) were made periodically after infestation. In previous evaluations clones CM 6Q6l3-3, CM 21n-2, SM 540-8, SG 250-3, SG 106-54 and CM 2177-2 were selected as promising sources of resistance. The last three were evaluated in replicated tríals with and without pesticide to determine yield depression as a resistance criterion. Yield'reduction was 10.1 .. 9.3 and 34'.2% far SG 250-3, CM 2177 and SG 106-54 resp. (Table 8.5; CIAT Annual Report, 1990). Adequate levels of resistance or tolerance were identified in the first two ciones to justify their incarporetion in a resistance breeding programo Clones CM 6069-3, CM 5263-1 and SM 5409-8 selected during 1990 are currently under evaluetion far yield depression. During the 1990-91 cycle, 243 clones were evaluated far P. harreni resistance in carimagua and Villavicencio. CM 2600-2, CM 5948-1, CM 3581-17 and CM 5620-3 were selected to enter the yield depression evaluetion scheme. Mealybug populations were higher and damage was severer at Garimagua than Villavicencio, indicating that the former site is superior for resistance screening. Additional greenhouse and lab studías have been inmated to determine whether the resistance expressed is tolerance associated with plant vigor or due to antibiosis mechanisms. 8.4 Cassava Root Mealybug: Pseudococcus mandlo ,Areas of Paraguay, Argentina and southern Brazil (especially the state of Sta. Catarina) haya reported root damage due to the mealybug, P. mandlo. This species feeds on underground plant parts including the swollen roots and below-ground stems, causing clorosis and defoliatlon of lower leaves and darkened spots on Infested roots. Damage reduces commercial value of the roots, and culinary quality is affected. Continuous cassava cultivatlon results In increased mealybug populations. A callaboratlve project, with research contracted with EMPASC, Itajal. SC (Brazil), Is studying the occurrence, behavior, damage and control of this pesto Preliminary results are presented from 1989 to date. 159 Table 8.5. Yield '" pmtected and mealybug f¡>. hemlnij-inI",,1od pIoIs In Carimagua yield-depresoion tri .. Protectod YieId Infeoled Vi.ld Yleld Loss Clone t/ha !/ha % CM 2177-2 29.5 e' 26.7<1 9.3 $G250-3 39.2. 35.2b 10.1 M Ven 77 30.3 e 20.4 e 32.6 SG lQ6.54 25.6<1 16.8 19 34.2 CMC40 14.2g 4.6h 67.0 MCcl22 18.6 ef 2.1h 86.3 1 Mean. with the _ leller aro nO! slgn1flcantly dlflereOl (P • O.OS). 8.4.1 Blology Developmenta/ studies 01 P. mancJio were done in the lab (25±OC; 80± 5% RH). Aftar eclosion, temales complete 3 nymphal stages befare adulthood. Ma/es have 4 nymphal stages (the 3rd and 4th stages occur in tM pupal coceon); the 5th stage is the winged adult. Avg egg-to-adult development time for the fama/e is 25 days; adults live an avg ot 17.8 days. Avg development time for males is 17.9 days and adults live 1 to 3 days. Fertilization is necessary for reproduction; unfertilized fema/es do not produce eggs. 8.4.2 Yield losses Yield loss studies were done during two cropping cycles. During the first cyale yield reduction was 17% (significant, T-test; P = 0.05); during the second cycle there was no significant yield reduction. 8.4.3 Control Research on control of P. mandio has been inítiated wlth emphasis en crop rotation and biological control. 8.5. The Cassava Burrowing Bug The cydnid burrowing bug Cyrtomenus bergl was first recorded as a pest of cassava in Colombia in 1980. C. bergl is reported trom Colombia, Costa Rica, Honduras and Panama. Subsequently, several other species of Cydnidae, which cause similar damage to cassava were identified. This complex includes Pangaeus piceato, P. aethíophs (Colombia); Tominotus communis, Prytidorsporus ídentatus and Dallsielus lugubres (Cuba). 160 Surveys in Colombia and other areas have revealed that onions flúlium tlsw/osum), peanuts (Arachis hypogea), maize (Zea mays), sorghum (Sorgum vulgare), sugarcane, coffee, coriander, pastura grassas, potatoes and numerous weed species are also hosts of C. bergi. Yleld losses in peanuts and onions are considerable, and repeated pesticide application are required for effective control as other control maasures are not available. This pest apparently has the capacity to switCh from a preferred host suCh as maize to other hosts such as cassava, making control of C. bergi a difficult chal/enge. The insect feeds by insertíng its stylet into the fleshy root (parenchyma). Feedíng damage combined with infection by soil pathogens results in the appearance of brown to black spots 12 to 24 h after feeding is ínitíated. From 70 to 80% of the root may be affected, and a ~ 50% reduction in starch content is common, resulting in serious reduction in commercíal value. 8.5.1 Biology, ecology and behavlor Development and behavior studies were ínítíally hindered because of the inability to rear C. bergí in the lab. This problem has been solved by replacíng cassava with maize as a culture medium. Lab colonies of 10,000 to 12,000 índíviduals of alllita stages (egg, 5 nymphal instars and adults) have been established. C. bergi feeding bahavior can be studied on 'artificial' cassava roots--pellets consisting of low-HCN cassava flour encased in paraffin wrap. These pellets can be usad to evaluata nutriant raquirements or pesticide effectiveness. C. bergi develops taster on maize than on cassava and preters maize over cassava in free-choice feeding tests (78 vs. 22%). OViposition was 300 times greater on maize than on cassava (105 va 0.4 eggs/female); intermediate on onlons (CIAT Annual Report, 1990). The LO .. on maize was 95 days, compared to 69 on onions and 66 and 64 days, resp., on a sweet (eMC 40) and bitter (M Col 1684) cassava cione. Nymphal development was shortest on maize (92 days), longest on onions (119 daysl, and intermediate on cassava (111 days) (CIAT Annual Report, 1989). Field tríals suggest that C. bergi feeding preferences may be related to the HCN content of the root. Previous lab studíes showed that aduits and nymphs fed on a high HCN cione had longer nymphal development, reduced adult longevity, reduced fecundity and increased mortality comparad to insects fed low HCN clones. Recent studies under controlled conditions compared nymphal mortality and feeding on a low (CMC 40) and a high HCN (M Col 1684) clone. On CMC 40 nymphal mortality was 56%, occurring during the first 2 instars. Nymphal mortality on the hígh HCN clone in two trials was 94 and 84%, resp. (Table 8.6). AII mortality occurred duríng the first 3 ínstars, with híghest mortality in the first two. On eMC 40, 50% mortality occurred at 35 days (r"=0.96); for M Col 1684, at 28 days (r"=0.98). The aarliest instars are most susceptible to root HCN content, and even low HCN levels cause considerable mortality. Nymphal development time for the 1st through 3rd instar was 43.5 days on CMC 40 and 81.7 days on M Col 161 Table 8,6, Mortalíty 01 C. Mlfili nymphs 1 •• dl.9 on Iow· (CMC4O) and hlgh- 1M Col 1684) HCN _il rOofS. % Mortallty Oone Inotar CMC-40 M Col 1684 26 40 1I 30 42 111 O 2 IV O O V O O Total % Mortalíty 56 84 n= 50 X2 test oIlndepe.oence lo, Instars I ano 11 algnifleant, P $ 0.05 1684. This supports previous observations that cassava is not a preferrad host for C. bergi and that an additíonal host must be available for poputation growth to occur in the fietd. Surviva! is highest when C. bergi feeds on the amall storage roots of sweet varietíes and lowest when feeding on the root peel. Fietd studies also show an increase in swollen root damage following rain events. This is supported by lab studies, which indicate greater C. bergi surviva! and more activity during moist conditions. lt ís possibte that C. bergí migrate deep into the soU and diapause during dry periods until conditions become more favorable. The use of ground covers or mulches, which increase soil moisture condítions, coutd increase C. bergi activity in cassava. In tield trials on the Colombian North Coast, C. bergi damage recordad on 7 clones averaged 24.3% during the May-to-May vegatatíve cycle when rains occur during the tirst 6 mo. Root damage during the Oct.-to-Oct. (dry season) cycle was only 0.6%. These results support observations that soíl moisture may influence C. bergí attacks. 8.5.2 Control of C. bergl 8.5.2.1 Croo manaaement. Previous research has shown that intercropping of Crota/aria sp. (sunne hemp) with cassava is an effective means of controlling C. bergí through allelopathy. lab studies compadng feading behavior on severa! erops showed that Crata/aria was least preferrad. Trials in farmers' tields showed that when Crata/aria is sown between eaeh cassava row, root damage was only 8%; whereas in control plots of cassava monoculture, damage was 77%. Pesticide applications are less effective than 162 intercropping with Crata/aria (CIAT Annual Reports, 1988-89). As Crata/aria has I1ttle or no commerciaJ value and as the intercrop results in reduced eassava yields, farmers are not inclined to adopt this technology. Nevertheless, allelopathy may eventuaJly contribute to the control of C. bergí if a commercially acceptable system can be developed. 8.5.2.2 Chemical control. The insecticide chlorpyriphos was incorporated in cassava pellets (see previous description) at 4 concn. (2.5, 5, 10 and 20%) and placad in plastic boxes in the lab. First, 3rd and 5th instar nymphs and C. bergi adults readily feed on these. At the 3 highest concn., 100% mortality of all instars and adults occurrad within 24 h of initiating feeding. At the lowest concn. (2.5%), 100% mortaJity occurred for the 1 st and 3rd instars during the first 24 h and for the 5th instars and adults at 72 h. The use of cassava pellets containing pesticides as poison baits may be feasible for controlling C. bergi, elimínating the broadcasting of soil pesticides presently employed in certain areas. 8.5.2.3 Biologjcal control. The potential for biologícal control of C. bergi with nematodes is under investigation. Steinernema carpooapeae was evaluated for pathogenicity to C. bergi under controfled conditions in the lab. Plastic containers with soil and maize kernels were infested with 70 C. bergí nymphs and adults. The soí! was inoculated with approx. 20 nematodas previously reared on artificial media. Six days after inoculation, 20% mortality of C. bergi was observad with a mean of 19 nematodes recovered per dead insecto Analysis of soil samples reveaied considerable numbers of S. carpooapeae. Thase results suggest potential for using entomopathogenic nematodes to control C. bergi. Other nematode species will be evaluated and field trials will be designed. 8.6 Whiteflles Numerous species of whiteflies are reportad en cassava. The predominant species in the Amerieas are AJeurotraehelus socia/is, Trla/eurodes variabills, Bemisia tubereulata, AJeurothrlxus aepim and Bemisia tabae/. In recent years high whitefly populations have been reportad trom Paraguay (B. tuberculata), Brazil (A. aepim), Colombia (T. variabilis and A socia//s, and the Dominican Republic (B. tabec/). Results trom early research show a correlatíon between duration of whítefly attack and yield loss; an 11-mo attack resulted in a 79% yield loss. These and other results have providad an estímate of probable yield Iosses in areas of high whitefly populations. This has justified a continual aJthough moderate effort in whitefly research, with emphasis on identifying resistant germplasm and developing resistant clones. Recent developments including increased whitefly populations in severa! areas and the appearance of a new blotype of B. tabeci has led to an increase in research efforts. 8.6.1 Bemisia tabaci Whiteflies are known to transmit severa! virus diseasas (see Chapo 7). B. tabeci transmits African Cassava Mosaie (ACMO) in Atrica, a disaasa of considerable importance, which 163 has not yal been reportad in the Americas. Although B. tabaci is presant in the Americas and is a major pest in several important crops (beans, soybeans, etc.), until recently it had not been reported feeding on cassava, suggesting the existence of a distinct biotype from that feeding on cassava in Atrica. The absence of ACMD in the Amerieas is thought to be related to the inabílity of ita vector, B. tabaci, to colonize cassav8. In recent years, B. tabae! has been reported feeding on and transmitting virus diseases on numerous crops that it did not previously colonize. It has been reported feeding on eassava in the US (Florida), Puerto Rico and the Dominican Republic. This indicates the possible development or introduction of a new biotype in this area. If thls biotype is capable of transmitting ACMD and thls dlsease enters the Ameneas, it could devastate cassava production. Whiteflies collected 1rom cassava in a 1991 survey of Florida were identified as B. tabaci (Source: Louise Russell, USDA) and vermed by electrophoretic studies at CIAT. These findings warrant an increased emphasis on whitefly research. 8.6.2 Whltefly control HPR studies wlth A. social/s, T. variabilis and B. tubarculata--the most economically importa'nt species in Colombia-were initiated nearly 10 yr ago. More than 2000 clones have been evaluated. Clones M Ecu 72, M Col 336, M Bra 12 and M Col 339 were selected as resistant or tolerant 10 whitefly attack. M Ecu 72, which has consistently expressed the highest level of rasistance, ware used in a crossing program with other salected dones to produce whitefly-resistant, high-yielding clones. Using astimates of yield depression and plant damage ratings 10 compare resistance levels, 4 clones were selected (CG 489-34, CG 489-23, CG 489-31 and CG 489-4) (Table 8.7). These studies, done in collaboration with ICA-Nataima, Tolima, should eventually lead to the official release of these dones. More recent evaluations 01 elite cassava germplasm have k:lentified several additional source of resistance (CG 959-1, CM 2156-3, M Bra 191, CM 1203-13, CG 1141-1, CM 2136-2, CM 4042-4, CM 4157-34, SM 301-3 and CM 2156-13). These dones are under evaluation for yield depression, and an additional 300 clones are being screened as posslble sources 01 resistance. Resistant materíals have been sent to liTA 10r evaluation for resistance to B. tabaci, the predominant species attacking cassava in Amca. Crop resistance to whiteflies has not been frequently reported and Is not aasily achieved. Results with cassava are encouraging, and it is hOPed that HPR will be a useful mechanism for reducing the incidence of virus diseases. Reports on cotton suggest a relationship between rasistance and low leaf-tissue pH. Thirty cassava clones were evaluated for leaf pH; three fully expanded leavas were removed from 5-mo-old eassava plants. From 25 to 30 9 of lea1 tissue were macerated and mixed with deionized water (2X the wt of leaf tissue), and allowed to stand for 1 h; pH readings were made at 5,10 and 15 mino 164 TobIe 8.7. Yleld reduction in 10 ca ... va clones causad by whitefly _ allCA/Natalma. Telima. Colombia. Yleld wilh Vi.1d WilhoUI insecticida _. '!lo Prolection Protection Aeduction Clone (I/l1a) ~fhal in Yteld ca 489-34 30.3 fg' :29.3 gh 3 ca 4811-23 39.1 e 35.4 de 9 ca 4811-31 27.9 Ohi 24.0 íJk 10 ca 48!1-4 32.9 el 27.1 hij 17 CMC76 22.8 k 15.8 I 31 Reg. Cuind. 25.1 iJk 17.3 31 M Me. 59 36.3 d 24.1 jk 33 CMC40 53.5 " 25.7 íjk 52 CMCS? 33.' el 7.7 m 75 H 305-122 42.1 b 8.6 m 79 Data __ from _ •• pe~ments. I Moana wiIh !he samo iefler .... not oigníflcantly dlflerenl (P = 0.05). Six resistant clones identified through field evaluetions had pHs below 6.0 (range of 5.51 for M Ecu 72 to 5.72 tor M Sra 12); the other 24 clones had pHs > 6. TI16se results indicate that leaf pH may be an indicator of whitefly resistance and could be a useful tool in screening procedures. 8.6.3 Intercropping Traditionally grown cassava is often intercropped with other species. Many evaluations have been made of insect populations under these conditions (see SE :tlon 8.5 on C. bergí). The most recent have examlned the effect of intercropping maiza and cowpeas on whitefly populations. Intercropping cassava with cowpeas reduced egg populations of two whitefly species (A. socialls and T. variabifis), relative to those in monoculture. These effects were residual, persisting up to 6 mo after hervest of the intercrop (CIAT Annual Report, 1987). Intercropping cassava with maize did not reduce egg populations. High populations of cassava whiteflies led to significant reductions in growth and yield. Yield losses in cassava/maize, cassava monoculture and mixed variety systems were ca. 60%; whereas in cassava/cowpea intercrops yield losses were 12%, and yields were superior to those obtalned in the other systems. Intercropping may offer the small-scale farmer a valuable means of reducing pest problems. especially in areas where cowpeas have a high commercial value as in NE Brazil. 165 8.6.4 81010glcal control Natural enemies of cassava whiteflies include the predator De/phastus pusillus and the parasites Amitus aleurodinus and Eretmocerus afeurodiphagus. Predators playa minor role in whitefly population dynamics (Gold et aL, 1989)"'; however, parasitism of A socialis by A a/eurodinus and E. aleurodiphagus ranged from 49.1 to 54.3% in experimental plots, indicating that parasltes may be an important mortality factor (CIAT Annual Report, 1989). 8iological control of whiteflies will be given more attention in future research. 8.7 Insects Attacking Sexual Seed The Cassava Program recently decided to invest in research on the potential of sexual seed far propagating the cassava crop. In arder to identify potential pest problems in seed storage and field germination, experiments were initiated in 1991. Measured quantities of untreated cassava seed were placed in paper bags and stored in unprotected facilities at CIAT. Seeds were inspected at 15-day intervals. After S mo, no insects or insect damage was detected. These. studies will be continued in cassava- production areas: The ability of Tribolium castaneum, Sitophílus granarius, Acanthoscelides obtectus and Zabrotes subfasciatus to feed on cassava seed was evaluated: Fifty sexual seeds were placed in plastic containers and infested with 4Q adults of aach species. Oviposition on the seed coat was observed after 2 to 3 wk. Although eclosion was observed, none of the species was able to penetrate the seed coat although Z. subfasciatus causad some rasping damage. These preliminary results indicate that the cassava sead cQat prevides considerable protectlon against insect attack. These trials will continue. 8.8 Insects Attacklng Orled Cassava The increased utilization of processed cassava for animal feed has resultad in the commercial storage of dried cassava for 1 to 3 mo. Oried cassava as flour or chips is vulnerable to numerous stored products pasts. 8.8.1 Pest infestatlons Cassava chips are susceptible to insect attack while the chips are exposed on drying surfacas and during storage. To evaluate the first case. 500 9 samples of bitter (M Col 1684) and sweet (CMC 40) cassava were removed from the drying patio at 1-h intervals. placed in glass containers. and sto red in the lab for 30 days. Samples removed while the B.. GoId, e.s .. M.A. A1t1erl and A.C. Bellottl. 1989. The effects of the Intercropplng and mlxed varletles 01 predators and parasltolds of cassava whlteflles (Hemiptera:A1eurodidae) In Colombia. Bul!. Ent Res. 79, 115-121. 166 humidity was aboye 14% were susceptible to mold; those from 11 to 14% showed no mold damage. No insects or insect damage were detected in any of the samples. Cassava chips and flour produced from CMC 40 and M Col 1684 were evaluated in 4 packing materíals: sisal, paper, cloth (cotton) and palypropylene. Of the 18 species of arthropods found in the dried cassava products, the mast frequently recovered from large chips were Araacerus fascicu/atus, T. castanaum and Rhyzoparta dominica. From small chips, the major species were T. castaneum, Lasiadarma serricorne and Dinoderus minutus. Cassava flaur was primarily infestad with T. castanaum, L. serricorna and Sitophi/us orizae. Pest populations were not significantly different in the 2 varo lhe best packing material tor avoiding storage pests was polypropylene. During 1991 several drying plants on the Colombian North Caast were surveyed to determine the extent and nature of arthropod pest infestation5. Samples taken from drying plants were infested with T. castaneum (most common), Sitophagus h%pte/oídes si/vanus sp., Psocoptera sp. and Sítophilus orizae. Significant infestations were found in cassava stored for more than 20 days. 167 9. CROPPING SYSTEMS Due to the widespread practice among small-seale farmers of íntereropping eassava and maize, the Agronomy Seetíon has dedicated substantial effort to gain more knowledge and develop better technology that ean contribute to improving the productivity and sustainability of this assocíation. Some research actívitíes are cerried out on farms with farmer participation in the desígn, ímplementation and evaluation of teehnology; other activities are conducted on experiment stations. The development of methodology to evaluate the performance of new maize and cessava varo in intercropping systems has been an important activity for too last six years. Results on Ihe performance of new varieties and on nutrient cyeling in the maize/cessava intercrop are presented. In several regions where cassava is an important crop, fallow is practiced to malntain soil fertitity. Population pressure and new market opportunities for cessava have increased demand fer roots. High demand means more production ánd consequently a more rapid dep/etion of soil nutrients. Given the socioeconomic characteristics of the farmers who grow cessava, a long-term, low-cost strategy should be developed to sustain production. Meanwhila low levels of fertilizers applied to the most common cassava-based cropping systems will help sustain yields. A summary of results obtained in farmers' fiald using fertilizers for the cassava/malze association is presentad here. Lana preparation and weed control are closely related. These are among the most labor- demanding activities in cessava production. Raduced tillage is so alternative te mechanized lsod preparation that should help reduce soil arosion and production costs. This research is long term in nature; preliminary results corresponding to the first three years of research are reported. In NE Brazil, cowpeas are commonly intercropped with cassava. A1though there were some available research results that could be applied immediately to improve the performance of this crop combination, some specificcharacteristics of the cropplng patterns usad in the semiarid areas of the world have not been studied. Results are given of an experiment station with the cropping pattern usad in NE Brazil. Technology components developed by Ihe Cassava Program sod/or national programs should be valldated under farmer management and across a wide range of socioeconomic and environmental conditions to assess their viability. This activity will provide feedback to research and apeed up the technology development procesa, as well as the transfer of this technology to farmers. Results are given of pre-production trials and of a study to validate selection and chemical protection of stakes before planting. 169 9.1 Intercropplng Cassava and Malze 9.1.1 Intercropplng cassava with improved and traditional varletles of malze In most countries where cassava is an important crop, maize is one of its most common Intercrops. Cassava Is seldom grown as a sale crop in NE Brazil, coastal Ecuador, the AtlanDc Coast of Colombia, or several other important cassava-producing areas. In Latín America, national programs release improved maize varo more frequently than improved cassava varo Therefore in regions were the cassava/malze association Is cultivated, a few traditional cassava clones are generally intercropped with severallocal and ímproved maize varo The increasing availability of improved malze varo bred orlginally for monoculture justifies the indusion of field experlments to test their performance when íntercropped with cassava. To develop a generally applicable methodology for testíng maize var. with cassava, a set of field trials was conducted from 1985-90 in coIlaboration with the ICA on-farm research team in the North Coast of Colombia. Newly released maize var., malnly V-156 and v- 109, were compared with the traditional or local maize varo A 5-yr experiment uslng a split-plot design with association and sole crop as maín plats and maize var. as subplots was conducted on more than 30 farms. The local cassava var. "Venezolana' was used as a check. Researchers planted and harvested the plots; all other aspects of crop management were the farmers' responsibility. For both maize and cassava, either in association or in sole crop, differences in yields between years were highly significant. Yields also differed significantly between locations (farms). More variability in yields of cassava and maize in pure stand and in association Is expected between years than between farms in ralnfed areas. Avg casseva RYs in monocrop were always higher t±. 2t) than in association with maize (statistically significant). Yields of maize in assocíation with cassava or in monocrop were nat significantly different (Table 9.1). If maize is as important as cassava to farmers, this small difference between yield in monocrop and yield in association justifies the widespread practice of intercropping. There was a significant interaction between types of maize and association or not with cassava, indicating that cassava does not respond equally in tarms of yield when íntercropped with different types of maize. This interaction, significant at the 10% level for maize, means that all maize varo do not perform equally when intercropped with cassava. The avg yield of the 2 improved maíze varo was significantly higher than that of the traditionalone. 170 Table 9.1. AV; yleld$ (I/IIa) al ImprOlllld and local mal .. Yar. and ......... as sole CIOP and In18«:rOPped wflh maíz •• First Second Thlrd Ma1Z8 varietles2 V·1156 V-lOO Criollo Or local Malle GY X CroppIng syált1l" eassava/V-1156 Cassava/V-lOO Cassava/Criollo Of loeaI 14.10 a' 13.75 a 8.50b 12.34 a 10.59 b 10.91 a 10.97 a 9.91b 1 Amounts foIlowed by !he ....,. _ are no! otall!lllcaUy dllferenl (1' < 0.05 DMAT). 2 Meana of monocmp and __ mal.o. • Mal.e vat. are con&Idored as _ In plols wh~re cassava woa assoolaled wflh mai.e. Maíz. Yiold 1.68 b 2.10 ab 2.67 a 2.32. 2.25 a 2.51. 2.42. 1.91 b The Competitive Aatlo Index (CAl)'" provides an estímate of the relative ability of one crop to compete with another when planted in association. This index was in favor of maize in 74% of the 134 cases analy;.:ad, indicating that C".assava AY depends heavily on the performance of the intercropPed maize. The environmental conditions during malze development and/or the agronomic management ofthe maize determine the performance of the intercropped cessava. As 75% of the cases involved improved maize var., the management of improved varietien is a key factor determining cassava AY on the North Coast of Colombia, where more than 39% of the farmers were using improved maize varo in 1990. If farmers continue to grow this crop combination, the system cen be modified most easily through maize management. This Is probably also true in other regions where thls crop combination is Common. The slow initial growth of cassava and the comparatively rapid early development of maize, particularly of the improved varieties, probably account for maize predominance when these two species grow together. land Equivalency Ratio (LER)", values aboYe 1 were obtained in 93% of the cassava/maize plots, indicating m()re efficient use of land by the cassava/rnaize intercrop than by either species planted in pure stand. In regions where fanners have limited } 1 Lelhner, D. 1983. Management and evaJuatlon of Intercropplng syst&ms with cassava. ClAT, 70 pp. 171 access to land, land use efficíency is of great importance; e.g., the North Ceast of Colombia and many other regions 01 the world where cassava is an important crop. fmproved maize varo yielded significantly more than the traditional var., both in monocrop and in association with cessava; conversely, cassava in association with improved maize yielded significantly more than in association with the traditional varo (Table 9.1). This represents an additional beneflt of new maize var., which were originally developed to perform better as a monocrop. No significant differences in yield were registered between the two improved malze varo tested, but on avg they significantly outyíelded the traditional anes. The stability of different cassava¡maíze intercrops across varlous environments was tested, and the sensitívity of these cropplng systems was estimated using methods developed by Eberhart & Russell·... Gassava monocrop ls less sensitive to environmental variatíons than.cassava associated with any of the malze varieties tested. Monocrop cassava RVs, whlch were afways > 10 t/ha even In the less favorable environments, did not respond positively in the more favorable ones (see the almost horizontal líne corresponding to cassava RV in Fíg. 9.1). 60th traditional and improved malze var. performed well in favorable environments, but their performance under less favorable conditions was relatively poor (Fig 9.2). The general performance of maize as sole crop and in association with cassava is similar in terms of environmental sensitivity"". Within the range of varieties tested, their performance in lntercropping can be predicted from performance as asole crop; however, lower yield can be expected in assocíatlon with cassava than In monoculture. From the standpoint of cassava, association with malze increases its sensitivity to the envlronment. In less-favorable environments, yields of cassava in assocíation with malze are significantly lower than in monoculture. In more favorable environments. yields of cassava in association with maize could be higher than yields of cassava in monocrop. In favourable environments expected differences in yields between cassava monocrop and cassava in association with maize may be lower than in lass favourable environments. (Fig.9.1). Improved maize var. have high Hls and thus allocate more DM to the grain. Theyare frequently planted at high densities to achieve their full yield potential. If traditional malze ls replaced by improved var., nutrient balance within the farm will change as more nutrients will be exported from the farm system (Fig. 9.3). When soíl fertility is restored ... Eberhart, SA; Russell, W.A. 1966. StabHity Parameters for comparing varieties. Crop Science 6(1):36-40 . •. 3 Cortes, M.H. 1991. AnAlisis de ensayos con el Intercultivo de yuca (Manihot esculenta Crantz) y malz (Zea mays L) en la costa norte de Colombia. B.Se. Thesls. Universidad del Valle. 139 pp. 172 Y.=le~ld~(t~/h~a~) ________________________________________ , 25, Trlala -+- CAS8AVA/V-109 CAS8A\a/V-168 -e- CA8SA\M/V-155 ~ CASSRIA/LOCAL --+- CASSIIM. SOlE CROP Figure 9.1. Estimated yields of cassava monocrop and assoclated w~h dlfferent malze varo in favorable and less·favorable environments. Vleld (tlha) 5~~~=---------------------------------, 4 3 .. - -_ .. -- ,- ------- 2 ---f:"'f •• ;jf- 1 ¡'<:ii:::=r'~"'.- --._-- -_. . ::::¡--- o 2 2 2 1 2 3 3 2 5 8 4 9 o 5 3 2 2 2 1 8 Trlala i- Y-168 -+- V-loa ---*- V-268 2 2 2 3 3 3 3 3 3 7 9 8 o 1 2 4 ~V-16& ~LOCAL I Figure 9.2. Estimated yields of maize monoculture in favorable and less·favorable envlronments, maize varo frequentiy intercropped w~h cassava. 173 Uptak .. (kg/ha) 200~----~~--~--~---------------------------, 150 100 50 O N p K Ca Mg s Uptak .. (kg/ha) 1 0.5 oc......)-- Fe Mn Zn Cu Nutrients Malze Varlelles el V 156 _ Tradltlonal Figure 9.3. Nutrlent uptake from the soiI by ImproVed and traditional maize varo 00 the North Coast of Colombia. 174 Table 92. Biomass ptOduced by CUllava vat. CMe 40 and uptake of nutrian .. In assoelation wllh dlffe,ont!ypH 01 mal.e and as so'" crop al ClAT HO. H-211' V·258' Clavo' Umeñoz Sole erop H-211 V·258 Clavo' Umoño Solocrop 1 Improved maia var. 2 '--1 mal .. var. Total 18.9 17.48 11.54 20.48 22.07 N P 111.9 19.6 101.8 17.3 102.6 17.5 124.4 20.3 130.9 21.1 Biomass (lfha) Lnves $tema Roo .. o.es 8.63 9.es 0.44 7.19 9.85 0.64 6.90 9.99 0.55 1.75 12.18 0.72 9.12 12.23 Nutr'oents \kgfhal K Ca Mg S 118.6 82.2 63.7 14.6 94.0 72.8 es.l 13.7 104.0 72.2 57.2 13.4 121.8 60.8 es.o 152 128.6 9&.8 83.0 18.5 mainly through tallow, more nutrients leaving the farm means either a longer fallow recovery penod or lower yields. The total amount of nutrients removed by improved maize varo was significantly greater than in local varo Total biomass productjon and nutrjent removal and distributlon in a cassava¡maize association were studied at CIAT HQ tor traditional and improved maize varo grown as sole crop or jntercropped with a single cassava varo Removal of nutrients by cassava monoculture was higher than cassava in association with. maize (not statistically significant). No influence of the type of maize (traditional or improved) was found en removal of nutrients by the intercropped cassava. Total cassava biomass production was also higher in cassava monoculture than in association with malze (no significant difference due to the maize var.). No difference in total biomass accumulation was found between maize vaneties intercropped with cassava (Table 9.2). Maize varo differed significantly in the amount of biomass and nutrients allocated to the grajns, improved varo allocating more nutrients to the grajns (42% of total) than local varo (21% of total). Improved var. were similar in terms of nutrient and biomass allocation to the grain, but there were statistically significant differences between the traditional var., probably because of theír genetic backgrounds. Associatlon with cassava did not affeet 175 the ability of improved varo to alloeate more resources to the grain. High grain-formation capacity was more evident in hybrids than in open-pollinated varo Hybrids performed well in terms of grain yields when they were intercropped with cassava in a1ternate rows. Intercropping hybrids in the same row with cassava did not result in good maize yield. It is not known whether this is a general characteristics of all clones or specifie to those reported here. Traditional var. alloeated relatively more nutrients to leaves and stems. If leaves and stems are left in the field after harvesting the cobs, reeireulation of more nutrients would be expected with traditionai than improved maize var. A similar situation would be expected if manure from animals fed leaves and stems of traditional varo were returned to the field. Intercropped maize always extracted smaller amounts of nutrients than as a sale crop beeause total biomass and the pereent of eaeh nutrient in the tissues, except Mg, was always less in the intercrop than in the monocrop (Table 9.3). In monoculture improved malze var. tended to use more N than traditional var. In monoculture, traditional maize var. take up more P, K and Ca than improved varo although they use them rather ineffieiently, at least in favorable environments sueh as that of CtAT Ha (Table 9.4). Table 9.3. Nutrient uptake by malze In association with cassava and sole crop; avg of 2 Improvad and 2 tradlllonaJ malze varo al CIAT HO.' Ctopplng Syetem lnIo,croppod s.Ie crop Total biomass 8078 a 10789 b N 117 a 158 b p 32b • Dala loIIowod by lile aame letto, aro nol olgnirlC8nlly differon! (P $ 0.05). Nutrlonts (lI1I. ''"1'. on-farm research is usually conducted with farmers and seeks their active participation in the management of experiments, a follow-up phase in which technology components are adapted to existing production conditions is still required. Agricultural production systems consist of a yearly sequence of avents that begin before the onset of the rainy season and end with postharvest actMties. The testing of new technology components that would everltually replace some of the existing production practices should be done In approx. the sama sequence so that the production system being tested is a mix of traditional and new production practices. The adoption of any technology component by farmers depends strongly on the interaction between new and existing technology. Use of manual labor, particularly in peak demand periods, is a key factor determlning whether a technology component will be adopted by farmers. On the North Ceast of Colombia, several technology components to improve production of cassava and associated crops have been developed by on-farm research teams over the last 8 yr. These components have yet to be fully validated in large plots, under the complete management of farmers, to ascertain their compatibility with current farmer production practices. Selection of these technology components was done 3 yr ago by ICA and CIAT scientists; they were then discussed with a group of farmers to determine if they would be willing to test them on their farms. During the cropping seasons of 1989- 90 and 1990-91, 40 and 36 plots were planted, resp. (Table 9.14). Farmers were instructed on how to apply the new technology, but all aspects of crop management were their responsibility. 187 TabIe 9.14, Componenl. of improved and trsdltional tedlnoIogy lh1ed In pre-ptCd_trIala wilh cassavaand mai .. in_atioA en !he CoIombien North eoasl Improved Ram T .. clmology Son preparation NA' Seed. Maíze Improved varo, Cassava selectlon & treatrnent" Weed control Pree"""1lenc::e herbk:ld .. Inseot contrOl Mal .. LDrsban 2.5"'" Fortiliz.!ion Mal ... 40 kg N¡'l1a4 t Improved technology nol availabl •. 2 Stak .. Ir •• led only !he first year of lrIeI., o Lorsban = chlorpyrifos; hlgh maize plan! denslly usad only duñng VI 2 of _. • Appllad enly If necessaty. T,adltionel T echnology FIIfIT1GfO' practlc::es l<>cal VIIt., selectlon. no Ireatrnent No hetblQldes No control No fertilizar Plots sizes were always larger than 0.1 ha. Two samples (100 m") were taken at harvest time to determine crop yields. Nearby pIots planted by the same or different farmers were previóusly selected as checks. One or two 100 m2 samples were taken from check farmers to assess yields obtained with traditional technology. In the 1989-90 cropping season, maize yields obtained with the improved technology were similar to those obtained by the farmers. Cassava AYs in this cropping season were equal or aboye farmers' yields. A major limitation for comparing levels of technology was the difficulty encountered in sampling farmers' fields. In 1990-91 maize and cassava yields obtained with the improved technology were superior to those obtained with traditional technology. Yields in the second cropping season were higher than those of the first year (Fig. 9.5). Manual labor requirements for the ímproved technology were similar to those of the traditional technology, both in terms of amount and temporal dístributlon. Total production costs of improved technology were 8% aboYe what farmers normally spend to produce cassava and maize (Figs. 9.e & 9.7). 188 MalD Ylelda (tlh) 3.0 __ 1'.'-1"0 1"0- 1.'1 3.0 2.0 2.0 ... 1.0 0.0 0.0 • O • 7 • O • 7 12 2O~C~."~~=.=Y=I="~d~.;(1~/h~';')~====~======~~~~=======-l I ___ 1 •• '-1.'0 __ "'0-1"1 lO 10 O O 4 • • 7 4 O • 7 12 Localities Technology ~ Improved o Tradltlonal Figure 9.5. Ylelds 01 maize and cassava In association with improved and traditional technology on the North Coast 01 Colombia. 189 lmproved Technology 1¡8¡-U~gQ Total Oava Labor. 124 ?1&"lIn" In .L'~"~M.,," n.· ,¡go-l¡¡, Total Caya Labor • 110 Fu...,. T~hnobgy 18aa-l¡¡O Tobd Da~ Labor • 126 _0-,"" Total Daya Labor • 82 Figure 9.6. Hand labor use lar activities in th maize/cassava crapplng system. 190 Improwd Tecnnclo1¡y 1989-1990 ~ ......... WO .•.• d:''''''''''' u.,o Har ........ _ Pr$)uaUon 11'1'0 -$32,355 Farmer', Technology 1989-1990 r.gl"",,~""'" Figure 9.7. Production eoSlS as pareent of total eoSl by activities. 191 lamí PHJo .... tlon U4.~920'1 10. PLANT NUTRITION ANO SOIL MANAGEMENT Over the past 5 yr, the Physiology Seetíon has eonducted researeh on nutrition and soU management at both Santander de Ouilichao (Cauca) and at Media Luna (Magdalena). Researeh focused on (1) evaluation of cassava germplasm for adaptation to soils low in P; (2) long-term effects of continuous cassava cultivation on productivity; (3) evaluation of cassava germplasm under moderate levels of applied fertilizer in poor sandy soils; and (4) the effect of plant mulch on cassava productivity and stabilityon sandy soils. In addition, a collaborative research project to elucidate mechanisms associated with the crop's adaptability to low-P soils was ínitiated in 1988 with a doctoral student from the Swíss Federallnstitute of Technology. Highlights are given here of research findings that shed Iight on the potential of cassava in peor soils as well as the response of the crop to soU fertility management. 10.1 Screenlng Germplasm 10r Adaptatlon to Low·P Soils (Santander de auilichao) In continuation ofefforts to characterize cassava germplasm for adaptation to low-P soils, several advanced breeding lines and clones were screened (4 reps) at the Santander de Ouilichao station from 1986-90. The accessions were planted (10.000 pI/ha) in plots with Iow available P « 4 ppm) that had received no P fertilizer for the last 8 yr. Half the plots received 75 kg P /ha annually (8 yr). Adequate levels of N and K were applied at planting to the whole experimental area. P fertilizer in the form of triple superphosphate was banded around the stakes at planting. Tables 10.1 to 10.3 show the RYs of the best adapted clones at 10 mo afier planting along with the calculated P adaptation indices. At both levels of P, there were significant differencas in AYamong genotypes (P<0.01). The overall avg RY of all genotypes at low P was significantly lower than the avg RY with adequate P in all seasons. Moreover, yields have besn stable at both levels of P ovar the past few yaars. These reasonable yields without P application might indicate that native P uptake by cassava was effective. Cassava is known to be highly depandant upon VAM association for P uptake. It is possible therefore that cassava roots wera effectively colonizad by efficient strains of the fungus. These highly adapted genotypes Qndices > 1.5) includad 2 CIAT clones that were recently released as new commercial varo (CM 523-7 and CG 2177-2, Tables 10.1 & 10.2) and sorne traditional clones from Brazil. The remainder of the screened accessions showed eíther íntermediate Ondices trom 1.49 to 1.0) or low Ondices <1.0) levels of adaptation. Taking into consíderation that cassava was continuously cultivated in this sita tor the last 10 yr, the moderate level of productivity maintained without P application suggests that acid soils with low available P but high QM can support sustainable yield. Nevertheless, the significant response to P fertilizer observed in these tríals indicates that cassava still banefits from applied fertilizer. Those genotypes highly adapted to low-P solls can be used as genetic resources for breeding new materiais. 193 Tablo 10.1. AY and low-P adoptation Indox 01 11 clono. (out 01 71 accesolons) bost adaptod lo Iholow-P soils 01 Santandor do Cuillchao, 1986-87 soasan. Varlety Panamlña CM 489-1 CM 507-:r7 M Van 321 CM 516-7 CM 523-7" M Bra 226 M Bra 41 CM 305-41 CM 975-5 MEcu68 Avg 01 71 accessions , P appllcation In kg P ¡ha • lDw - P adoptation indox z Frosh AY ~¡ha) P , P7!I , o 56.0 57.4 37.3 73.5 36.2 57.3 32.9 62.6 26.4 71.3 25.2 72.6 36.6 45.5 37.5 45.9 34.9 46.7 34.2 46.3 35.6 45.8 26.8 39.3 Vleld Po x Vlold P,. XVleld Po x XVleld P,. 3 Released In 1990 Ior Ihe Eastern Plains under!he nema "lCA-Catumaro." Adaptation Indox lo low p2 3.13 2.60 1.97 1.96 1.79 1.74 1.67 1.63 1.61 1.57 1.55 1.0 Tablol0.2. AY and low-P adoptation indox 01 varieties best adaptad lo Iholow-P soils 01 Santander do Oullichao (1987-88 soasan). Fresh AY ~¡ha) Adaptation Indox Varlety Po P,. IoIowP CG 913-4 43.9 46.4 1.99 CG 2171_2' 31.3 52.2 1.59 CG 1370-5 32.1 49.2 1.54 Avg of 33 aceesolons 27.0 36.0 1.0 , Released In 1990 Ior Ihe Eastern Plains under Ihe namo "lCA-Cebucán." 194 rabie 10.3. AY and low·P adaplatlon Inde. ofvatietio. best adaptad lo tholow-P ",,11. of Santander de Ouiliohao (1989-90 ... ...,n). Dry AY ~/ha) Adaptatlon Indo. Varloti .. Po P715 10 lowP CG996-6 13.1 16.9 2.0$ CM 305-41 13.9 14.6 1.86 M Sra 383 11.9 16.2 1.78 M Bra 191 10.9 16.9 1.67 Avg of 33 8""",",",ns 8.8 12.5 1.0 1 0.2 Ralationships Batween Leaf Gas Exchange, Total Biomass, RY and Adaptation to Low-P Solls Using a portable infrared CO. analyzer, leaf gas exchange (CO. uptake and H.O loss) was measured on upper canopy, exposed, fully expanded leaves of the 33 genotypes screened for low-P adaptation during the 1989-90 season. Measurements were conducted 6 times between 3 and 6 mo after planting with a total of 24 leavas measured/genotype/P treatment when incident sunlight was > 1000 11 mol m"'s in the photosynthetic active range. Gas exchange data were averaged throughout the measurement periad and correlated with RY and total biomass at 10 mo after plánting. Low-P adaptation index and some other growth parameters were also included in calculating simple correlations. Among this group of genotypes, leaf photosynthesis was significantly correlated with RY, total biomass, top wt, storage root no. and low-P adaptation index (Table 10.4). Mesophyll but nol stomatal conductance was also significantly correlated with these yield and growth parameters. These patterns of correlations confirm earlier findings at CIAT that a direct and positive retatlon exists between single leaf photosynthesis and yield of cassava (CIAT Annual Reports 1988, 1989). Another important finding is the significant positive associatíon between RYand both top wt and storage root no. It appears that when HI is high On this group of genotypes, Hls were > 0.6), improvement in yield could be achieved by improving the crop's photosynthetic capaeity through higher leaf area and higher leaf photosynthetic rate (carbon assimilation souree) and by increasing storage root no. (sink strength). Adaptation to low-P soils also appears to be directly related to both the assimilation capacity and storage root no. High assimílation capacity and high sink strength would probably lead to higher Puse efficiency (PUE), particularly in soils low in P. 195 Table 10.4. Co"elation "".mol.nto _ •• n le.I g ... e.changs eharact<>ristlcs. RY. bioma •• and 10w·P adaptatlon Inde. 01 33 clone •• 1989-90 5& •• on.' .... lilal photosynlhosia lila! conduotance (H2Q) Mesophy1l Conduelan"" (C02) ~ DryRY Top DryW! ToIal Bioman Sloraga Roo! No. 1.Dw-P Adaptallan Ind •• 1 Mean 01 Ihe two P levela. NS = No! sign_ a' 5"- • • Slgnifocant al 5$ •• - SlQnHIcant at ,"- Leal Conduotance (H20) 0.81- Mesophyll Conduotance (C02) 0.97" 0.69 .* Dry RY O.4S- 0.29 NS 0.53** lop Dry W! 0.39 • 0.25 NS 0.40 • 0.58 •• l.Dw.p Tola! Storage Adaptallon Biomau RooI No. Indo. 0.49 .. 0.37 • 0.51 _. 0.31 NS 0.11 NS 0.30 NS 0.56 .* 0.49-- 0.57 *. 0.96- 0.112 .. O.99 h 0.76** lUlO .. 0.58 ** 0,65 ** 0.96 .. 0.67 .* HI 0.12 NS 0.03 NS 0.ll8 N.S. O.33NS 0.52 •• 0.09 NS 0.07 NS MINa 10.3 Long-Term Response to NPK Fertillzer In Infertlle Acid Soils (Santander de Qullichao) For the last 8 yr cassava was contínuously grown on infertile acíd soils at Santander da Ouilichao to assess the effeet of this parmanent cultivation system on cassava productivity and íts response to applied fertilizer. Two cassava clones (M Col 1684 and CM 91-3) wera plantad annually at the same sites. The fertilizer treatments (4 reps) consisted of 3 levels of NPK (Le., O, 50 and 100 kg/ha each of NPK). Each of the three elements was also varied índependent!y at 3 levels of O, 50, 100 kg/ha whíle the other two elements were kept constant at 100 kg/ha. Treatments were allocated at random in a complete randomized block design within the site. AlI fertilizer treatments were applied at planting. Harvest was always condueted at 11 mo afler planting. Figure 10.1 iIIustrates the long-term response of cassava to NPK. It is elear that cassava is highly responsive to K, particularly when the soíl is poor in this element; on the other hand, there was little response to N and P. Moreover, in the absence of adequate K levels, no benefit is achieved by applying high levels of N and P. In cassava a large portian of absorbed K (> 60%) is removed with the harvestable roots; whereas significant amounts of absorbed N and P are recycled 10 the soil through fallen leaves and crop rasidua. It is known that a crop of cassava can return to the soil trom 3 to 6 t of dry leaves during its growth cyele of 10-12 mo (see Chapo 3, sec.1). In addition to tha nativa OM in the soil, thasa relativaly larga amounts of crop residue can serve as a source of nutrients. In conclusion, it can be stated that cassava produetivity could be maintained at a reasonable level in acid soils high in OM, provided that moderate levels of K fertillzer are applied to compensate partially for the soil K removed in the harvestable roots. Howevar, when soils are poor in OM or in sandy soils, other nutrients such as N and P would IImit productivity. 10.4 Response to NPK Fertlllzer in Sandy Soil (Media Luna) in contrast with the Quilichao soils, the Media Luna soils are sandy with extremely low OM and nutrient contents (CIAT Annual Aeport, 1988). Yields of cassava in that region declined rapidly in the last few years. Therefore, field trials were initiated two years ago to evaluate yield response 10 moderate levels of NPK fertilizer. Trials were condueted on a private farm using 13 cassava clones, among which were were local varieties as well as some CIAT advanced clones. Split applications of fertilizer (50-20-43 kg/ha NPK = 330 kg of 15-15-15 compound fertilizer) were made at 30 and 60 DAP. Table 10.5 shows data of dry RY and biomass production tor the 1989·91 seasons. The avg increases tor all clones due to fertilizer application were 88%,140% and 109% for dry AY, top growth and total biomass, resp. This indicates that by continuously growing 197 '" el o KgNma • 50 • 100 O~~-L~ __ ~-L~ .. - '.> Q KgPi'I'Ja • 50 • 100 CM 9103 2: 3 .. S 6 Ve .. Figure 10.1. Long-term response of cassava 10 NPK fertllizer in a low-fertllity soll al Santander de QuiUchao, Cauca. cassava on this peor sandy soU, productivity will decline and the need to fertilize the soU will become more crucial. Besides the substantial gains in yield, top growth was also greatly enhanced. This is of a paramount impertance to cassava growers as sufficient good-quality stakes are essential. The cost of added fertilizer is insignificant compared to the large gains in yield and the ensured planting materíals. There is also a possible reduction in weeding costs as fertilizer enhances cassava growth and reduces weed pepulations. 198 Table 10.5. Respon .. of casaav. 10 NPK fer1lYz ... ln lile aandy 8011 of_ luna (Magdalena, CoIombla); avg of tila 1981H11 growlng ... oons • . - F!!!lIIl:!I!I UnfertlUzed VetIeIl •• - T_ TOI8I - TOe! Total Cry AY (I/IIa) 00 12:20-2 5.9 6.2 12.1 2.3 1.8 4.1 00 1365-2 8.2 4.3 12.$ U 3.0 8.2 00 1372-5 6.1 6.1 12.8 2.8 2.8 5.4 00 1411-1 4.8 4.6 9.2 1.9 1.3 3.2 00912-8 8.3 5.9 12.2 3.3 2.5 5.8 CM 332Q.4 8.3 4.2 10.$ 2.7 1.7 4.4 ..... CM 4181-1 7.0 5.0 12.0 3.1 1.9 5.0 ~ CM50N!7 5.0 M 11,(1 3.3 1.7 5.0 M era 191 8.9 5.1 12.0 4.3 3.0 7.3 M era 383 5.5 5.4 10.9 3.2 2.3 5.5 M CoI1SOS' 4.4 6,(1 10.4 2.7 2.8 5.5 M Col 2215' 5.1 4.3 9.4 3.6 2,(1 5.5 M Q!l! 22111' 5.ll 6.2 11.7 2.8 2.8 5.4 Avg 8.0 S.3 11.3 3.2 2.2 5.4 %Incr .... 88 140 lOS LSO 5% 118(. 0.9 0.8 1.5 Tr.almenta 0.53 0.7 1.13 I iJ:ocaI_. 10.5 Soll Management Svstems at Media Luna A 3-yr trial was conducted on a prívate farm at Madia Luna (Magdalena) during 1988·91 to determine whether cassava productivity could be improved in poor sandy soil with fertilizer and¡or plant mulch in combinanon with the tillage system. In general the soil is poor in OM and in nutrients (P and K $; criticallevels), illustrating the very Iow fertility level typical of this cassava-growing area. The local varo M Col 1505 was used as a test material after selecting the planting stakes from a well-fertilized field. The treatments consistad of two tillage systems Q.e., zero tillage and conventional tillage), two fertilizer levels Q.e., no fertilization and 330 kg¡ha of the compound fertilizer 15:15:15 in a split application at 3Cl and 60 OAP). Mulch was applied at arate of 10-12 tfha green material ( .. 3-4 tfha OM) at planting. The source of mulch consisted of weeds, grasses and crop residue. Planting was done after the onset of the rainy season in May;harvesting, 11 mo afterward .. Figure 10.2 illustrates yield responses in the three consecutive seasons. Irrespective of tillage system, there were strong responses to fertilizer and mulch applicatlon with cumulative effects during the 3 yr. Mulch greatly enhanced RY without fartirlzer application, irrespective of tillage method. The effect of tillage was more apparent in the first year when combined with mulch. In absence of both fertilizar and mulch, RY remainad stable at a Iow level around 8-10 tfha fresh root. As an avg of al years, RYs were greatly enhanced by either applying mulch or fertilizer (Fig. 10.3), with the greatest response to mulch occurring without fertilizer. These data indicate that cassava productivity in this poor sandy soil is greatly enhanced by appIying either chemicat' fertilizer or mulch. The latter appears to be beneficia! in improving the physicochemical properties of the soil; moreover, it can a11eviate water stress in thase sandy soils by reducjng water evaporation from the surface soiI normally exposed to high temp (Fig. 10.4). Reducing evaporation from soil is of a peramount importance as sandy soils are characterized by a low water retention capacity. Another advantage of mulching is the great reduction in HCN content of cassava roots in the absence of fertilizer (Fig. 10.5). In regions where cassava is normally grown in marginal poor soíls coupled with prolongad drought such as NE Brazil and Sub-Saharan Africa, application of mulch, if available, could be beneflcial for cassava production. Altematively, cassava may be grown in rotation with other field crops that yield sufflcient residue to be usad as mulch. In cases whare feIIow can be practiced, natural vegetation may be used tor mulching. On hillsides, where cassava is grown in degradad lands, mulching would be advantageous te minimize soil ereslon as well as improve productivity. It is therefore warranted te evaIuate this management system on steep lands combined with other practicas effective In controlling soil eresion such as live barriers, which could be used as source for mulch (Sea Chapo 11). 200 Figure 10.2. • tO _ ... ______ ... ________ • ! • l"'g. ........- ª • í 25 ,. '. ,. ------.-------.... • --0;-, ----!-. ----;. • , V"'" Responses of cassava 10 fertilizar appllcatlon (330 kg/ha of 15-15·15 NPK). surface mulch and tlllage sySlem on sandy soll at Media luna (Magdalena). Note the cumulatlve posItlve effect t:Ner y4!8rs of both mulch and chemical fertilizer. _ Fortized 21.0 m Uftlerlilized Figure 10.3. Three-year (1988-91) avg response of cassava to chemlcal fertilizer, mulch and tillage system on the sandy soU al Media Luna (Magdalena). 201 3 .0)- i ~ 1-ToII 2.TiI+~ 37 - 3. No TilI 4 .. No nu + MuIch l 3 ti 2 • • 2 34 - • 11 f • ~ , 3 2 "'r • 2 • .1 ~ 3 • 9 •• '5 '. Figure 10.4. SoR temp at Media Luna (Magdalena). as affected by surfaca mulch and lIIaga system. Note lha Iarge decrease In sol temp une!er mulch at midday. i 11 ~ j z u ::r: ~ 300 250 200 150 100 50 O i! o z _ Fertilized ¡.<~·.'l Unferiilized Agura 10.5. Total HCN in cassava roots at 11 mo afler planting as affected by chemicaI fartlizar, mulch ane! tBlaga systam. Note the larga decrease In HCN under mulch in the absence of chemlcal fartlizer. 202 10.6 Varletal Dlfferences in Response to P Trials conducted previously in Santander de Ouilichao indicated a wide range of response for adaptation to low P-soils in cassaya germplasm (CIAT Annual Reports, 1986-91). Varieties well adapted to low P-soils have good yield at both low and adequate P supply. The objectives of this study were to identífy mechanisms involved in varietal response to P and to determine the relative importan ce of P uptake (P acquisition) vs. internal use. Through a coIlaborative project with the Swiss Federallnstitute of Technology, a 2-yr. trial (1988-90) was conducted on a private farm in Santander de Ouilichao. The soil-P rever at the experimental site was about 2.5 ppm « the criticallevel). The experiment was laid out in a split-plot design with three reps. The following fertilizer treatments were assigned to the majn plots: (1) unfertílized; (2) 100-0-100 kg/ha NP; (3) 100-SO-100 kg/ha NPK; and (4) 100-100-100 kg/ha NPK. In the subplots, 4 cassava varo were planted (10,000 pI/ha): CM 523-7 released in 1990 as "ICA Catumare: CM 489-1, M Col 1684 and CMC 40. Sequentíal harvests of B protected pl/plot were made every 2 mo. Plants were separated into roots, stakes, stems, young leaves, mature leaves, fallen leaves, patiales, flowers and fruits in order to determine fresh and dry wt and P contento Total P uptake was estímated as the sum of P accumulation in the different plant parts. For fine root density determination, 12 soil samples/plot were taken at the mid- distance between 2 plants. A hand auger (433 cm3) was used to sample within the 0-20 cm soil layer. Rootlets were separated from soil by flotation, and root length was estimated using the grid fine method. 10.6.1 Vleld response Figure 10.6 shows dry RY at final harvest for both crop cycles. The two most contrasting clones, M Col 1684 and CM 489-1, maintained different responses; Le., significant response to P application in CM 489-1 and a lack of significant response in M Col 1684. Under low P, however, these 2 clones had similar RYs in both years. 10.6.2 P uptake Figure 10.7 iIIustrates the total P uptake over time as affected by fertilizer levels. AII varieties accumulated more P at high P levels (SO and 100 kg P/ha) than the control. Between 2 and 8 mo after planting, P uptake was nearfy linear in all clones and P treatrnents (Fig. 10.8). In all clones, the P uptake rates were higher with higher P supply. At alllevels of P supply, the ranking order among clones was CM 523-7 > CM 489-1 > M Col 1684 > CMC 40. The same pattern of ranking was observed with fine root length density (RLO) (Fig. 10.9), suggesting that RLO is 01 a paramount importance in P uptake. This is further iIIustrated by Figure 10.10 where P uptake rate was plotted as a function of RLO. At both low and high P supply, uptake rates increased with increased RLO. At a given RLO, all clones absorbed more P at high P supply. It appears therefore that the 203 30 -- 0'990 ,.89 "",- LSO '%¡ CM523-7 CM 489-1 MCOL 1684 (19M} '"'~ .. f CMC<40 l " t. § tSDS'4 t -, I 1'990) 1 lO~ si o 1 2: 3 4 1 2: 3 " 1 2: 3 • l 2: 3 ",- Figure 10,6. Responses 01 cassava to P fertlllzer In low-P soll at Santander da QuUichao. (1) No fartilizar; (2), (3) and (4): zero, 50 and 100 kg/ha P wllh 100 kg/ha N and K .. CM$2j.7 " " " " " i • i " % WCd1&N . " 30:- 1$8$ ·to " " " " Figure 10,7, Total P uptake al different P lertilizar levels as a functíon 01 time alter planting. (0,0) No fertilizer; h 1, (., j, (',10,50,100 kg P/ha with 100 kg/ha N and K 204 I l!l e i ... 5 4 ... 3 f- 2 i- f- o o Okgf'/ha _ 50kgP/ha e 100 kg f'/ha , T '--- CM 523-7 Cm 489-1 M Col 684 CMe 40 FigunlfO.8. P uptake rates (2-yr avg) at dlfferent P fertUlzer levela 1 - 0.8 g ~ 0.6 -b 'í?! Q) "C 0.4 ) J CM 523·7 CM 489·1 M Col 1684 eMe 40 Figure 10.9. Flr;6 RLD 01 4 cassava clones (2-yr avg 01 all P levels). Note the same ranking among clones as with P uptake rates (Fig. 10.8). 205 5 4 R'=O.93 f 10%, a significant portion of cassava production occurs in hillside regions with greater slopes. Since 1979 tha Cassava Program has put major emphasis on investigating soíl erosion in cassava-based cropping systems, as well as on production management systems effective in reducing son erosion while maintaining productivity (CIAT Annual Reports, 1982-86). To strengthen this research effort, a collaborative research project was ¡nitiatad with the U. of Hohenheim, Institute far Plant Production in the Trapies and SUbtropics, Garmany, in 1986. Two doctoral stooents and, more recently, a postdoctoral researcher haya collaborated with the Physiology Section in executing this project. The project has focused on (a) collecting fundamental data on son eroslon in the Inceptisols to assess long-term affects on soil degradation and (b) evaluating different cultural practicas in relation to soil erosion and productivity. The trials were established at 2 sites: The CIAT Experiment Station in Santander de Quilichao (alt. 1000 m) with an avg slope of 10-15%, and a private farm in Mondomo, Cauca (alt. 1450 m) with an avg slope of 15-20%. The soils (Oxic Dystropept and Oxic Humitropept, resp.) are acidic, high in Al saturation (50- 85%) and low in nutrients in the thin top soil (15-25 cm). Basad on the results of this on-station research, the project is currently extending its activities to farmers' fields, in close cooperation with local institutions and organizations operating in the northern parts of the state (Cauca). Collaboration with local tarmer groups and local institutions will constitute a besic principie of further research done by CIAT's CassavajSoil Conservation Program in order to: .. Bridge the gap betwean research and practice .. Assure that on-station field research meets the necessities ot local farmers .. Accelerate the application of adequate 5011 conservation practices on the farmers' fields .. Achieve a more sustainable impact on farm practices able to reduce erosion risks In this report some findings are presented that indicate the potential of soil erosion on hillsides and the effects of different crop management systems on soilloss and cassava production. 209 11.1 1987-1989 Trials 11.1.1 Soillosses and cassava productlvity Total annual soillosses as affected by crop management systems are presented in Table 11.1. Monthly cumulative soil losses are shown in Figure 11.1. lhe overall avg annual soíl losses from the bare plots were > 100 tfha, which shows the high erosíon potential of these soils. Depending on the amount and intensity of the rainfall (Fig.11.2) soíllosses can exceed 300 tfhafyr (Table 11.1). At both Ouilichao and Mondomo, the majar portian of soUlosses occurred between Oct. to Dec. when ralnfall was highest (Figs. 11.1 & 11.2). The impact of highly intensiva rainfall is better iIIustrated by the estimated kinetic energy of rainfall (Fig. 11.3)--the largest amount of raínfall beíng assocíated with the highest kinetic energy value. Thís informatlon has important practical implications for soU conservation on hillsides where cassava often ís growo. It is likely that soUlosses will be large when planting coincides with intensive ralny periods due to the fact that cassava canopy development is slow during the first 3 mo; through proper management, however, soil losses can be minimized. Compared with the traditional cassava cultívation on flat lands, growing cassava in contour ridges or with grass barriers has led to much Iower soíl Iosses. On the other hand, growing cassava in down-slope ridges or in associatíon with early-maturing grain legumes resulted in much higher soilloSses than the traditional practice (Table 11.1). Too much soil disturbance might have occurred during the planting, weeding and harvesting of the legume crop. Although minimum tiUage (preparing only holes for cassava stakes) greatly reduced soU eros ion, cassava productivity in Ouilichao was much reduced (Table 11.2), probably due to competition from the natural weed vegetation and because of soil compaction. . Other mahagement practices, however, did not show large variations in productivity. It appears therefore that growing cassava either in contour ridges or in association with live barríers is not only effective in reducing soíl losses but also in maintaining cassava productivity. The choice of a specífic management practica-- particularly with respect to its socioeconomic implications--would probably depend on the farmer's conditions. Nevertheless, the long-term benefits of controlling soíl erosion and henoe maintaining soil fertility (Table 11.3) should be assessed against the extra costs required for son conservation measures. It is therefore warranted to evaluate these management technologies in partidpation with farmers, taking into account their traditional practicas in an integrated farming system. 11.1.2 Mlnimum estlmates 10r runoff Uke son losses, runoff was measured after erosive raln storms on a regular basis using permanent water col1ectors. Given the fact that the water collecting system did not always function properly, particularly during intensive rain storms, the runoff volumes measured in these trials should be taken as minimal estimates (Table 11.4); higher volumes of runoff are likely to occur on larger slopes. Nevertheless, these mínimum estimates of runoff 210 Table 11.1. Sol! _. during 11187-89 ctopping 18_ al Ouillchao and Mondornc (t dry soiljha). Ouilichao Treatment 11187/88 1988/89 Bate Fallow 49 197 CUsava/F1al S 17 C&ssáva/Contour-ridgH 4 8.S CUsava/D2wMl9pe 31 88 ridges Cassava/Cawpeas 8 29 Cassava/Grass sIrips 5 15 CassavafMln. tillage 2 2 '¡- McndOmo Mean 0U111ch .... '~1187/88 123 11 6.3 49.5 18.5 10 2 ---A~ .. AII 45 10 1.4 34 20 1.5 3 .. CasoHv*.~ ~ o CáúaVlL mg.doWn -iJ .!IttI" ....... '" Canav$~ 1- ~graastnpt Jr.~ 1988/89 311 -40 2.6 13 47 3 2.5 Mean Mondomo 178 25 2 23.5 33.5 2.:' :2.3 2.; 2.8 FIgure 11.1. Cumulatlve soilloss at Mondomo and OuBichao as alfected by sorne caasava- based cropping systems (2-yr avg). 211 . , 600 1 5001 400~ i MondQm(I o 1987·88 .1988-89 • Long-tem'l Figure 11.2. Monthly ralnfaJl at Mondomo and OuUlchao durlng !he 1987-89 seasons, along wlth long-term ralnlall pattems. show that growing cassava in down-slope ridges results in larger volumes of runoff as well as soíllosses (Table 11.1). The two management practices~-í.e., growing cassava in contour ridges or with Uve barrlers-were effective in reducing runoff and consequently soil losses. These findings indicate the importance of crop management systems in conserving both soíl and water. Taklng into account that cassava is a long-season crop (8-18 mol, conserving water by reducing runoff during the rainy season should reftect favorably on productivity when the crop has to endure long periods (2-3 mol with sporadic or no ralntall. The comblned effect of reducing soillosses and runoff would lead to improved soil water status. 212 12"-1 --------_ 1 11 L- Mondomo '0r- 01986-97 9¡- .1967,sa 8~ .1988-89 ~ O'--~-L_~~~~~~~~_L-~ ]¡- i~ 12r---------=-----~ r 11 Quük:haQ J Figure 11.3. Kinetlc energy of ralnlall at Mondomo and Quilichao, 1986-89. 11.1.3 Predlction 01 soillosses uslng the Universal Soil Loss Equatlon (USLE) As stated, one 01 the objectives of this research effort was to collect fundamental information on the characteristics of rainfall, soUs and crops in order to test the applicability of the USLE for predicting soU erosion in the Andesn regions of the humid tropics. Originally developed for temperate zones, the validity of the USLE needs to be tested in the tropics, where limited information exists. More data on a longer term basis are required before the model's utility can be judged). Short-term results (2 yr) were inconsistent in relation to actual and predicted soillosses. Only in the second year, were there significant correlations between the erosivity indices utilized in the USlE and the 213 " 1: Tabl. 11.2. Cassava ""oh RY in .",sien lríals at Cl\IllicI!aa and Mondomo (I/hA). Cl\Il1i1:hao _mo CM 523-7 M Col 1522 Trealmenl (3-yr 8"11) (198U9) Cassava·Aal 31.1 19.7 Cassava-Contout ridges 29.7 15.3 Cassava-Down·slope rldges 'Z1.7 15.4 Caaaava.cowpus 21.3 16.8 CUsa.va-Grasa atrIps 'Z1.1 18.2 Caaaava·Mln. tilla98 9.0 15.7 LSD5% 3.7 2.9 TabI. 11.3. Ñlnual _. of CM and of somo nulriento In eroded 0011.; avg of _ """,plng .. asona (1987-89) at __ (. SO). CM Mg K P Trea1ment (lJl!a) kg/ha Bate fallow 9.0 * '1.5 1.lho.n 8.2.3.9 0.6.0.4 Cassava·Aat 1.2.0.9 1.2.0.86 1.7. 1.6 0.23.0.2 Cassava-ConlOur 0.24.0.2 0.3.0.2 0.39.0.2 0.08 t 0.08 rI 25 mm/h Intsnsity 1,$ • max. 1~1n.lntontlly o! an orost ... BIOIm El,. • produCl of Idnatic _rgy and 1,. 1m • m .... 7.~ln. Intentlly of an orosive BIorm Elm « ptOduCI of Idnetic anorgy and 1m Al.., - ptOduCl ot amount of erosivo ralntall and 1.., TabI. 11.6. Cusava dry RY under dlfferen! cropplng ays\emI al Santander de QuHiQhao. 1987-89 .......,n .. l/nfertillzod Fertillzod Ctopping Syatem CM 507-37 M Col 1684 CM 507-37 M Col 1684 Ory RY ~fha) Cusava. _. rOlll1Ollfld 7.2 5.8 18.0 10.5 Ctsoava + mulch (weod.) 11.0 9.1 19.0 13.6 Ctsoava + zomIa 6.9 5.6 15.3 8.3 Caaatva + kudzu 8.5 3.2 12.2 4.9 Ctsoava + _OIIema 6.7 2.0 14.2 6.8 Caaatva + skaIro 8.3 2.3 11.5 6.2 Caaatva + _Ium 4.1 2.0 10.7 5.7 Cusava + Arach/$ pInfoi . 5.2 2.3 12.5 8.1 LSO 5% (!lar. X Syatem) LSD 5'" (!IlII. X Sy$Iem) Unfet1Itizod - 1.4 Fertilizad - 2.1 Table 11.7. AMual dry aoII _ (I/ha) and runo!! (mm) In dlfIerent ..- cropping ays\emI 1990-91. CluHiQhao Mondomo Runoll Runoff T_ SoiII.c .... (10 mol SoII I.caeea (8mo) !!are Fallow 142 144 228 123 Cossava-Rat 8.3 48 12.8 50 Ca_va-Contour ridgat 3.1 44 2.0 37 Ctsoava·Kudzu (No.99OOl 15.4 43 Cassava-Zomia (No. 8283) 2TA 50 6.9 78 Caaatva-CentmsemallClJlilbllum (No. 5568) 12.8 38 2.5 Ctsoava- Cet>tmoemo mac""""l'um (No. 5740) 5.5 56 Caaatva-Elephant g .... strIps 4.0 50 1.0 49 Ctsoava·Vetiver g'aos strip. 1.3 45 4.7 53 217 • Cassava-Flat A~rridgn o Cassava-__ (8VQ o, all specilS) • Casnva V.tilo'$( grass • Cassava~E.IecIhant graa 'l \Aoodomo :2 'O~ a~ t i ;r 1 O. c - -o 0_ I auiliChao ! lS ,+ o e O O e "......-" o -~ ,. 10~ 5 ~ Juno iIo,Jg. Oc!. _ . .... '1190 II1I>1 flat planting > bare plot (Fig. 11.6). It decreased with increasing soilloss, time of soil exposure, no. of weedings and tillage operations. In the contour ridges, mixing of top and subsoil during field bed preparation has probablY led to an aggregate size distribution similar to the original soil. Proportion of sizes < 0.25 mm, induding grains 01 the most erodible size fractions, .is about 18.5% in the 4-yr-old bare plot and in flat planting. In contour ridges and in original soil, thepercent of these small-sized grains is much less. 11.6 . Conclusions In conclusion it may be stated t/:lat soil losses in the traditional flat planting system of cassava on hillsides are aboye the tolerance limit. Other practices such as growing cassava in contour ridges or in association with live barriers are more ,effactive in reducing soil loss and in maintaining cassava productivity. In view 01 CIAT's recent majar emphasis on natural resources management and on reducing environmental degradation, soil conservatlon and fertility maintenance should be integrated into the farming system as a whole, taking into consideration both short- and long-term consequences. The current efforts in the Cassava Program will continue in collaboration with Hohenheim U. and with several national institutions. rabie 11.9. Changos In sorne ehemlcal propertles of suñace soll In dlfferen! cropping $)'$IIImo al SantsOOer de Oulllchao. 1987·91. Accum. Soll Loss P Yoar rrealmenl l/ha pH %OM ppm K Ca Mg />J Sal. meq/l00 9 lIOÍl 1987 Original lIOÍl O 4.6 7.3 1.4 0.16 0.85 0.3 75 1991 Sare soU1 553 4.0 5.4 1.0 0.07 0.21 0.05 93 1991 Cassava·Aat" 34 4.4 7.2 10.8 0.13 1.76 0.13 53 1991 Cassava-Contour" 15 4.2 7.4 17.7 0.18 1.69 0.62 51 1 Unfertilized. 2 Urna (500 kg/ha) and fertilize, applled. 221 Aggregate Stabllity in Ouilichao (Voder 1938, modifled) 11m Virgin Soil Q Cassava- Flat ~ Cassava-Countour 1991 ~ Bara Plot 1991 --- ---- --- -----------~----- - - ------- ----- I---------------------------------------------~ --- ---------- --------------- -----.----.---.----1 Aggregate size Agure 11.6. Soll-aggregate stabllity al Ouillchao as affected by soma cassava cropping practlcas_ 222 12. PROCESS ANO PROOUCT OEVELOPMENT Research in the area of process and product development over the last decade has made an important oontribution to the Cassava Program's demand lead strategy. This research was justified for the following reasons: • Stagnation of existing cassava markets has reducad the possíbility of adoption of improved productíon technology and of welfare improvement tor small farmers. • New or improved cassava products have high potential for linkíng farmers to growth markets and provide opportunities for rural inoome generation. • There were no alternative instítutions conducting research and extension activities in this area, especially in Latin America. • CIAT has a comparativa advantage in linking procassing and marketing research with cassava production actMties in integrated éassava projects. 12.1 Objectlve To create new markets tor cassava through the development of novel or improved cassava-based products, with an emphasis on small-scale processes suitable tor generating income for smaJl cassava farmers. 12.2 Methodologles As a direct result of the R&D experiences gained during the 19805. a methodology for process and product development has evolved. The methodology consists of four stages: .. Identification of product opportunities through idea generation and selection .. Aesearch under controlled conditions to develop the experimental process and product, and to define market potential further .. Pilot-scale testing of procass, product and market under real conditions and under the management of small farmer groups .. Commercial expansion and replication of the pilot experienca Collaboration with other research institutes in the research or experimental phase of projects has been a feature of CIA T's approach: national expertise and Interest In cassava product development has thus expanded during the decade. The pilot phase is always 223 carried out in Colombia, within the context of integrated cassava projects. TIle active participation of the Integrated Rural Development Fund (DRI) of the Colombian Ministry of Agriculture has been instrumental in facilitating the coordination of activities with a range of Colombian agencies (e.g., CORFAS, SENA, ICA, INCORA), whose participation is essential to ensure success. TIle role of the Cassava Program is therefore greatly reduced once the commercial or expansion phase is initiated although the absence of a Colombian institution responsibfe for research and technical assistance in cassava postharv!3st utilization activities has been a major Iimitation. This chapter will cover the tirst 3 phases of the product development process. TIle commercial or expansion phase will be reported in the Colombia country report (Chap. 14), and an pifot and commercial activities in other countries will be reportad in their respective chapters. . 12.3 Prloritles A wide range of potentially viable cassava-basE¡ld products can be listed a priori. The selection of priority products for research at CIAT was necessary in order to tocus the Iimited resources avaiJable on research with the greatest potential benefits within a reasonable timeframe. Cassava products can be divided into primary or intermediate products (e.g., flour, starch, chips) and derived or secondary products made from the intermediate ones (á.g., high fructose syrups, animal feed concentrates, extruded snack foods). Product development has tocused on the former. TIle four primary products that were initially identified at CIAT were: • cassava chips for animal feed • fresh, conserved cassava .. cassava flour far-human consumption • cassava starch It was decided not to focus on "farinha,' important in Brazil, ar an cassava leaves for animal feed. Table 12.1 shows the type infarmation used to decide the research priorities for activities during the 1980s. Dried cassava for animal feed was an obvious tirst choice for action, with a simple low- cost technology already developed and tested, and good market potentia! in Colombia and other Latin American countries. TIla biggest constraint was tha weakness of farmer organizations capable of managing and operating small-scale processing cooperatives in cassava-producing areas. Cassava flour tor human consumption was seen as an attractive proposition, although with a more uncertain demand due to the novelty of the producto An efficient, sma"-scale process to obtain high-quality cassava flour was needed to be developed. TIle potential 224 Tabla 12.1. eheeldi., 01 research priofiUe. in cassava proceso and product _opment, as defined early 1900S. """""- ~-, "- """"""'" CassaYa $taren (AnImal FHd) (Human Cer.1l.lr(>plkm); ""'- -~"II Oemand oitualion +++ ++1 ++ 1 Raw material supply +++ +++ ++ +++ Physíeal environment (*) +++ ++ +++ Organizatiooal fador. + + + + Existing activity no no no yos Consumer /oIlonl ++1 +1 1 + ."""ptance Capital neod. low medlum low medium Labor availablllty +++ +++ +++ +++ Technofogy availability • +++ + + +++ Benefil distribulion Fat'I'MI'S + latldless FáttMI'S ... JanclIess Fatmef'I .. tuJaI/UfbcI'¡ 7 ,...,... ........ Iaho< OeoIslon FIlMaren ¡:H'OICeS$ & - ....... & On hoId PToceed. plld ~ ........ ....... Notes: + + +, + +, +, • ",Iati"" advanlaga or disadvantage of each product. ? t .... "'" required 10 resolv. doubls or supply Informalion. (*) Availabillty of water; oIlmatic factors, etc. of fresh conserved cassava to alter radically the prevailing market decline of the fresh product in urban situations had been demonstrated. Consumer acceptability of storad roots was unknown, however; and further technology development was requirad. The market potential tor cassava starch was unknown and dependent upon process improvements aimed at greater efficiency and a higher quality product, Environmental considerations--namely, the requirement tor large volumes of water and the pollution hazards generated thereby--also contributed to the decision not to dedicate core funds to this producto In the late 1980s special funding from CEEMAT (CIRAD-France) permitted research to initiate in this area. 12.4 Orlad Cassava tor Animal Feed 12.4.1 Background During the 19705 CIAT carried out basic research on cassava drying; and a chipping machine was designed, adapted trom a Thai model. In 1980 economic studies demonstrated the feasibility of using dried cassava to replace imported sorghum in Colombia. A pilot plant was constructed under a DRIjCIAT agreement, with funding from the Canadian International Development Agency (CIDA). The plant, whích was operated 225 and managed by a small farmer cooperatíve in Betulía (Sucre), rapidly provad the technical, market, economic and social feasibilíty of the project. Since 1984 ORI has been coordinating the expansion of the dried cassava industry to a large number of smal1 farmer groups in the Atlantic coast regían. CIATs role was gradually reduced, terminatíng in 1989 (see Chapo 14 for ínformation on expansion of this agroindustry). The success of the dried cassava project in the Atlantic coast region of Colombia has in recent years producad feedback on new research activities needed to salve certain problems taced by the farmer groups. Since 1987 the following research activities have been carried out in response to this demand: 12.4.2 Equipment improvement 12.4.2.1 Root chipper. The expansion of the dried cassava industry in Colombia and elsewhere led to a demand for a chipping machine with increased capacíty. An improved chipping machine was designed in collaboration with UNIVALLE, Cali and field testad. The cutting surface was imprQvad by replacing the perforated disc with 8 interchangeable blades with a trapezoidal cross-section, mounted on a disco The interchangeable nature of the blades increased the useful life of the discs, and chip geometry was also more uniformo Tha new machina was field tastad on the Atlantic Coast of Colombia in 4 ditferent drying plants. Over 20 trials, the throughput of the machine varíed from 8.2 to Ü!.3 tjh (mean of 10.0 tjh)--an increase of over 100% compared with the prevíous model. Other advantages of thís new machine include shorter drying timas and reducad transport costs because of greater chip uniformíty and densíty when packed. This chipping machina has now become standard on the Atlantic Coast, and the plans have been distributed to several other countries. 12.4.2.2 Pedal-operated chippar. In several regions of Colombia and in other countries, farmers are too disperse to make associative forms of drying teasible. A small-scale chipper was therefore designed for use by individual farmers for farm-Ievel drying 01 chips. A pedal-operated machine was built and tested, based on a Filipino designo Improvements were made to the seat, pedal and handle positions; diso and blade dimensions were also altered to optimize the use of force employed. The disc (55-cm diameter) has 8 blades arranged radially. Chip dimensions are 6Ox7x5mm. Two operators taking 1O-min turns can chip 400 kg roots/h. The cost of the machine ís approx. US$235. The machine has been evaluated by 4 small tarmer groups, where it has proved useful not only in facilitating individual drying of small amounts of cassava, but also to groups involved in fresh cassava storage, who need to process small volumes of reject roots. Some groups have adapted small motors to drive it. It is now commercially available in Colombia and can be marketed through coops. 12.4.2.3 Moisture conten. (MC) determination. Oried cassava chips are accaptable to industrial clients (animal feed companies) once the MC has been reducad to below 14%. Higher values carry risks of microbial contamination and reduced storage times. Lower 226 MC represents an economic cost to the drying coops, not only because it reduces the total volume sold but also results in fonger drying times. It is thus important for coop members to be able to estimate accurately and rapidly the MC of the chips as they are drying to determine when drying can cease. At present subjective measures such as chip appearance and texture are used. A more accurate and objective method was sought; however, it was found that no standard lab method existed against which to calibrate any rapid method. Animal feed companies use oven drying at over 100·C; whereas at CfAT longer drying times at SO·C are used for lab determinations. A study was therefore conducted to identify a standard methodology for MC determinations in arder to select a rapid method for potential use by drying coops. Freeze-drying chips at ambient temp gave the most accurate results as regards residual moisture and changes in total carbohydrate contents during drying. Calibration equations were obtained for oven drying at 60 and 70"C against freeze drying. In the second stage of the study, three rapid methods were evaluated against oven drying at 7!r.C: infrared lamp, distillation with vegetable oil, and a Brabender MC determination oven. Calibration curves were again obtained. Both infrared lamp and oil distillation offered potential at the drying plant level; however, trials wit~ some drying coops using the latter equipment were not successful as the operation was too complex. In a second study, humidity measurements made using natural and forced-ventilation ovens, infrared lamps and an electric system (Motomco), compared to the standard measurements obtained with a freeze dryer. Unear regression models indicated that the Motomco and Despatch forced-ventilation oven and the natural-convection oven provide reliable humidity measurements; measurements with infrared equipment varied in time due to fluctuations in voltage and amperage. 12.4.2.4 Artificial/mixed drying. The Atlantic Coast of Colombia is characterized by a well-defined drying periad of about 20 wk between Oecember and April. Recently, the drying season has been extended as the farmers have leamed to manage plants more efficiently and have reduced loading rates. Artificial drying offers an alternative during periods of high rainfall when natural drying is impossible. It is also an option tar other regions where a high and evenly distributed raintall pattern makes natural drying difficult. Although artificial drying incurs additional costs for energy use, it can result in labor savings. Product quality is improved. but the animal feed market offers Iittle or no moneta,ry reward for improved quality. Mixed drying systems are an intermediate option: an initial day of natural drying is tollowed by artificial drying to reduce MC to the requlred level. Mixed drying results in lower drying costs than the full artificial drying option and could·be used to increase plant drying capacity, by allowing drying at night after a day of natural drying. An artificial coal-fired drying system was added to a 1500m2 natural drying plant to constitute a mixed drying system, which was pilot tested in 1990-91. Results as regards costs and benefrts were compared with natural drying (Le., a 4OOOm2 drying area, equivalent to capacity increase when adding an artificial drying unit). The results (Table 227 T able 12,2. A eomparison of natural and mlxad dryíng of eassava chips al coop dryíng plan!. en tIIe Allanlle Ct>uI of Colombia. 1991. Cap.acity (lfy) Orying infrastruc1ure FFR (%) Not .amingo {\'lo) Production ceS! (Col $/l) 1 Capital investment (Col $) ' • 'US$I • "".500. Natural 432 Roor 1500m2 56 14 72,500 7.200.000 Dryíng System Mixad Natural 1150 1150 Roor 15OOm::t "" 30m2 FIoor 4000m' artificial dryíng cI1amber 71 72 13 15 73,239 71.071 15.700.000 15, 100.000 12.2) were that the financial rate of return (FRR) and capital costs of fhe mixed and natural drying systems were remarkably similar. The former permits out-of-season drying, with a more regular cash flow and bettar use of fixed costs. 12.4.3 Mllllng and mlxing Added value at the rural level can be increased through the cooperative produ~ion of balanced feed rations from locally available ingredients, including cassava, and their use to raise livestock (chickens, pigs, etc.) by local farmers. A pUot project involving the construction of a small-scale feed mili, consisting of hammer mili, vertical feed mixer and equipment far measuring and incorporatíng molasses into the feed ration, was funded by DRI for operation by a farmer coop. Drled cassava was milled and incorporated ¡nto feed at levels from 35-55%, depending on the poca and availability of other raw materíals. Experimental diets were from 9-18% cheaper than locally available commercial feed concentrate. However, feeding triais with chickens, conducted on three farms in the area, showed that feed consumption and finalliveweight were less for the cassava-based feed than the commercial one. Fead conversion rates were only 5% lower. Qne possible reason is the difference in presentation of the two feeds; the commercial feed is a pellet while the experimental cassava-based feed is a meat, a less efficient means of providing nutrients. Further feed tríals and semicommercial operation of the pilot plant are envisaged (a DRI-ICA project proposal). 12.4.4 Futura resaarch Continued expansion of the dríed cassava industry in Colombia has produced more topies for research. Although animal feed concentrate producars are uniformly concerned 228 about aflatoxin contamination of cassava chips, this has been confirmed in only one case during the last ten years. poorly driad cassava with MCs in excess ot 14% have routinely been analyzed at CIAT for aflatoxin presence and found negative. Studies are required to determine, under controlled conditions, the ambient and chip moisture and temp conditions under which aflatoxin are produced. Observations to date would suggest that cassava chips have an advantage over sorghum and maize in this respecto The opening of many Latín American economies to world market torces could have implications far the viability of the dried cassava industry based solely on the market for animal feed. Market research to assist diversification efforts is an urgant necessity.- This ls under way in Ecuador, where diversification into industrial markets is already well advanced (see Chapo 15). 12.5 Fresh Cassava Conservatlon 12.5.1 Background The rapid postharvest perishability ofthe fresh roots results in a poor-quality, inconvenient foodstuff for urban consumers. Large marketing margins due to the risks of commercializing a highly perishable product frequently make cassava more expensive than competing carbohydrate sources in the urban environment. Research on physiological deterioration of fresh cassava roots, conducted by CIAT and tha NRI during the late 1970s, led to the davelopment 01 a simple storage system besad on root curing in polyethylene bags to prevent the onset of physiological deterioration, and a treatment with a thiabendazole-based chemical to prevent secondary microbial deterioration. Thiabendazole is a permltted agent for postharvest use in many fruits and vegetablas. Residues in cassava tissues haya beén analyzed at less than 20% of permitted levels. The storage system was field tested in several edaphoclimatic regions of Colombia with success, except in high-altitude areas where slower curing resulted in higher stofage losses. Consumer testing of the stored root acceptability and tarmer trials with the storage technology were initially conducted In Bucaramanga (Santander, Colombia), an area of high cassava consumption and production (see CIAT annual reports 1985-86). As this DRI-CIAT project was expanding to semicommercial operations, the deteriorating public order situation in the cassava-production region (Magdalena Medio) torced the closure ot this project; however, DRI was sufficiently convinced of the potential of the storage technology to finance a pilot project on the Atlantic Coast region. Originally, it was thought advisable to keep the fresh cessava projects separate from the dried cassava project; however, the Bucaramanga experience showed that the two products were complementary and could be run in the same area or by the same farmer groups, to their mutual beneflt. 229 12.5.2 Barranqullla pilot project The Atlantie Coast region has the highest per eapita consumption of cassava in Colombia (53.3 kg/yr; natíonal avg 25.5 kg/yr). The dried cassava project had by this date (1987) a1ready produeed over 30 small farmer coops and strong support for cassava in rural development agencies. Barranquilla (population 1,000,000) is the largest eity in the regían. 12.5.2.1 Market and consumer studies.The principal results of a series of wholesaler, retailer and consumer surveys were: • Small neighborhood shops are the most important retail outlet for fresh cassava. Only ene market exists, retalling 12% of total volume. Although supermarkets retall only 9% of total volumes, nearly 50% of the cassava purchased by the upper income stratum is from this outlet. Of the cassava purchased by middle- and low-income consumers, 75% is from small neighborhood shops. • Fresh cassava is more expensive and of lower quality (Le., more deteriorated) in small shops than markets. • Small shopkeepers purchase 20 kg of cassava 5 or 6 times a week. Consumer purchase frequency is higher for low-income strata, but purchase volume is lower. • Low-income consumers purchase foodstuffs daily due to cash limitations (daily wages). Storage of fresh cassava Is not an economie aption for them. , • Wholesale preferencas for cassava were based on production region or certain varieties. Retailer and consumer preferences were related to visual characteristics (peel and parenchyma color). Basad on these results, two marketing strategies were formulated: • Cassava in 4-kg slzed bags for sale through supermarkets to middle- and high-income consumers. • Cassava in 12-kg slzed bags for sale to small shopkeepers, who will benefit from improved quality, fewer losses, better availability of cassava to ctients, and fewer visits to the wholesale market. Low-income consumers who purchase from such neighborhood shops would benefit from quality and less waste, but not from convenience. 12.5.2.2 Consumer panel. 25 middle- and high-income consumers were supplied with a 4-kg beg of recently treated and packed cassava, and asked to evaluate the product after 1 and 2 wk of at-home storage. A fresh root sample of the same varo was also 230 evaluated as a control. losses of stored cassava were only 2.6% after 2 wk, comparad with a postharvest life of 24 h for var. Venezolana under normal conditions. Eating quality of the fresh roots was evaluated as excellent or goOO by 96%. Roots stored for 2 wk were of comparable quality: 100% good or excellent. Only 9% noticed a sweeter taste of stored roots, which were generally rated as easier to cook and peel than fresh roots. Similar results wera obtained with 21 shopkeepars: 95% toought this systam would result in increased sales. A final trial was conducted· in a supermarket. where 2-kg bags of treated and packed cassava were placed on sale. AII 70 bags were sold in one sayo Follow-up interviews showed that 95% of consumers rated eating quality as good or excellent. Mean storage time at home was 4.6 days for 2-kg bags and 6.4 days for some 4-kg bags that were also sold. Optimum bag size was 4 kg, although 2 kg was useful for product introduction. 12.5.2.3 Farmer trials. A region producing high-quality cassava as continuously as possible throughout the year was sought. COOPROMERCAR, a DRI-supported coop of farmers in Repelón (Atlántico), was commercializing tomatoes at a loss, and members also produced cassava for individual sale to traders. The coop is situated in a irrigation district, thus off-season cassava productión was possible. Venezolana, the preferred varo in Barranquilla. was widely grown, and eating quality was excellent. Farmers organized a production team for the fresh cassava storage operations, producirig all the bagged cassava used far the consumer testing in Barranquilla. Costs of the packing and treatment process were monitored. Following the end of the consumer- tasting periodo prices for the supply of bagged; cassava were negotiatad between a supermarket chain and the coop. The results of fue first 2 mo commercialization of over 50 t of cassava are presentad in Tabla 12.3. Although the bagged cassava was more expensive than the normal cassava (Col$60jkg vS. Col$52jkg), this was acceptable to consumers as a higher proportion of the purchased product was consumed and quality was better. 12.5.2.4 Project expansiono In collaboration with DRI, a project was developed with the objective of expanding the voluma of cassava in bags marketed in Barranquilla to 4000 l/yr (approx. 10% of the market). To meet this obJective, activities were ¡nmated on several froms: • Identification of cassava-production regions capable of supplying the Barranquilla market with high-quality fresh cassava. As no one production region produces high- quality cassava 12 mo ayear, it is necessary to switch production region at certain times. The best harvest months for each region were identified, and farmer organizations already existing in each regían were contacted. 231 Table 12.3. Resulta of Inltial 2 mo of marketing 51 1 01 """""va in bago by COOPROMERCAR. Repelón, Alfántioo Colombia (Aug./Sopt. 1987). V81Iable oos.t. Raw materia' !.abor PoIyeIhyiene bag Thlabendazole Oto., Tren.port lo Barranqullle Total Income from sale of ca_va l.Ma casll fIow, worldng capital 00$1, depreeiation Nel prom mar"ln Total profit on 51 t ""Id 1 CoI$ 252 • US$I, Sepl. 1987. Cost/kg Cassava Col.$' 2.25 26.43 1.84 2.39 0.98 1.69 2.83 38.45 44.44 6.31 5.66 288.875 • Training of cassava coops in storage technology, To date a total of 15 coops and 2 second-order organizations have be/m trained. Of these, 11 have used the storage technology commercially. • Organization of a distribution enterprise in Barranquilla. A second-order federation of the cassava coops in Atlántico, Magdalena and Bollvar (the states closest to Barranquil1a) was formed in 1989 (FAGROCOL). Although the Federation's main actMty was to commercialize dried cassava, the coops were also interested in using the Federation to set up a distribution enterprise to coordinate sales of fresh cassava in Barranquilla, essentially to act as an intermediary. DRI obtained a space in the new wholesale market in Barranquilla for FAGROCOL, which opened in 1990. DRI and CORFAS provided FAGROCOL with a small amount of working capital to inmate operations. Orders for over 100 tjmo were rapidly obtained from supermarkets and restaurants. Problems were soon encountered, however: • Quality control. As the no. of coops treating and packing cassava increased, the variability in quality became marked. This was due to both variations in inherent eating quality of the cassava and in the execution of the treatment and packing operations. 232 • Lack of public awareness of the advantages of stored cassava • Shortage bf working capital The Federation decided to take a more active role in the treatment and packing operations, and experiments were conducted to see if delayed treatment wes a feasible option. This would allow operations to be carried out at one central lacation under Federation control, rather than at the level of the individual coops. 12.5.2.5 Modifications to the storage technology based on the pilot expedence. Four basic problems were identified: • Lack of quality control by individual coops. The solution was to: " pack cassava into large pOlypropylene sacks initially (no treatment} " transport to central lacation .. treat with thiabendazole " repack into polyethylene bags of required size There were no significant differences between treatment at harvest or after 24 h; delaying treatment 48 h did affect storage success, however. • Extreme susceptibílity of Venezolana to physiological deterioration at certain times of the year (deterioration < 12h after harvest). The solution was to prune the aerial part of plant 5-8 days before harvest, which reduced the severity of deterioration without affecting eating quality. ' • Secondary deterioration due to bacteria not controlled by thiabendazole .. The solution was to include sodium bisulfrte in the root treatment, which gave excellent control of bacterial problems. • Large nO.of roots with excessive mechanical damage, unsuitable for storage. The solution was to cut away damaged area, cover exposed surface with calcium carbonate powder (desiccant). The exposed surface dries cleanly; after removal of superficial tissues, the remainder of the parenchyma is of goOO eating quality 12.5.2.6 Recent Developments. This series of experiments provided a range of technology options suitable for all conditíons. FAGRQCOL started to use the new methOOology in early 1990, usíng the space at the wholesale market for treatment operations. On arrival, cassava was selected for treatment and storage, based on actual orders; the remainder was sold in the normal fresh market. FAGROCOL soon gaíned a reputation for selling high-quality cassava, becoming the price setter for a11 intermediaries. At this time, most of the 35 intermediaries in Barranquilla switched from sisal sacks to polypropylene sacks for the transport of roots from the fleld. This was a direct result of 233 the prOject's experiences with polypropylene saeks alone, permitting a longer storage life of varo Venezolana. During 1990 FAGROCOL supplied supermarkets directly with bagged eassava. and an urban distributor was hired to reach the small shops in lower income barrios. Sorne sales were also made directly from the warehouse in the wholesale market. Total volumes traded reachad 10-12 tjwk. Unfortunately, trom Oetober 1990 onward, FAGROCOL was beset by a series of financial problems resulting from losses taken in eommercializing and exporting other products (yams, tomatoes, etc.). Since early 1991 they stopped eommercializing fresh cassava for lac!< of funds (DAI did not approve a new cassava- relatad project for an organization in debt). The organization is currently being liquidated. Fortunately, the supply of fresh cassava in bags has éontinued in Barranquilla throughout 1991, as private individuals have set up operations. One individual in Barranquilla is marketing approximately 15 tfwk. In addition COOPAOMEACAR is onca again commercializing bagged cassava directly to both Barranquilla and Cartagena. Several wholesalers are now supplying untreated cassava in polyethylene bags to small shopkeepers '(12 kg cassava/bag), inereasing storage lite 2-3 days. 12.5.2.7 Future actívities. DRI-CIAT activitíes will termínate early in 1992, given the absence of an organízation to execute projects of a cooperative nature, and the reeent success of the private sector in taking this technology to a commercial level. The 9ituation will however be monitored to ensure that any further problema are resolved. The affects of the project on>the fresh cassava market·in Barranquilla will also be evaruated, given that some of the technology eomponents have recently been widely adopted. The lessons rearned in this project will be applied to similar projects now under way ir! Santander (DAI-Colombia) and in Paraguay. 12.6 Cassava Flour tor Human Consumption 12.6.1 Background High-quality cassava flour for human consumption has the potential to be used as a substitute for wheat flour in a no. of food products. Several institutes have researched the formulation of bakery and other foods using cassava flour, and the substitution rates are known in many cases. Many tropical countries import significant volumes of wheat, which could.be replaced by cassava if price competitive. This requires a low raw material price and efficient low-cost processing. Industrial·scsle processes for obtslning cassava flOtl' have been developed in Brazil and elsewhere, but are little used in practica. From 1984-86 IDRC funded a proJect at CIAT in which an efficient, sman-scale process for \he production of high-quality flour was developed, in collaboration with UNIVALLE. Bread formulated with cassava flour was tested under lab conditions at the Institute of Tec:hnological Aesearch (liT), Bogotá and with bakeries and consumers in Colombia. 234 Finally, an evaluation of the wheat flour system in Colombia was made to assess the feasibility of a cassava flour industry and to recommend market channels for the producto The cassava flour process developed at CIAT consists of root reception and selection, washing and removal of the outer bark layer, chipping, artificial fixed-bed drying using a coal or coke fired burner wíth heat exchanger, followed by premilling to reduce chip size. Premilled chips can be milled successfully in a normal wheat flour mili (90% conversion rate to high-qualíty flour). The conversion rate from fresh roots to pre-milled chips is approx. 3:1, comparad with 2.5:1 for cassava chips for animal feed and 4.5:1 for cassava starch. The efficiency and low cost of this small-scale procass result in a highly price- . competitive producto 12.6.2 Pllot project In 1989 IORC funded a second (pilot) phase of this project with CIAT and ORI as the joint executing agencies. This proJect, in which UNlVALLE is also collaborating in a research- support capacity, is now in its final year. The cassava-production component of this intégrated project, invoMng the tarmer testing of production technology packages, is reported elsewhere (Chap. 9). A discussion of the three main objectives of this second phase follows. 12.6.2.1 Objective 1. To implement, adapt and evaluate, on a pilot scale in a rural context, the technology developed for production of dried cassava chips and flour. After evaluating six sites on the Atlantic Coast according to relevant eríteria, Chinú (Córdoba) was chosen. The tarmar coop (COOPROALGA) in this cassava-growing area also operates a natural drying plant. ' The cassava flour pilot plant was designed by an architectural firm in collaboration with CIAT. It was subsequently scaled down (to an annual capacity of 200 t) to reduce construction costs to US$ 50,000. A UNIVALLE-Ied team worked on improving the root- washing machine, coupling of washing/chipping machines, and chipping machine. Design criterla emphasized reliability, efficiency and securíty. The original plant personnel, including a plant chief, production chiaf, threa workers and a watchman, were selacted by the farmer coop's administration. Personnel changes have occurred; new personnel receive in-service training conducted by CIAT. The current plant chief is also the community leader. Because of intervillage rivalries, the plant was deniad a stable water supply; therefora, COOPROALGA had to build its own well and pipeline at a cost 01 US$9,000. As the root- washing step was initially impossible, the plant produced chips for animal leed in 1990. Once water supply limítations were overcome, the plant ¡nitiated productian 01 cassava chips tar human consumptian in January 1991, producing 30 t in the first semester. This was sent to a MedeUfn wheat mili tar milling. The target extraction rate, already achieved, 235 is 90% ot first-grade cassava flour plus the two by-products, second-grade cassava (6%) and peel (3%). About 1% is lost in the milling and screening processes. Chemical analyses of the different flours have been obtained. During the first months of plant operation, chip quality of all lots was closely monitored. Table 12.4 presents representative results. Samples have been slightly aboye the 50 ppm limit tor total HCN, but this has not been seen as important by industrial clients. Stareh, fiber and ash contents are all satisfactory. Microbial counts have not consistently met standards, however--the major coneern of potential industrial clients. The conversion rate of fresh eassava to premilled chips has averaged 2.7. With an extraction rate of 90% to obtain high-grade flour, the fresh root-to-flour conversion rate is 3.3. PUot plant operation as of Aug. 1991 points to the following three main challenges: (a) to obtain a continuous supply of freshly harvested. high-grade roots. (b) to lower drying costs through improved burner and heat exchanger designo and (e) to produce chips that are microbiologically acceptable to the tood industry. To meet the first ehallenge, a network of eassava wholesalers 15 being established in several regions to supply freshly harvested, high-grade roots during the penod from June to Nov. Neighboring farmers can supply roots from Dec. to May. This network will also TabIe 12.4. Ch.mica! composítion and mlCifoblal quallty 01 2 samples 01 eassava produoed by 111. pilO! planl. comparad will1 Colombian standard •. ICONTEC Standard Sample 1 Sample 2 Chemleal Com!1Qsítion Me ('J.) 12 (max) 6 5 Starch ('J.) 52 (min) 89 87 Ash ('J.) 2 (max) 1.4 1.5 Aber ('J.) 2.6 (max) 1.4 2.0 Total HCN (ppm) 50 (max) 36 61 Aftatoxins O (max) O O Microbial contents Aerobic mesophilous bacteria 200.000 (max) 6,300 827,.200 CoIHorm bacteria 100 (max) O 1,200 E. coll O (max) O O Salmonella O (max) O O Fungl and yoam 100 (max) O 100 Note: !CONTEC. CoIomblan Standardslnstltute; standard sel for cassava flour for human consumption. 236 help meet ehallenge (e). Challenge (b) can be confronted by building the burner wlth briek instead of iron and by mOOifying ehamber dimensions. The burner and heat exehanger should also be protected from cool night winds. Improvement ot microbiological quality Is mainly SQught by washing the eassava roots with sodium hypoehlorite-treated water for 5 min. In addition, the perlOO between cassava harvest and processing is minimized to avoid deterioration. 12.6.2.2 Objectlve 2. To identify and research eassava flour markets and promc,e the use of ~ssava flour in these markets. A market study was conducted at 3 geographic levels: locally (Chinú area of influence), regionally (Atlantic Coast) and nationally (Medellrn and Cali). More than 200 small, medium and large food processing companies representi~g multiple tood eategories were surveyed and given cassava flour samples to anable flour substitution trials. A second survey obtained feedback on substitution trial results and intent to purehase. This study identified toOO segments where cassava flour utilization has potential. To astimate volume potential, this information was complementad . with seeondary information on wheat imports, and experts were Interviewed in firms participating in key tood categoríes. The maln eonelusíons were as follows: • A eonservative estímate of markef demand in the medium term (5-10 yr) for cassava fIour as a substitute of other flours or starchas In tha food industry Is 22,000 tfyr. This estímate assumes fow substitution and adoption rates in the bread-making segment, by tar the largast one in the toad industry. The maln foOO categorías where Ose of cassava flour is most feasible are: processed meats, sweet cookies, pasta and soup noodlas, cakes, spice bases, bread-making, meat pies, porridge mixes, soup mixes, soft sweets, lee-eream conas, breading and sauces. In sorne of thesa categoríes, cassava flour exhibits functional advantages over competing flours given Its high water- absorption capacity, binding potential, erispness-enhaneement characteristícs, etc . • The main raw materíal to be substitutad would be wheat flour, as well as corn flour and sweet.cassava stareh. Consequently, the reeommended market-penetration price for cassava flour could be equivalent to 80-90% ot the priee for wheat lIour in Medellrn, eurrently at Col$230/kg (US$O.38). This price Is acceptable to maJor clients. Of the two largest cities nearest Chinú (MedeUrn and Barranquilla), the former exhibits a greater and more coneentrated demand and has been selected as the site of inltíal market penetration. Initial promotion of cassava flour was eonducted in the eontext of the national market study and consisted of personal contact with the interviewer, a promotional pamphlet and a sample. Subsequent promotional activities have eoncentrated on the target market, Madellrn. Twelve large- and medium-sized firms representing key foOO categories, most of whieh had expressed a positive buying intention after the substitution trials, have been visited twice, and so- to 200-kg samples have been provided ter further tria/s. 237 A eassava flour distribution system in Medellrn is being established with the participation of TECNAS, a processed-meat consultant, and Harinera Antioqueña, the wheat míll. The former is focusing on the proeessed-meat markets, whila tha lattar will concantrata on other markets such as cookies, cakas and braad-making. By-products will be sold to tha animal feed industry. Negotiations with these two distributors involved declsions on buying and salling prices, scope of responsibilities, and costs of services provided. TECNAS has proposad to purchase the first-grade eassava flour and resell it, while the mili would eharge for milling and distribution, without purchasing the producto The financial modal of the eassava flour pilot plant has been very useful in negotiations. A brand name for eassava flour, YUKARIBE, is being registered nationally in the flour category. This name will appear on the SO-kg polypropylene bag, which is larger and denser than the ones used for packaging wheat f1our. In general, it can be concluded that food-processing companies tend too be quite conservative when adoptlng new raw materials. The microbiological quality of the flour is important to potential clients in MedelUn; this has been the major limítation to proceeding with commercial sales. In the area of product development, formulations using cassava flour were developed for a typical sweet ("manjarblanco') and porridge f'coladas"). Manjarblanco can be prepared, maintaining good organoleptical quality, substítuting 25 to 100% of the standard rice flour with cassava f1our. In coladas, other flours were substituted suceessfully with cassava fteur at 5 to 20% levels. Preparations prasantad 24-h stability; consistancy was softar and texture slightly granular. 12.6.2.3 Objective 3. To estimate feasibility of establishing acassava f10ur agroindustry Sfld make recommendations, if warranted, on implementation on·a national basis. A feasibility study should eonsider technieal, financial, marketing, social and organizational aspects. Given that most of these aspects have been diseussed alraady, financial aspects will be diseussed here. Financial models of the pilot plant were daveloped to analyze the impact of multiple variables on financial profitability (as measured by the FRR) and annual cash f1ow. The basie model ineludes investment costs, variabla and fixed costs, sales price and a eash flow estimate for 8 yr. It was demonstrated that (a) selling cassava flour is more profitable than seUing ca,;sava chips; (b) the most profitable míllíng a1ternative is in-plant roller milling (FRR >50%), followed by subcontracting milling externally in a wheat mili and by in-plant hammer milling; and (e) finaneial profitability is highly sensitive to plant capacity and capaeity utilization, root price and fresh root-to-dry chip eonversion factor. The financial model was very helpful in negotiating cassava flour prices with potential distributors and elients. The model indicated, for example, that it is more profitable to seU to TECNAS than to sell through tha wheat mili, under current conditions. 238 It must be notad that tha govarnmant's currant fraa-market policy in Colombia may affect the project negatively if the real price of wheat decreases greatly; a minor priee reduction can be absorbed easily by reducíng profitability margins. In the case of an unanticipated, radical prlce reduction of wheat, alternative industrial markets can be pursued. 12.6.3 Support research 12.6.3.1 Small-scale mili development. Although the pilot plant Is eurrently producing only premilled chips, the provision of a small-scala milling capacity, which would permit the in-plant proouction of flour, would increase the.FRR by 10%. Trials with existing small hammer milis resulted in conversion rates of chips to high-quality fIour of only 65%. Research is currently under way to evaluate several alternative systems for small-scale milling, and to design and test a prototype mili based on the best option. Eleven different combinations 01 milis, screens and different sizad meshes were testad. For each system, the content of fiber, ash, HCN and extraction rate were taken. The best option was found to be a system consisting of (a) roller mili, with the eylínders set to different speeds (420 and 230 rpm); and (b) two screens in series, with 2.38mm- and 2.50mm-diameter mesh. This has given extraction rates 01 90% in ínitial tríals, with fiber and ash contents 01 1.5 and 1.3%, resp. A prototype mili incorporating these features will be evaluatad in the pilot plant during 1992. • 12.6.3.2 Flour storage studíes. To study stability of cassava chips dllring storaga, batches wer~ stored in government warehouses in three cities exhibiting different elimatie conditions. Samples were analyzed monthly to determine microbiological and physieochemical quality. Results indicata little variation with time regardíng the latter aspecto Yeast and mold populations wera within norms; aflatoxins and pathogenic bacteria were not found. In addition, initial aerobic bacteria counts were high but decreased to acceptable levels after 20 days. 12.6.4 Future plans A third phase of this joint CIAT ¡ORI projeet has been submitted to IORC for funding. If approved, activities over the next 3 yr will tocus on: • Expansion 01 pilot plant capaeity to eommercial level • Evaluation of in-plant milling system and local marketing of cassava flour • ldentifieation 01 sites for repUeate plants in Colombia • Continued market promotion of cassava flour • Training of national institutions in the technical and other aspects 01 cassava flour processing 239 • Redesign of the plant and equipment to reduce costs • Dissemination of project results outside Colombia The World Food Program of the United Nations has offered financing for new cassava flour plants in Colombia, based on the project progress to date. A similar project is currently under way in Indonesia, where the national agricultural research program (CRlFC) has been developing small-scale cassava flour processing equipment, which is now being pilot tested. At the same time, private sector food companies have been initiating their own cassava flour production far manufacturing a range of cookies and cakes for the natianal market and exporto The Cassava Program is collaborating with CRIFC in evaluating flour quality (analytical methods) and in exchanging information on processing equipment. 12.7 Cassava Starch 12.7.1 Background Native and modified starches are important raw materials for many industrial products; e.g., in food processing, paper manufacturlng, textile, adhesive and oil drilling industries. As a glucose polymer, starch is also a raw material for producing rnany derived products in sugar chemistry (glucose, fructose, maltodextrins, mannitol, etc.), each of them with specific properties and specific uses in food, chemical or pharmaceutical industries. Afthough cassava is one of the best sources of starch (ca. 85% of root parenchyma DM), itrepresents only 4% of the starch used by industry. Annual production ot cassava starch far industrial !Jses is about 800,000 t, malnly in Brazil (tar the national market) and Thailand (tor export to Japan and the EC). Many small-scale cassava starch industries exist in tropical countries where the product has specific uses in traditional foad industries corresponding to a specific market niche; e.g., "krupuk" in Indonesia, sago in India, "pandebono" in Colombia, "biscoicho' in Brazil and 'chipa" in Paraguay. In Colombia and Brazil, 'sour" starch is produced, a naturally fermented starch whose specific functional properties are irreplaceable in the manufacture of traditional cheese breads. This smalf-scale industry has a high socioeconomic importance in specific regions of these countries; nevertheless, the sector presents many problems in terms of production, processing and commercialization, which limit its development. In order to research and develop new technologies far this industry, a collaborative project between CIAT-CEEMAT /CIRAD on ·Production and utilization of cassava starch" was initiated in 1989. 240 A technical diagnosis of the traditional process in Colombian cassava extraction plants CraRander[as"), which process from 1 to 5 t of fresh roots daily, was carried out to determine the main problems of this industry. These were identifiad as: • Low procass efficiency with starch losses of 25% • Large variation in product quality, with no objective parameters for quality evaluation of sour starch being available • Lack of knowledga of the effect of the raw material on starch quality and extraction yields The following priority areas of research were established in agreement with producers and users of starch: • Standardization of analytical assays • Characterization of sour starch .. Equipment tachnology improvement .. Study of tha influences of raw material and the process on product quality .. Study 01 the mechanisms involved in natural fermentation (fór sour starch production) These topics have been developed in collaboration with Colombian institutions: lIT, The Consulting and Development Service for Coops and Small Enterprises (SEDECOM), UNIVALLE (Food Saetíon, Chemistry Department, Dept. of Mechanical Engineering) and the U. Aut6noma del Occidente (Depts. of Mechanical and Industrial Engineering) in Cali. Some national agencies are supporting this project in Colombia, Ecuador and Paraguay by províding technical assistance for cassava production, commercialization and farmers and starch produéar organization in order to transfer the first results of research (sae respective country reports). 12.7.2 Improvement of the sour starch extraction process The traditional sour starch extraction process includes the following steps: .. Washing and peeling the roots to remove adhering soil and peel .. Grating the roots to destroy cell walls and ralease starch granules ... Extracting the starch under running water and screening the pulp to separate the starch granules from other components (mainly fibrous residue callad "afrecho") .. Saparating the solid starch from the starch milk by decanting into tanks and running off the supernatant after a one-day decanting period (tha yellowish upper layer of sedimented starch ("mancha"), which contains proteins and some impurities, is removed before a new sedimentation or the transfer of starch to fermentation tanks at the end of the wk) lOo Natural fermenting of sedimentad starch in tanks for 20 to 30 days • Solar drying of wet starch to 12-14% for long-term storage 241 • In a first stage, equipment technology design was based on the same principies as the existing equipment, keeping costs to a minimum. The improved equipment is now being evaluated with the producers in some pilot rallanderras. In a second stage, the possibility 01 scale reduction 01 equipment using advanced technology is been studied. This will involve using a grater with water, vacuum filter and separation to offer a new small-scale technology (5-10 t 1resh rootsjday) with the possibility 01 continuous operation. 12.7.2.1 Washinq /peeling. The presence of external peel affects end product color and increases the quantity 01 mancha. The traditional washer is a rotating drum (diameter 0.80m, length 1 m). The 101l0wing modifications have been made: • addition 01 2 or 4 abrasive rollers • distribution 01 washing water from the central axle • direct power transmission by a reduction motor The peeling rapidly reaches 90%, which increases machine capacity to 1 tjh. 12.7.2.2 Extraction. The low eifficiency of the process is due to poor release 01 starch granules during grating and limited extraction of starch (an avg 18% 01 the starch is left in the a1recho). • Inefficient sieving, may result in a too high fiber content in the final producto • Extractor capacity (200-250 kgjh) is less than other steps in the process, which creates bottlenecks and decreases productivity. The traditional grater is made 01 a wooden drum covered with a perforated iron sheet. The extractor is a rotating drum equipped with internal paddles and a cloth fixed around the drum is used as a sieve. The 101l0wing improvements were made: • Grater: cutting blades fixed on a plastic drum with a high rotation velocity. To reduce costs a wooden drum has been built and will be tested this year. • Extractor: a traditional cylindrical water extractor was equipped with 4 mechanized mixing screws to improve contact between water and the cassava mash, and an external metallic sieve (60 mesh). • Sifter: a vibratory sifter with 2 sieves (SO & 120 mesh) was tested. Initial results have shown ca. 20% improvement in the starch extraction rate, reflected in a decrease in the starch content 01 the afrecho. Additional modifications will further improve this. 242 12.7.2.3 Starcb separation. The natural decantation of starch milk in settling tanks takes 24 h. This is long enough to permit the development of microorganisms and induce chemical changes in the starch. To avoid removal of starch every day, successive sedimentations are carried out during one week. The natural fermentation which thus occurs modifies drastically the functional properties of native starch; in these conditions the final product does not meet industrial specifications for sweet (nonfermented) starch. During starch separation, 20% of the starch is lost with the removal of the supematant. " • The improved system consists of sedimentation in settlíng channels (100-120 m long, 40- 50 cm wide, with a 0.5% slope, as used in India and BraziQ. This system has signíficantly reduced starch losses, increasing recovery trom 57% to 69%. This system also permits classifying starch by purity and granule size. Settling tanks and channels have similar capital costs (US$1000). Starch prOOucers have seen the advantages of this improved system, and it has been already adopted by many producers in Colombia. Demonstration plants with this s.ystem are now operating in both Ecuador and Paraguay. 12.7.3 Improvement of starch quality The main quality parameter that sour· starch users (small-scale bakeries or larga companies) are searching far is "expansion power" during baking; however, no quality test with a scientífic basis exists for this parameter. Empirical tests are used; e.g., acid taste or the color of a f1ame in presence of starch dust. The first priority was to astablish a simple methOO for evaluating sour starch quality, to correlate quality with sorne sour starch properties, and to explain the mechanisms involved in producing expansion power. The other, more objective criterion is the microbiological quality of sour starch: Coliform bacteria are sometimes found, mainly because of contamination of the extraction water or of the starch during drying by animals. This can be controlled by the adoption of basic hygiene measures. 12.7.3.1 Baking test. The "expansion power" of sour starch is defined as"the ability of a fermented starch to increase the vol. of a dough contalning that starch and submitted to a process of baking." The simplest parameter to measure this is the specific volume of the breads after baking; nevertheless, this expansion power depends not only on starch quality but also on the cheesa usad (presents many fluctuations in quality), !he product formulation, !he dough texture and consistancy, the shape of the breads and the baking conditions. For setting up a reproducible and sensitive baking test, severa! assays using good- and poor-quality starch samples were carried out under different baking conditions and formulations of chaese, fat, salt and yeast. The following protocol gave !he best differentiation between good an<1 bad starch quality, with goOO reprOOucibility: 1> Manual preparation of a dough made of sour starch (1 part), commercia! white ('campesino") cheese (1 part) and cold water (0.5 parts) at ambient temp .. Baking of 6-8 small round breads (15 g) at 4500 F for 7 min in a domestic electñc oven 243 • Determination of the avg specific vol. of the cheese breads by the seed displacement method This test can be used at the level of the lab to determine the influences ot raw material or process steps, as well as in the extraction plants as an element of process-quality control and for product price negotiations. 12.7.3.2 Evaluation of different sour starches. To giva a quality scale relatad to specific vol., 35 samples were collacted and ranked in 3 classes by clustering tha values of the specific vol. determined by tha baking test (Table 12.5). Between specific vol. and starch characterization parameters (biochemical, physical and functional properties), the following significant correlations (p < 0.05) were found: total organic acids, lactic (+) and acetic (-) acids, color (brightness), swelling power at SO"C, water absorption at 40 and 500 BU, viscoamylogram characteristics. The physicochemical parameters and starch functional properties vary greatly from one plant to ananother and between fermentatíons at the same plant. Soma factors such as ambient conditions, the variety, water quality and traditional knowledge may be involved in this variability of the natural fermentation. The fermentation step is the key to obtaining good-quality sour starch. 12.7.3.3 Baking power property of sour starch. The studies carried out on the natural fermentation of cassava starch and the modilications of its functional and physicochemical properties that occur during this step have shown: • A dominant lactic microflora with amylolitic activity • Production of organic acids, mainly lactic acid and CO. • Attack of starch granules by amylolitic enzymes (at end of farméntation, some granules are pitted and porous • Drastic modificatíon of the functional propertias of the starch (decrease in max. viscosity during haating, no gelling tendancy during cooling) Table 12.5. CI __ 01 sour starch clusters' . 1 2 3 Oualily Starch Good Foir Bad 1 Values with the same letter represeJ1t a significant difference, p <: 0.05. 244 Mean Specific Vol. (SO) 4.89 ml/g (0.05) a, b 4.21 ml/g (0.170 a. e 3.38 ml/g (0.15) b, e It is possible that during baking, organic acids and gases producad during the fermentation and absorbed by 1he starch granules are volatilizad. Partial gelatinization of the starch gives a viscoelastic strueture to the dough. which is able to confine the discharged gases. The baking power may therefóre consist in a competition between the gaseous expansion and the viscoelastic properties of the amylaceous network; however. other hypotheses can be presented. Molecules may be formed that associate with the starch to produce the adequate viscoelastic structure; e.g.: • Formation of an amylose~omplex during baking • Produetion of polysaccharides such as pentosan and dextran (whose beneficial effeets are well known in bread-making) by bacteria (Leuconostac mesenteroides). which have been isolated during fermentation; a significant increase of sugars during fermentation a1so occurs. • Produetion of phenolic compounds esterified with these polysaccharides; a complex could occur during solar dryiOO by an oxidative reaetion catalyzad by UV rays. This hypothesis would explain the faet that solar drying is indispensable for obtaining good- quality sour starch according to all the sour starch producers. Based on these different hypotheses. research is now facused on the following points: " Influence of solar drying on sour starch expansion power " Presence of arnylose complexes. polysaccharide. phenolic compounds and changes in their concentration during fermentation " Isolation of the microorganisms involved in modifying the physicochemical and funetional properties of starch (arnylolític laetic bacteria) and the production of polysaccharides (L. mesenteroides) " Identification of the isolated microorganísms, and metabolic and enzyme system studías 12.7.4 Varietal suitabillty According to the starch producers in Colombia. there are large differences between the 2 main cassava var. as regards extraetion yíelds and sour starch quality. M Col 8 (Blanquita. long cycle) yields better but the quality is worse; the contrary is true for M Col 1522 (Algodona. short cycla). 12.7.4.1 Varietal jnfluences on extraction yíelds and starch quality. Five varo were processed in a rallanderla (Table 12.6). The harvest period plays an important role in extraction yields. mainly for var. susceptible to root rots (CMC-40. M Col 8). M Col 8 and CM 523·7 gave the best-quality starch. Other tríals have been undertaken thís year to evaluate the suítability of varo tor sour starch production. 245 Table 12.6. Extrac\ion yield. and oour starch qualily lor 5 cassava varo Varlety AQe (mo) Ex!ractlon YI.ld (%) Specific Vol. (ml/g) Oualily CMC40 10 24.4 3.9 Bad CMC40 12 16.8 4.0 Bad MCol8 10 21.8 4.4 Fair MCol8 12 17.8 4.2 Fair M Col 1684 12 17.9 3.8 Bad CM 523-7 12 18.3 5.2 Good M Col 1522 16 20.4 5.6 Good 12.7.4.2. Sour starch.QualitY vadation among varieties. To understand the vadations in quality. the fundional properties of sour starches from different varo were determinad, Viscoamylograms (Fig. 12.1) showed differences among tha starch paste of different qualities¡var. 300 iiSCoSitY (Brabender Uníts) 250 I ::: r 100 1- ... _~~ ... _________ •• ~. ______ ww O~~~~~~--~-~~~-~--~ o 10 20 30 40 50 60 70 80 90 100 110 120 eMe 40 CM 523-7 Time (min) M Col 1684- Mcol 1522 M Col 8 Figure 12.1. Viscoamylograms of sour starches extrac1ed from dlfferent cassava varo 246 • The va/ues of max. viscosity and viscosity after 20 mln at 90°C decreased as sour starch quality decllned. • The gelatinization index is lower and near zero for goad-quality starches. This confirms the importance of functional properties in the creation of the víscoelastic network duríng baking. As mentíoned before on the characterization of native starches (4.4.1). varíetles must play an Important role in explalnlng the differences in the behavior of fermented starches. A study of the structure and the degree of polymerlzatlon of starches during fermentatlon will provide more information on the changes In starch propertles and on the mechanlsms of attack of starch granules by microorganlsms. 12.8 Futura Actlvlties Of the four priority products researched up to 1991, only two will remaln active beyond 1992: cassava flour and starch. Activities in dried cassava and fresh cassava conservatlon will continue only as required in support of Integrated projects involving these products. No new research will be undertaken in these areas. Research on cassava flour will continue through 1994, when the Colombian project wlll be' reachlng a conclusion. Research on cassava starch will continue as an important element of ongoing CIAT -CEEMAT collaboration. No new research will be initiated on other cassava productsat eIAT, glven the increased emphasis being placed on quality issues. Nevertheless, a great deal of research remalns to be done. The Cassava Program will actively encourage and support process and product development research by other Institutions in the developing and developed world. A research network Is in formatlon, and a specia/ project to obtain flnancing for other insUMions to carry out this research is being written with a no. of collaborators. Emphasis will be on developing derived products from flour and starch, for tood and other industries. Brazilian institutions will be encouraged to undertake product development research on tarinha, especially on the potential for ímprovíng markets through a better quality producto A vacuum remains with regard to realizing the potential of cassava leaves as animal or human foad, which CIAT is unable to fill at present. 247 REGIONAL COlLABORATION 13. REGIONAL COLLABORATION IN LATrN AMERICA 249 14. COLOMBIA 259 15. ECUADOR INTEGRATED CASSAVA PROJECT 273 16. BRAZIL 297 17. PARAGUAY 313 18. SEED SUPPLY SYSTEMS 325 19. REGIONAL COLLABORATION IN ASIA 335 20. GERMPLASM IMPROVEMENT 343 21. SOIL CONSERVATlON AND FERTILlTI MAINTENANCE RESEARCH IN ASIA 355 22. COLLABORATION WITH liTA IN SUPPORT OF AFRICAN NATIONAL PROGRAMS 375 23. EXPANSION OF THE GERMPLASM BASE FOR AFRICA 379 24. COLLABORATIVE STUDY OF CASSAVA IN AFRICA 399 25. ADOPTION ANO IMPACT STUDIES 407 13. REGIONAL COLLABORATJON IN LATIN AMERICA A1though only 22% of cassava is produced in Latin America, the continant of origín of the crop, it is still a major tood erop with a per capita production of 74 kg in 1986-88, 75% of which is produced in Brazíl. Cassava production declined throughout the 1970s and early 1980s as a result of several factors, notably the growth of wheat flour subsidies, which caused a decline in consumption of farinha (a toastad cassava flour) in Brazil; and the declining consumption of fresh cassava in rapidly urbanizing populations. There are now signs that this trend has been halted. Table 13.1 shows the production data for cassava in Latin America from 1987 to date. If the preliminary data for 1990 are confirmed, cassava production in Brazil, Colombia and Latín Ameriea will be at its highest since 1986, 1981 and 1972, resp. The traditional utilization of cassava in Latin America has been dominated by the production of farinha in Brazil, which accounts for 50% of the cassava produced in Brazil and 40% of the total for the Americas. Fresh cassava for human consumption accounts for approx. 25% of production; animal feed, also from the fresh root, accounting for another 20%. Starch and other industrial uses total only 5% of production, while about 10% is wasted (Fig. 13.1). The srnall amount of cassava dried for animal feed does not yet aceount for a significant percent of the total utilization of the crop in Latin America although it is now locally important in several regions. 13.1 Modes of Collaboratlon The Program operates two dístinet modes of collaboration with Latín American countries. Tha fírst is the more traditional approach of interacting with national research programs, with the aim of strengthening these important but often underfunded institutions. Most countries now have cassava research programs, partly as a result of CIAT efforts over many years; however, these programs remain small and of relatively low status within their organizations. Nevertheless, there are many dedieated professionals with years of experience in cassava researeh and extension. Collaboration with such persannel consists in (a) training courses and workshops on speeifícthernes at CIAT and in-country; (b) joint researeh projects, particularly in areas where for ecological reasans it is not possible to eonduct the research in Colombia; and (e) applied and adaptive research aimed at the cassava-production regions in each country. In the area of cassava utilization and product development, national programs do not have and are not developing a researeh capacity. Research collaboration in this area has therefore besn mainly with universities and tood technology institutes. Several national research and extension programs are now undertaking adaptive research and extension activitíes in the area of utilizatian. 249 Table 13.1. Cassava production In latIn Amorlea ('0000 tj. 1970 1980 1987 1988 1988 Latin America 34,735 29,940 30,590 29.109 30.843 Brazil 29,464 23,466 23,464 21.612 23,247 Colombia 1,200 2,150 1,260 1,2l!O 1,509 Paraguay 1,580 2,031 3,466 3,890 4,000 Wo~d 97,597 124,707 136,801 141,109 147,800 ~ FAO produotlon dala (1990, prellmlnatyj, excapt Colombia (Mlniatry 01 Agrlculturej. Fresh consumotion 25% Farinha de mandioca "0% Staroh and others 5% Waste 10% Figure 13.1. Cassava utillzatlon In Latín Amerlca 1990 33,700 25,400 1,939 4,000 150,000 The second methodology used by 1he Program is that of collaborating with a variety of national and local institutions in integratad cassava production, processing and marketing projects. These projacts have been designad to anable improvements in cassava production, processing and marketing to be a vehicle for rural development in specific regions of priority countries. Through the davelopment of batter markets far cassava- based products (novel or improved) and with the involvement of farmers in organizations dedicated to processing fresh cassava, significant income and welfara improvements can be generated. As these projects require actions in areas as diverse as organisation of 250 farmer groups, credit provision, institutional strengthening, etc., in addition to the technical aspects of cassava production and processíng, coordination with a wide variety of collaboratíng institutions is necessary. Four stages have been identified tor integrated cassava projects: • Macro planning (at the national level): identífication of an appropriate production reglon, product to be marketed, etc. • Micro planning (at the regional level): ~ characterizatlon of cassava praduction, processing, marketing in the selected region; identífieation of sites far pilot activlties, etc. • Pilot stage: partieipatory testing of all project components in real world situations, including production, proeessing and marketing aspects, plus the ~ institutional, organizational and eredit actions required to obtain success. • Replication, expansion or commertial stage, in which the technologies and olber aspects tested in thé previous stage are expanded and consolidated at a commercial level. 13.1.1 Country prloritles Over the pest 5 yr the Program has focused its attention on 4 countries: Brazil, Paraguay, Colombia and Ecuador. Together these 4 countries account for over 90% of the cassava produced in Latin America. As the largest cassava-producing country, Brazil is an obvious priority far the Programo The size of the eountry is such, however, that several distinct cassava-production regions can be identified: from the semiarid NE region to the subtropical South. The NE region has been given the highest priority, both by CIAT and by Brazil itself, which located the National Cassava Research Program CNPMF in a state in this region. The NE produces 50% of the eassava in Brazil and has the largest concentration of rural paverty on the continent. Secondary priority has been given to the southem subtropical region. Paraguay, which produces 4 million t of cassava yearly, has tIle highest par capita consumption on the continent (> 100 kgfcapitafyr, even in urban areas). It is the most widely eultívated crop in the country, found on over 200,000 of the country's 250,000 farms. Colombia produces over 1.9 million t of cassava per year, mainly for fresh consumption. As the host country tor CIAT and the site of pilot testing of many of the Program's utilization and production technologies, it is in a good pasition to benefit from these activities. Economic studies have also demonstrated the potential for inereasing cassava 251 . ~ production and utilization in Colombia. The Atlantic Coast, which is the majar cassava- producing region in Colombia, is among the poorast regions of the country. Although Ecuador is a minor producer of cassava (150,000 tjyr), it is a major crop in the Manabf region. The feed requirements of the growing shrimp industry in this coastal region provided a ready market for the smaJl-scale rural industrialization of the crop. Contacts have also been maintained with Panama, Cuba, Costa Rica and Mexico through training and occasional technical visits. Argentina, Bolivia, Venezuela and Nicaragua are countries with an increasing interest in cassava development, and contact is expected to increase in the future. 13.2 Training and Networks 13.2.1 Training Table 13.2 gives a breakdown of training events by type and by country. In the early 1980s the training emphasis of the Cassava Program .vas placad on cassalla production and utilizatioh courses held at CIAT. By 1987 the emphasis had changed to participating in courses hald in the countries. Since 1988 annual courses have been held in Colombia, organized by DRI and PNR, resulting in a larga body of trained professionals in all institutions involved in integrated projects in Colombia. . Similar in-country courses in Brazil, Ecuador and Paraguay have all supported the growth of integrated projects in these eountries. Courses of a more specialized nature have also been stressed; e.g., diagnostic skills, experimental design and analysis, HCN analysis methodology and integrated pest control. Sinee 1990 training in integrated project formulation, implementation and evaluation has also been important. Other events (seminars, workshops, etc.) have been held at CIAT and in-country on topics such as cassava stareh, integrated project methodologies and the cassava-maize association. An annual cassava researeh seminar has been a feature of the imegrated project in Ecuador. Table 13.2. Cout$éS, workshops seminars and meeting held in Latin Arneriea. by country¡ 1987-91, Typo 01 Event c.otombiaJ """" Paraguay Eov""," ""'"' Pe .. ........ ""-' Course 6 6 4 3 4 2 3 Workshop 2 2 2 O O O Seminar O 2 O 2 O O 1 O Meeting 2 O O O O , Ineludes "".nls al CIA T. 2 Panama, Nicaragua. Guatemata, Bolivia, 252 Now that an adequate corps of trained professionals exists, Mure priorities will be focused on forming 3 regional training teams: Southern Cone, NE Brazil, and Central America, Mexico and the Caribbean countries. 13.2.2 Networks Many of the activities inherent in international collaboration ean best be earried out through networks of interested parties. In 1988 the Program organized a meeting at the request of several countries to found a eassava network for Latin Ameriea. Unfortunately this exercise was premature in that the human and financial resources required to operate such a network adequately did not exist. Since then, however, several smaller networks have evolved, focused on specific research or development topics, with more limited objectives. . 13.2.2.1 Pan-American Cassava Breeders' Network. The Cassava Breeding Network is described in Chapo 2, Sect. 2.2.2. 13.2.2.2 Southern Coné. A meeting was held in Paraguay in Oct. 1990 to discuss the possibility of collaboration among Argentina, Paraguay and the southern Brazilian states of Río Grande do Sul, Santa Catarina and Paraná. These cover one relatively homogenous subtropical eassava-production area with many similar production and utilization problems, which are frequently different from those of· other regions of the continent. CIAT HQ cannot carry out research of relevance to the subtopics. At the initial meeting, the current status of cassava production and utilization and the areas of R&D of interest to each party were presented. A further meeting in Rio Grande do Sul in May 1991 helped define these ideas in concrete proposals for forming a network with several components: germplasm, production, processing and utilizatíon, socioeconomics and technology transfer. Funding for the germplasm component, based at ltajai, Santa Catarina, Brazil, exists within an IFAD financed project. Also, within the technology transfer component, a project to form a subregional team for training technology intermediaries has just got under way with financing from the Interamerican Oevelopment Bank (IBO). For the other components, some specific topics of ¡nterest to the group of countries were agreed upon, and a further meeting is planned tor 1992. 13.2.2.3 Integrated proiects. Integrated projects exist in Brazil (Ceará), Colombia, Ecuador and Paraguay, with incipient projects in Panama, Bolivia, Argentina and other states of Brazil, (including Rio Grande do Sul, Bahia, Paraiba and Pernambuco). Informal contacts between projects have been occurring for a no. of years; but as a result of a course on integrated projects he Id at CIAT in 1990, it was decided to continue contacts on a more regular basis in the Mure. At that event, participants gained immensely from the experiences, both positive and negative, of other countries. A study visit to the Colombian integrated project was highly successful in opening up issues for discussion and in giving the participants an appreciation of technological advances relevant for other areas. A second workshop was held in Portoviejo, Ecuador in July 1991, at which the 253 pro/ect advances during the previous year were discussed and 2 special issues dealt with: tarmer organization and monitoring, evaluation and impact of projects. One feature of this event was the leading role of FUNDAGRO·-the Ecuadorian agricultural development institution supporting the project there--in organizing and funding the meeting. The Ceará project funded the participation of extension agents and tarmer leaders. In this way the Integrated Projects Network has already developed certain financial independence. A further meeting in 1992 was tentatively arranged to take place in Ceará. 13.2.2.4 Utilization research. Given that research in processing and utilizatíon of eassava fal)s outside the mandate of many national programs in Latín Amerlea, the Cassava Program has attempted to build links with research institutions of a diverse nature in Latin America. Strong links have now been established with UNIVALLE (Cali, Colombia) and with the State U. of Sao Paulo (Brazil). UNIVALLE has been a collaborator of the Cassava Program in product development of cassava flour and starch since the early 198Os. The Botucatú campus of the State U. of Sao Paulo has been active in encouraging cassava utilization research at a wide number of Brazilian universities and research institutes. A working group now exists to coordinate research on cassava utílízation in Brazil. Other universities in Ecuador, C()lombia, Argentina and Costa Rica have more limited interests. In the developed world, CEEMAT /CIRAD of France and NRI of the UK are both earrying out and supporting research on eassava utilizatíon focused on Latin Amerlca and Africa. These instítutions have close contacts wíth CIAT. An informal network of institutions interested in research on cassava starch has existed since an intemational workshop on this subJect was held in Santa Calarlna, Brazil in 1989. A second workshop was held at CIAT in 1991, with the partícipatíon of starch producers and industrial users as well as researchers. This grouping of interested par1ies is the 'basis upon which a research network with a broader.scope can be built. The new CIAT strategic plan envisages a reduction in the Cassava Program's product development activities. A significant potential for product development research still remains, especially as regards the development of secondary or derived products from starch and fIour. It is the Program's intentíon to ensure that the necessary research is undertaken by other institutíons, through the financing of special projects where necessary, so that the demand lead progress already achieved with eassava during the 1980s can continue beyond the year 2000. CIAT will continue an active role in the interface between production and processing (Le., raw material quality) and in determining research priorities based on market potential. 13.3 Country Reports Chapo 14·17 describe in detail the progress mada in Colombia, Ecuador, Brazil and Paraguay, countries on which the Program has focused its collaboration for the past five years. The situation in Panama, Mexico, Cuba and Costa Riea is briefly described below. 254 13.3.1 Panama Cassava ís of great economíc ímportance for small farmers and rural-urban consumers in Panama. Several typical Panamenian dishes are based on cassava. As most of the population is concentrated around the. capital, both the poultry and swine industries have increased significantly in recent years, resulting in a demand for balanced feed rations. The national R&D institution (IDIAP) has dedicated significant effort to cassava, despite political problems. As a resuit of development and promotíon activities undertaken in collaboration with the CIAT Cassava Program, cassava now plays an ímportant role as a component of feed concentrates far poultry and swine; and demand is on the increase. In addition the private sector has becoma more interasted in developing the crop for agroindustrial purposes. Technical assistance provlded by the Program has concentrated on supporting the Integrated development of the crop. This requires simuitaneous work on production, transformation and co¡nmercializatíon. Initially, IDIAP, and CIAT economists conducted a macroeconomic studly on the potential of .cassava in the country, as well as that of other carbohydrate-producing crops. This study damonstrated the importance 01 tresh cassava in the human diet and identified tha potential for using cassava In rations for poultry and swine. Subsequent regional analyses indícated that production was concantrated in Ocu (Herrera), leading to the selection of this region for R&D activities. The principal production systems were characterlzed, and problems/opportunities far improving production were ídentified. Possibla sites for lo.cating a natural drying plant to be aparatad by larmers on a pilot basis were also identified; and trom thase "Los llanos de Ocu,» a farm run by a farmer coop, was selected. An in-dapth study 01 potential markets was initiated. Possible bU)Íars 01 dry cassava chips within the feed industry were identified, and studies on raw material production costs were IXldertaken en the basis of real production figures and thaoratical purchase prices for the producto Simultaneously, studies were conducted on the physical characteristics of production. One 01 tha principal problems identified on traditíonal Jarms was the Iow fertility 01 the soils and a high incidence 01 weeds. Yields of roots and of planting material were low; leading to serious problems of availability of planting material. Based on these findings, it was .concluded that the potential demand for cassava was much greatar than what larmers could produce, given their current levar of technology and access to land. In'dustrialists' lack 01 knowledge 01 cassava as a raw material was identified as a limiting factor for marketing. In addition, the distan ce trom production centers to urban centers and feed industries was a critical problem for marketing fresh cassava and dried chips, given the relatively high cost of transportation in Panema. 255 Oespite relatlvely low commercialization margins, the pilot plant 'Los Uanos' began operating successfully and was imitated by other smal! farmer associations as well as prívate ínvestors. Today there are 7 drying patios in the region .. The local feed índustry ís purchasing chips despite the problems that have affected the country's entire economíc sector. The existence of a new market for cassava has stimulated production, and the farmers are demanding improved technology. Technícal recommendations on planting density, stake selection, fertilization and weed control have íncreased production considerably. Progress has been made on the use of minimum tillage to prevent surface erosion of easily saturated soils. Technical recommendations are also available for intercropping cassava with maize and cowpeas, without an adverse effect on yield. The introductíon of improved varieties has been successful. The local variety "Brasilet'la: introduced many years ago, shows symptoms of systemic pathogens. "Dayana,' introduced by CIAT in 1984, Is one of the best ylelding varieties in the country (20 t/ha avg vs. 15 t for the local check). Wlth technical assistance from CIAT, a complete germplasm testing scheme has .been established, together with the rapid multiplication of that germplasm based on tissue culture. Training technicat personnel has been a constant concem of the project. In the last five years, more than 20 technicians trom the state and prívate industry have been trained in CIAT courses. At the locallevel, IDIAP and MIDA have tralned many small-scate farmers in production, operation of the drying plants and commercialization of the dry chips. The challenge for the future líes in coordinating the public and prívate sectors so that they complement each other in production, transformation and commercialization of cassava. 13.3.2 Mexlco Despite being one of the supposed canters of origin of cassava, production of the crop in Mexico Is minimal (ca. 3000 ha). In the 1970s, however, at a time when Mexico was importing increasing volumes of feed grains, a great potential for cassava was envisaged as a partial substitute for these imports. Accordingly, a national cassava research program was created within the National Institute for Agricultural Research (INIA, now INIFAP) in 1977, based in Huimanguillo, Tabasco. The aim was to develop cassava production, processing and utilization technology for commercial-scale production on the acid. infertile soils of the savannas around Huimanguillo. A young and highly motivated team of scientísts was formed; and by 1980 it was considered that sufficient progress on variety selection and related production teehnology had been made to initiate an ambitious plan for promoting cassava production arnong the farmers of the region. After a rapid initial expansion in cassava production trom 100 ha in 1981-82 to a level of 2404 ha in the 1984-85 oyele. the area under eassava has slowly díminished; and today 256 tha crop has been virtually abandoned, both by farmers and by official state organizations. INIFAP's national cassava program has been virtually disbanded. This situation, which contrasts sharply with that occurring in other countries, has meritad an analysis of the technical, economic and institutional factors that led to the demise of a program that held such promisa 10 yr ago. Among the principal reasons identified as affecting the program's future were: • An overinvolvement of and control by state agencies of aspects relating to cassava production and processing. Production and processing input and output decisions, market outlet and marketing strategy definition and profit distribution decisions were all in the exclusive domain of program officíals. Participation of farmers in decision- making processes was almost totally absent, resulting in a rapid 1055 of commitment and motivation. • Cassava production arid profitability were lowar than expectad, principally because of untimely delivery of state-controlled inputs and sarvices. • Too great an emphasis was placad on production--both in research and in promotion activities--and not enough on processing altematives and market identification and development. Unks among and integration of these components were almost totally IaCking. The model for cassava development in Huimanguillo was quite diffarent from that being adopted at the same time in other Latin Amarican countrles, such as Colombia, Panama and Ecuador. Two factors stand out: (a) the very high level of stata intervention and investment, and (b) the absence of mechanisms for validating and adapting both production and processing technology prior to promoting it among a larga numbar of farmers. 13.3.3 Cuba CIAT's relationship with Cuba has been quite different as the Cuban R&D instítutions are stronger than those of the other Caribbean countries. Consumption of root crops, aspecially cassava, is a tradition in Cuba. Close relationships have been maintained with the national program, but afforts haya concentrated on training technical personnel and broadening the germplasm base. Cuba is an active participant in the Pan-American Garmplasm Natwork, organized in the last few years. Previous annual reports haya described tha so-callad "Colombian" system f9f producing cassava, employed in Cuba for many years. This system is a mixture of Cuban technology and tachnology developed at CIAT. As a result the country has become self- sufficient in cassava, and a good part of the production is being processed for domestic consumption. 257 13.3.4 Costa Rica In Costa Rica the consumption of cassava is relatively low. Together with the Dominícan Republic, Costa Rica dominates the export market of cassava for the US and Europe. The country exports more than 30,000 t!yr of cassava to the USo The roots are either treated with paraffin wax and exported fresh; or peeled, sliced, packed into polythene bags and frozen. There is strong government support for infrastructure, working capital and marketing contacts. The dependence on a single var. (Valencia) means a hígh degree of risk for exporters. The CIAT Gassava Program is collaborating with the National Gassava Commission on identifying and characterizing Costa Rican germplasm, which was mostly introduced from CIAT, in arder to íncrease its díversity. Parsonnel haya also been trained at CIAT HQ. Future cooperation will focus on designing research to identify environmental and veríetal factors that affect the quality of the fresh roots. Ex-ante economíc analyses do not indicate the feasibility of transforming fresh qassava into dry chips as the cassava- producing zones are located in areas with high rainfall. 258 14. COLOM BIA Being located in Colombia, the CIAT Cassava Program has naturally had a very spaclal relationship with national ínstítutions, different from that formad with institutíons in other Latín American and Asian countries. The diverse cassava productíon and utilization situations in Colombia have provided the ideal ground for testíng and adaptíng both production/processíng technologies and partícipatory R&D methodologíes. The practical work carried out in Colombia has served tha Program in its endeavors in other countries by providing a learning experience for Program staft and visiting research and extension personnel from other countries. In this respect, the Program acknowledges with gratitude the excellent collaboration that it receives from its Colombian counterparts in a wide range of institutions. 14.1 DRI¡CIAT Integrated Project, Atlantic Coast of Colombia 14.1.1 Background Colombia was the first country in which an integrated cassava production, processing and marketing project was initiated. The project now has a history of 10 yr in the Atlantíc Coast region of Colombia. The role of the Cassava Program has evolved significantly since the early 19805. The Colombian institution most closely involved in the project has been DRI, the Integrated Rural Development Fund of tha Ministry of Agricultura. CIAT has maintained a series of agreaments with DRI since 1981, through which all the project- relatad activities of the Program have been financed. DRI is the overall coordinatir)g institution for project activities although numerous other institutions are involved, for example: .. CORFAS: provision of credit, technical assistance in accounting and processing .. FINANCIACOOP: credit provision and supervision .. ICA: cassava production, research and extension .. INCORA: land reform .. SENA: coop formation and consolidation, some technical assistance in processing In addition DRI supported the formation of a second-order organization in 1986, grouping the coops and associations of small farmers into one Natíonal Association of Cassava Producers and Processors, ANPPY .. From 1987 to 1991, the Cassava Program collaborated with DRI and tha other institutions and farmer organízations in the followíng activities: 259 14.1.2 Marketing Untíl 1989 technical assistance was provided to ANPPY for dried cassava commercialization. Following the withdrawal of CIAT technical assistance in 1989, ANPPY has been essentially self-supporting in this area. Surveys of fresh cassava consumption and marketing were carried out in Cali, Medallrn and Bogotá in collaboration with CORFAS. These provided information on consumption habits (purchase frequency and amount, varietal prefarences) and, market informatlon (market share by retail outlet, margins) af relevance to the possible future expansion af the fresh cassava storage technology to other Colombian cities after the Barranquilla pilot experience. These studies, combined with those conducted previously in Bucaramanga and Barranquilla, provide a complete pidure of fresh cassava marketing and consumption in the principal urban centers of Colombia. 14.1.3 Processing Most of the CIAT activities carried out within the framework of the DRI/CIAT agreement have been at the pilot level (see Chapo 12 on product and process development). These include improvement to chipping machine design, animal feed ration formulation and testing at farm level, fresh cassava storage and cassava f10ur for human consumption. As tIlase activities formad an integral part of the DRI-coordinated actions in this project, thay received considerable institutional support. Technical assistance to the tarmer groups in pracessing and commercíalizing drled cassava was initlally a CIAT activity; but this was gradually handed over to natianal instítutions. CIAT ceased ta perform this function in 1989; however, it was not straightforward to find a suitable Institutian to take ovar this role. The marketing of dried cassava chips has been assumed by ANPPY, which organizes the price and volume negotiations with the animal feed companles on behalf of the farmer groups affiliated to n. As this now involves contacts with over 15 companies in at least 5 cíties in Colombia and the dispatch 01 cassava from over 50 drying plants, the operation has become exceedingly complex. Problems still occur with regard to chip quality (poor drying). Technical assistance has been devolved to CORFAS, as tIle agency with the closest and most fraquent contacts with the drying coops; however, this has not been entirely satisfactory as CORFAS is essentially a credit provision and supervision agency with a limited number of technical staff.· In addition, recent problems with the national system for distributing credit lines for small-scale agriculture have resulted in the reduced presence of CORFAS in the project region. Creative approaches to solving the problem of technical assistance to farmer groups in the project are urgently required. Durilg the 2 yr in which CIAT was involved in technical assistance to the cassava drying plants, 7 new plants were constructed, and many others expanded the area of drying Iloor. In total 41 coops, as well as ANPPY, received assistance. The presence of DRI- fundad CIAT personnel in clase contact with the drying plants provided an excellent 260 opportunity far firsthand knowledge af any remaining problems, which served as feedback for research at CIAT. In thís way, the need far an improved cassava chipper was identified, and the improved modal was testad with the farmer groups themselves for firsthand reactions. 14.1.4 Producllon Tha integrated project on the Atlantic Caast ¡nitiatad with a strang emphasis an processing and marketing given that the declining and unstable demand far fresh cassava was the immedlate constraint facing small farmers. By the late 1980s, hawever, this had led to increased interest in improved production technology by the small farmer beneficiaries af the praject. Whereas processing technology can be relatively straightforward to introduce and suppart within a project context, the devalopment and diffusion af improved production technalogy requires a longer timeframe. Wrth this in mind, during the mid 1980s ICA gave increased emphasis to cassava production research in the project regian. thraugh the formation of a warking group of researchers who jointly planned, executed and analyzed research results. togethar with CIAT. This Group of Cassava and Assoeiated Crop Researchers (GRUYA) has proved an excellent forum for discussing rasaarch methodologies and for obtainíng results with relevanca to the whole region, as similar experiments can be carried out by a no. of researehers over a wide area. One result af this has been the increasing emphasis an participatory researeh with farmers. especially for varietal development (see Germplasm Improvement. Chapo 2). Similarly. the widaspread use of PPPs--planted and managed in collaboration with farmers--to test promising technology reeommendations. has resulted in a no. of significant findings (~ee Cropping Systems, Chapo 9). In the stmes of Suere and Córdoba, cassava production has increased dramatically in the last few years. As most cassava is still produced by the small farm sector with Iimited land available for crop rotation or fallow, increasing pressure is being placad on the natural resource base of the region. The GRUYA group ls placing more emphasis on soil fertilitY maintenanee and other aspects of sustainable productions systems. 14.1.5 Farmer organization In 1986 DRI supported the formation of ANPPY, the second-order organizatian of farmer groups involved in the cassava integrated project. Initlally, ANPPY's major function was to organize the commercializ-ation of dried cassava produced by the first-order groups in the different end markets. It has been very successful in this activity, holding periodie price negotiations with individual companies and obtaining favorable results for the farmers. ANPPY is financed through a 1% levy on sales. Given the coneentration of farmer groups in the states of Suere and Córdoba, however, eoops in other states were not so well attended. Thus in 1989, several coops from the states of Atlántico, Magdalena, and Bolrvar formed a separate second-order organization, FAGROCOL. in order to commercialize their dried cassava, whlch was usually sold in a different terminal 261 market from thet 01 the remaining ANPPY coops. FAGROCOL also contained coops organized around the production and commercialization of other crops (yams, maize, tomatoes, etc.), as well as being the executing agency of the fresh cassava storage project in Barranquilla until 1990. During 1989 the tirst cassava drying plants were constructed by prívate individuals--all in Sucre and Córdoba. In 1990 ANPPY received the first two affiliates from this private sector. By July 1991 the no. of prívate-sector affiliates of ANPPY had increased to 31, compared with 34 coops and associations. The rapid and dynamic growth of privately operated and owned cassava drying plants, whife amply demonstrating the economíc profitability of cassava processing under increasingly free market conditions, has been a source of considerable tension within the project itself. Collaboration between the private and coop secters is advantageous in terms of price negotiations with the animal feed companies and for producer lobbying at a political level, regionally and nationally. Private-sector plants have rapidly improved process efficiency, and the coop plants can learn from these experiences; however, there is tension related 10 the distributíon of beneflts, wíth the coop sector favoring a high raw material price (small farmer benefi~) and the prívate plants a lower price in order to maximize margins (processor benefits). As a result ofthis tension, a no. of coops, especially newly formed ones, have left ANPPY to form a separate organization for commercializing dried cassava (ASOCOSTA); and institutional support to ANPPY has declined. Despite these problems, the no. of drying plants and their productíon has continued to expand at the same rate as in 1990. The Cassava Program has collaborated with both ANPPY and FAGROCOL in the development of the cassava flaur and fresh cassava pilot projects, resp. (see Chapo 12). 14.1.6 Inatltutional aspeCla Clase collaboration between CIAT and DRI has been a salient feature of the project, both in the Atlantic Coast region and at the national level (Bogotá). In each state DRI coordinates technical assisíance teams composed of representatives of each collaborating institutíon. CORFAS and SENA have been -the most active members of these teams, with functions of credit provision, supervision and technical assistance in processing, and in coop formation, organization and accounting, rasp. Technical assistance teams have worked well in the principal project areas of Sucre, Córdoba and Atlantico, but have been less active in other states where fewer plants existo Some regional institutions (e.g., FIDES, a Sucre development foundation) have recently become involved. ICA has taken on an increasingly important role now that demand for improved production technology, the release of a new variety in 1991, and the success of the GRUYA group in carrying out farmer participatory research have all combined to make cassava production a high-prierity issue. 262 14.1.7 Monitoring and evaluation (M&E) Initially CIAT carried out the project M&E in collaboration with DR!. In 1989 DRI and CORFAS became solely responsible. The dynamie nature of the growth of the cassava drying industry sinee then has not been fully captured by the M&E system currently in operation as CORFAS monitors the performance only of the plants to which it provides credit. Delays in publishing the results have reducad their usefulness. In effect, tha drying of cassava for animal feed has passed from being a projact to being an autonomously expanding agroindustry. The project M&E system has thus proved inadequate to cope with this expansiono In addition the spontaneous growth of driad cassava plants outside tha project region has also failed to be captured. Monitoring information is currently available to CIAT from the following sources: .. CORFAS: coop plants receiving eredit .. ANPPY: affiliatad plants··eoops and private .. Animal feed companies: volumes purehased nationalfy .. Processing equipment manufacturars: purehasers of equjpment The lack of an integrated M&E system is not only disadvantageous to the davelopment of the Industry itself; but jt also makes the assessment of project impact-both within and beyond thE\ project boundaries-·difficult to carry out. This wilf be one of the priority araas to coordinate actions with all interested parties in the coming years. 14.2 PNR/CIAT Collaborative Pro]ect Tha National Rehabílitation Program (PNA), which ís a special rural development prograrn operating out of the Ministry of the Presidency in Colombia, has as its responsibiiity those areas of the country that for diverse reasons are suffering from pubnc unrest. These are arnong the poorest, most isolated and least developed regions of the eountry. As such, cassava is an important or potentially important crop lo many of these areas, where poor market access Is a common feature. Following the success of the DRI/CIAT collaboration in developing a cassava drying agroindustry on the Atlantic Coast, the PNR proposed a similar agreement to expand dried cassava production to other regions of the country. Unfortunately the absence of a Colombian institution able to provida technieal assistance in cassava processing meant that CIAT was the only alternative. Accordingly, it was agreed that the PNR would finance technical assistance by CIAT in 4 regions of Colombia for 2 yr. At the end of this period, responsibilities would be handed over to national institution(s). 263 Tha areas for expanding the dried cassava industry wara agreed batwean CIAT and PNR, with the active participation of DRI, which also has national coverage, although with an amphasis on small-scala farmars. The states agreed upon were: .. North and South Santander .. Cesar .. Sucre and Córdoba (in areas not covered by DR!) .. Meta The activitias and achievements in aach ragion ara detailed below. Despite tha fact that all activíties were to be relatad to the expansion of dried cassava for animal feed, in effect potential for the fresh eassava storage technology and the need for improved varieties and planting material have resulted in the broadening of project actíons in all regions to inelude other aspeets of proeessing and production. 14.2.1 Cesar Cassava prodi.¡ction in Cesar fell from 121,000 t in 1980 to only 56,000 t in 1987. It then increased again. reaehing 127,000 t in 1990. The deerease in production during the early 1980s was due mainly to market problems. In contrast to other areas of the Atlantic Coast region, land for cassava production is not a limiting factor: activíties in the PNR/CIAT project were therefore focused on the potential of drying cassava for animal 'teed to provide a seeure marketfor cassava and hence encourage production ¡ncreases. Four drying' plants had been built and operated by farmer groups within the framework of the DRI project from 1984-88; however, these were operating at low levels of capacity for fack of institutional support, poor coop organization and financial problems resulting from poor administration and late arrival of credit. 14.2.1.1 Technical team. The team is coordinated by PNR; members inelude ICA, SENA, CECORA, Caja Agraria (the rural agrarian bank, which also provides extension services through EDO units). INCORA has bean an intermittent member. CECORA left the team in 1991; since then thare has beao no iostitution responsible for technical assistance in marketing of dried cassava. 14.2.1.2 Processing. Initially thera was a shortage of funding fer the construction of drying plants in Cesar; but a project was finally approved in 1990, and one additionaf plant initiated operations in 1991. The plant is operated by a coop of 200 membars (COOTRADECO), which has several years' successful experience commercializing maize. Lack of institutional assistance in this area is not a limitation. Technical support has baen given to the existing coops, and sorne reaetivation was possible although the lack of financing tor working capital has been a limitation. Interestingly, while the institutional 264 · financing for working capital has been a limitation. Interestingly, while the institutional sector of Cesar has been relatively weak, drying of cassava chips has been spontaneously expanding through the activities of individual small-scale farmers, who have constructed small drying areas and purchased manual chippers. COOTRADECO has been purchasing dried chips from these farmers and selling them, together with their own production, to animal feed companies in Bogotá. Projects far constructing 4 more drying plants with existing coops are under consideration by PNR. Experiments were also carried out with the fresh cassava storage technology, using cassava harvested from the Sierra Nevada region of the state. Storage times of 3 wk were obtained. A test market was carried out in Valledupar, in which 14 t of bagged cassava were sold in supermarkets. The project has not progressed further for lack of working capital for the coop. 14.2.1.3 Production. When the PNR/CIAT agreement started, no cassava-related activities or research was being conducted by ICA-Cesar. Two years later, cassava has been recognized as one of the strategíc crops in the state, and all the ICA stations (CRECEDs) and Caja Agraria EDOs have tríals in pregress. Trials have ínvolved the testing of 10-15 var., including local and CIAT germplasm. using the participatory research methodology. Several promising materíals have been identified as a result of these farmer-managed and evaluated trials. 14.2.1.4 Courses and other avents. A coursa on cassava production, processing and marketing was held in Valledupar in March 1991. with the participation of personnel (30) from institutions in all regions of the PNR preject. Field days have been held with farmers and institutions on 4 occasions. 14.2.2 Sucre and C6rdoba Although Sucre and C6rdoba are the 2 states in which the DAI cassava project has had the greatest concentration of activities, there remain areas (not covered by DRI) where drying plants have not been built even though cassava is an important crop. As the PNR program has different objectives and priorities, it was possible to expand preject coverage. Cassava production in Sucre and Córdoba has been increasing significantly. reaching 170,000 and 189,000 t in 1990, resp; however. these production increases have occurred in areas where drying plants exist or where access to the fresh market is good .. not those in which the PNR operates. 14.2.2.1 Technical team. The initial strangth of the technical teams in Sucre and Córdoba was recently weakened by the departure of the institution responsible for technical assistance and training in marketing (CECORA). In addition to providing support to existing coops. the technical teams have been active in promoting the formatíon of new coops and associations. 265 14.2.2.2 PrQCessing. During the 2-yr perlod of the agreement between CIAT and PNR, 4 drying plants were built using PNR finance. Five additional groups have been legally constituted, and projects for constructing plants prepared. Other groups are in the process of formation. The 4 drying plants have been operating at a 75% avg level of capacity. These plants are lacated in the least developed parts of the 2 states, including areas with significant guerrilla activity. The success of the project in bringing a viable economic altemative to these torgotten regions of the country cannot be underestimated. The lacation of drying plants in these isolated communities has also resulted in a greater presence of government institutions than previously. 14.2.2.3 Production. Given that the ICA/GRUYA group is active in these states, no specific activities were carried out in the PNR project on cassava production. A project has been prepared for financing seed multiplication plots far the PNR regions of the states. 14.2.2.4 Courses and other events. Personnel from Sucre and Córdoba attended the courses organized in Meta and Cesar in 1990 and 1991, resp. Many events were held with farmer groups related to the promotion and formation of coops. 14.2.3 North and South Santander These 2 states contain several important cassava-production regions: 186,000 and 90,000 t in South and North Santander, resp. When the project started in June 1989, cassava drying plants already existed in the Magdalena Medio region of South santander, a PNR area. However the tarmer coops operating these plants had suffered from consistently deficient technical assistance in almost all aspects 01 cassava production, processing and marketing, as well as in coop formation and management. In South Santander, a farmer group was experimenting with conservation of fresh cassava in bags using the CIAT- developed technology, marketing the roots in Bogotá. In North Santander, small farm production of cassava chips by hand chipping and drying on any available surface, had already been initiated spontaneously. Intermediaries were purchasing these small volumes of chips from individual farmers and selling them in Bucararnanga. 14.2.3.1 South Santander • Consolidation 01 technical tearns. There were two technical teams in the state: one coordinated by PNR for the Magdalena Medio region and the other by DRl for the region of Socorro. The former comprised the following ínstitutions: INCORA Qand reform), SENA (coop organization), ICA (cassava production), Agriculture Secretariat of the state, DAI, FONDISER (a regional NGO). Technical assistance in commercialization was and still is lacking. This has been a serious constraint to project expansiono The PNR/CIAT convenio provided technical assistance in cassava processíng.. The DAI-coordinated team far the Socorro region was similarly constituted, but with the addition of CORFAS tor technical assistance in 266 commercialization. The technical teams meet every 4-6 wk, with joínt sessions every 2 mo. • Processing. Technical assistance was provided to 3 coops, each with 500m2 of drying floor. An additional group, awaitíng the approval of a project to finance construction of a drying plant, has already processed cassava manually. In Socorro, assistance was provided to a coop originally promoted by DRI, SENA and ICA, in which the main activity is conserving fresh cassava for the Bogotá market. This group is marketing approx. 3-4 tfwk, year-round. Cassava is selected into three quality grades: the highest quality is treated and packed for the Bogotá market using the CIAT technology, Grade B is sold on the local fresh market, and Grade C (noncommercial) is chipped and dried far animal feed use. The total cost incJuding transport, finance charges and administration is Col$119fkg cassava, whíÍe the price received from the supermarket clients in Bogotá is Col$150fkg (US$1 = CoLPs.600). Evan at the low level of vol. traded in Bogotá, this coop has been able to make substantial profits in one year of operations. Two recent problems are the appearance of the Cyrtomenus bug (which affects root quality) and the lack of knowledge of the product and its management by the elients in Bogotá. • production. In collaboration with ICA, varietal trials with CMC 40 and CMC 76 were planted in the Magdalena Medio region. Wlth the support of a no. of PNR technicians, 16 plots (0.25 ha each) were plantad throughout tha Magdalena Medio region using local varo and improvad production technology (soil preparation, stake treatment and selection, weed control). These plots are the ¡nitial phase of a PNR project that will finance the planting of 92 plots with farmer collaboration. • Courses and other events. A course on cassava production and processing for 29 technicians of the state was held in June 1990. Talks and lectures were given to students at local universities, and informative leatlets on cassava processing were produced. 14.2.3.2 North Santander • Technical team. When the project started, there was no team although one drying plant had recently been built by a coop with funds from a regional development corporation, CORPONOR. The technical team is constituted by the PNR, DRI, ICA, the state Secretariat of Development and CORPONOR. CECORA (responsible for commercialization) and SENA (coop organization) have been intermittent members. Meetings are held every 2 mo. • Processing. In addition to the plant financed by CORPONOR, 5 small (200m") plants have been financed, 3 of which are currently under construction. Farms are disperse, which hinders associative forms of drying. Many small farmers dry hand-chipped cassava as a backyard actívity, selling to local intermediarías. The idea behind the 267 small plants is to facilitate some associative drying and also to permit the coops to take an active role in assambling dried cassava from small farmers and marketing this in Bucaramanga or COcuta (project awaiting approval). Trials of tha fresh cassava conservation technology ware also carried out with ICA using local varieties. Storage times of 4 wk were obtainad with good aating quality. • Production. In collaboration with ICA, 7 CIAT hybrids and 2 local control varo wera plantad in Oct. 1990. In addition plots of 2 local and 4 CIAT varo were planted for seed multiplication. Four damonstration plots (0.25 ha each, containing 3 local varo and 1 CIAT hybrid) were planted in the vicinity of drying plants in collaboration with the.Stata Secretariat of Developmant. Four more plots were planted in 1991. • Courses and other avents. A one-day course on cassava production and processing was held in Cúcuta in Feb. 1991, with 33 participants from 6 institutions. 14.2.4 Meta The area devoted to cassava production in Meta fell from 20,000 to 4,000 ha from 1980- 89. This was due to !he fall in demand in Bogotá, the principal market for fresh cassava, where o!her production regions of the country have increased !heir market share; however, demand for dried cassava chips in !he animal feed industry in Bogotá was estimated at 20,000 t in 1990. In order to supply this market, actions were required not only in the area of cassava processing but also in production. • Technical team. Upon the initiation of the project in 1989, a technical team was constituted under the coordination of!he PNR. Member institutions 'are ICA, the State Secretariat of Agriculture, CECORA and INCORA. 14.2.4.2 Processing. In 1989 a coop was formed to dry cassava for animal feed at Alto Casibare. Financed by PNR-Meta, operations began in Oct. 1990. The Piedmont area of Meta is characterized by high. well-distributed rainfall; thus the plant was designed to include artificial drying as an option for processing when conditions for natural drying were unsatisfactory. This plant was built in an area where tittle cassava is cultivated, but where farmers have good land availability to increase production. Thus, although the process has operated efficiently. there has been insufficient supply of fresh cassava to permit continuous oparation; however, now that farmers have a secure local market for thairfresh caSsava, production should increase. A further mixed natural/artificial drying plant was approved and built with PNR funding at Cerritos in early 1991. Trials of the cassava consarvation technology ware carried out together with ICA. The varieties tested included the two racently released by ICA. The bags of cassava were evaluated by housewives in Villavicencio, the main city in Meta. Results were excellent avan after 30 days' storaga. 268 • Production. Once the lack of cassava production to supply the drying plants was identified as a major constraint, considerable emphasis was placed on establishing seed multiplication plots. The scarcity of stakes was a major factor limiting expansion of production. Multiplication plots were established at 7 sites, including ICA experiment stations, the U. de los Llanos, the State Secretariat of Agriculture field station and in farmers' fields. Over 100,000 stakes were planted (25.000 in 1989 and the rest in 1990) for distribution to members of the coops and other farmers in the region, as well as for use by ICA and other institutions to increase further their supply of planting material. Stakes of the 2 recently released varo (Catumare and Cebucfln) were in particular demando • Courses and other events. A course on cassava production, processing and marketing was organized in Granada (April 1990) with the participation of 31 representatives of different institutions and members of coops, not only from Meta but also from other regions of the PNR/CIAT project. This course was repeated in June 1990 in Arauca, where COAGROARAUCA was responslble for organizing the event. A third course was held in Villavicenclo in July 1990, in collaboration with the U. de los Llanos and DR!. These events have served not only to improve the participants' technical knowledge, but also to assist in the formulation of projects for constructing and operating cassava processing plants by coop groups. 14.2.5 Future The Initial project with PNR terminated in May 1991. Although the original idea was to transfer CIArs responsibilities to other institutions at that time, this was possible in only one of the 4 regions: Santander, where FONDISER, an NGO whose objective is the agroindustrial development of the region, was identlfied. An agreement t:ietween PNR and FONDISER is now operational, and the area of activlties has been expanded to include adJacent regions of other states. In the Atlantic Coast region and Meta, no national Institution could be identified; thus it was agreed to contlnue CIAT activities on the Atlantic Coast for 6 mo more (Dec. 1991), att which time it Is hoped that PNR will contract ANNPY to take over this responsibility. In Meta the PNR-funded CIAT assistant left at the end of the original 2 yr, and it was decided not to replace him. Activities have therefore ceased in this state; but If a national executing agency can be identlfied, actlvities will be renewed. 14.3 Cauca Starch Project In 1990 an integrated cassava project was begun in the state of Cauca. This project is organized and financed by Colombian national institutions and NGOs. Cauca has a production of approx. 30,000 t of cassava, 95% of which is used as raw material for the small-scale production of starch, both sweet and fermented, sour starch. Over 170 farnily owned starch plants exist, providing direct and indirect employment for 5000 families in 269 rural areas. Th/s /s the third most important economic activity of the state, after sugarcane and coffee production and processing, with a large percentage of the benefits aceruing to the small farm sector of the population. Project objectives are to improve (a) the sustainability of cassava production through betler management of natural resources; (b) the efficiency and quality of the starch produced from cassava, hence expanding food and industrial markets; and (e) the marketing system for cassava starch so that farmers and processors obtain a larger percent of the margino 14.3.1 Production These activities are red by the Cauca Valley Corporation (CVC), a body responsible for the conservation of natural resourees in the cauca Aiver watershed. The Cassava Program is participating with the development of erosion control and soiJ fertility management technorogies, for testing at the farm level. CIAT germplasm is also being tested in this o/ea, and varieties suitable for starch extraction are being identffied. 14.3.2 Processing The Cassava Program is closely involved in the procass improvement component of this project, which is led by a local NGO, SEDECOM. The improved equipment devefoped in the CEEMAT ¡CIRAD collaborative project with CIAT is being testedin two pi/ot plants (a coop and a private plant). The basic research being conducted on sour starch by CEEMAT will also feed into this project. 14.3.3 Marketing This component is led by FINANCIACOOP, an institution that provides credit and business advice to coops. A coop has been formed to link the small-family based starch producers into one marketing organization, thereby reducing their dependance on intermediaries. The coop has also been able to improve and standardize quality of the final producto 14.4 Other Reglons 01 Colombia Cassava drying plants are in operation in several other states of Colombla. In cauea, a United Nations Coca Substitution project has financed the constrl!ction of 4 drying coops. After a sfow start these are now- operating at a high level of capacity. CIAT was involved in site selection and periodie technical assistance to this project. The drying plants are also used to produce dried plantaln for sale as animal feed. In the Urabá region of Antioquia, two drying plants are in operation and more are under construction. One pilot plant for artificial drying is under evaluation in this region, supported by a local NGO. Further coop and private drying plants are in operation in Caquetá, Arauca and Antioquia. 270 Although the Cassava Program has played a minor role in these regions of the country, many ofthese plants are using CIAT-designed equipment, now available commercially in Colombia. 14.5lmpact The adoption of production and processing technologies by farmers in Colombia is detailed in Chapo 25. 271 15. ECUADOR INTEGRATED CASSAVA PROJECT The Ecuador Integrated Cassava Project, initiated in 1985, was conceived as both a social and technical experiment. Cassava chipping and solar drying technology from Thailand had been tested and proven successful on the Atlantic Coast of Colombia in a collaborative effort by CIAT and Colombia's integrated rural development office (DRI); however, the institolional cos15 were quite high. The chaflenge for CIAT was to replicate the technological success of the Colombian experience successfully, but at a lower institutional cost. This required different institutional and organizational arrangements, emphasizing new actors: farmer organizations and national program staff in the field, from extensionis15 to farmer-promoters.'''' While sorne might consider the transfer of the cassava processing technology from Colombia to Ecuador as "merely extension, not research," quite the opposite is true, being decidedly experimental. The basic technology transferred was the same as that used in Colombia; bol the method of transfer and the new social and institutional environment were radically different Once transferred across tha border, a process of changa and adaptation began that continues today. The initial technology has evolved into new activities in response to new demands from farmer /processors and cassava product users. This in turn requires researchers to. address new issues such as product quality, marketing alternatives, greater efficiency and new processing technology and equipment. This report summarizes the changes in the social, organizational and institutional context embracing the processing technology in Manabl Province, highlights ongoing research activiti,es and results, and lays out possible pathways for the future. The project has expended to create a second site in Esmeraldas Province, but this report facuses on the activities in Manabí. 15.1 Cassava in Ecuador Nationallevel statistics on cassava production in Ecuador vary considerably. Estimates of total area of cassava ranga from 26,000 to 40,000 ha ""'. Estimates far Manabl range from 5,500 to 14,000 ha. The remainder is grown in the Amazonian regían, the southern provínce of Loja, the humid coastal province of Esmeraldas, and in scattered areas of various provinces wíthin the humid lowland western region flanking the Ecuadorian Andean cordillera. Yíelds average 10 tfha. , .. , Romanolf, S. 1991. The boltom Une or how farmer·promoters reduced the roS! of organizing producers in an Inlegrated rural development project. Culture and Agricultu[i, Spring/Summer. 41:1-6. ,5.2 Ministerio de Agricultura (MAG), Depto. de Programación, Portoviejo, Manabl, Ecuador, 1990. Instituto Nacional de Estadistica y Censos. Encuesta de Superficie y Producción por Muestreo de Areas, Resultados de 1990, Tomo 1. 273 During the decade of the 1970s. there was close collaboration between CIAT and the National Institute for Agricultural Research (INIAP) on testing cassava varieties and agronomic practices. Most of this work focused on the humid western zone and operated from the INIAP experiment staUon at Pichilingue. In 1981. however, a study indicated that without the integration of cassava production, processing and marketing. there would be little incentive for farmers to adopt technology to increase productíon and less chance for developing alternative uses for cassava. 'M The report also recommended that cassava chipping technology could be a viable alternative in light of a rapidly growing livestock sector (swine and chickens) and insufficient energy sources. The conditions for increasing cassava production were not economically favorable, however. until 1985.'''' This set the stage for launohing the cassava integrated project in Manabí. 15.2 Principies of the Ecuador Integrated Cassava Project As the project has evolved, three guidíng principies have emerged, becoming the "culture" of project participants and the oritería for good collaboration among participants: • The transfer of technology--be it social or technical--is more rapid, effective and efficient when users are directly involved and responsible. • A tarmer organization is the most effective intermediary agent between the large no. of farmer participants and the institutions designed to serve development interests and objectives. The organization can be an efficient channel for project services, eredit and information dissemination; and it can serve as the most equitable manager of resources, postharvest processing, marketing and distribution of benefits. lS,' As sueh. the resources, activities and funds destined to strengthen the farmer organization should be channeled directly through the organization because the learning and experience of managing these will contribute to the growth. maturity. wisdom and ultimate sustainability of the farmer organizatíon. • There is a need to establish and maintain good interinstitutional communícatíon and collaboration. The farmer organizatlon--not merely a recipient of the benefits of the project but an active partieipant--should be part of the institutional group supporting the projeet. The idea is that as a result of the project, the farmer organization beeomes an equal partnsr in the instítutional setting, able to operate its services and activities in collaboration with other development institutions, but not dependent upon them. ,5.3 Cock, J.H. el al. 1981. La yuca en el Ecuador: Recomendaciones para el desarrollo y ejecución de un proyecto da producción, secamiento y comercialización. l~' CIAT. 1988. Cassava Program Anoual Rapart 1985. CIAT Working Oacumenl No. 38:7-11. ,.. Romanofl, 1991, op. cit. 274 15.3 The Cassava Integrated Project In Manabí Provlnce As the result of several visits to Ecuador in 1985, Manabl Province was identified as an appropriate place to test processing technology because it had cassava smallholder production in excess of current demand, marketing problems, a high small-farmer population where family members often migrate in search of off-farm income, and dry weather for solar drying. In addition, key individuals in research and extension institutions promised necessary support for the experimento , ... From the start, farmer processors have been key teachers, technology transfer agents, promoters and leaders of the project. CIAT's role has been to guide and foster this process by providing time, space and access for the farmer-to-farmer transfer to take place. Institutional development professionals including CIAT staff have become partners and collaborators with farmers in transfer and testing activities. The tirst farmer-to-farmer contacts were batween Colombian tarmer/processors and potential tarmer /processors in Manabl. The former taught the latter how to dry cassava; farmers from Manabl then visited Colombia to' sea the technology in action (the "technica'" technology) and to learo how to organize and manage the processing plants (lhe "social" technology). A farmer leader and a mason from Colombia worked in Nlanabl on the design and building of plants at the beginning of the project. Manabl fermers from BiJahual and Jaboncillo, who were members of preaxisting farmer organizations, carried out their own experimenta to see if the technology worked and could be protitable. When convinced by their own actions, they set about forming other groups of fermers to test the technology and join them in marketing the resulting producto Members ot the ínitial two groups were joined by others from subsequent groups, forming a committee to market dried cassava, with assistance from MAG-Guayaquíl and CIAT. The tirst tarmer groups became legal farmer assocíations (APPYs), and the early marketing committee grew into a unjon ot producer /processor associations, UAPPY. As the needs ot the APPYs has grown, the services provided by UAPPY have increasad. Today the tunctions and responsibilities of UAPPY inelude: .. administrating and managjng loans and donations .. providing processing credit to the APPYs "commercialization .. product milling, finishing and transformation .. product qualíty control .. accounting .. transportation , ... For a complete descrlpllon ot the early steps 10 lorm !armer groups and establish the interlnstitutional base 01 Ihe Ecuador Integrated Cassava Projecl. see CIAT, 1988, op cit., pp. 7-9. 275 I • training • forming new APPYs • communication among members • R&D of new products, processes, markets Many of these activities or aerviees could have been channeled through existing public institutions; however, lodging them within UAPPY and training UAPPY leaders to manage them has strengthened the organizatíon. The organizational structure of UAPPY as of October 1991 is depicted in Figure 15.1. The administrator, who plays the key management role, has a fifth-grade education, similar to most of the APPY presidents. The people occupying the positions under the administrator have a very different educational profile. The training coordinator has an Agr. Eng. degree from the U. Técnica de Manabí (UTM) and is a former president 01 his APPY. The office. manager, who is currently the president of her mixed APPY, is completing a degree in Economics at the UTM. The special projects manager is an APPY member and a mechanical engineer. Most UAPPY paratechnicians have completed high school. . . The strategic use of qualified and trained people from within the organization to carry out specialized tasks is one way in which UAPPY optimizes its heteroganeous human resourcas end malntains a cartaln independence from supporting institutions. This enables UAPPY to collaborata with development institutions, to negotiate on e more equel toOOng with othar instítutions, and to obtain the sarvices and support its members require. From the beginning, the MAG, INIAP and CIAT have collaborated closely ín the project. In 1987, when FUNDAGRO, en Ecuadorian private foundation for agricultural development, took over the support of cassava research, extension and eduction through a USAID-funded project, the FUNDAGRO cassava program coordinator was assígned to coordinate the interinstitutional committee. A key element is that UAPPY ís a member of the committee on equal footing with all other instítutional members, having an equal "voíee and vote" with other representatives in defining problems, desígnating priorities end assigning resources. The committee served initially as a device for communicating activities in order to stimulate participatíon and reduca overlapping efforts. In 1990 the committee took on the responsibility for evaluating project actívities internally and planning for the comíng year. With the second round of annual project planning by the committee under way, results to date indícate that research, extension and education activities are more in line with tarmer problems, there is a greater degree of interinstitutional collaboration within activities, and there is a better allocation of project financial resources to priority problems than there was when actívitíes were planned by individual institutions. 276 INTEAINSTITUTIONAL COMMITTEE·MAG, INIAP,ITM, UAPPY,ETC. FUNDAGRO COOrdlnator CIATA '! 11 ¡JI i il~íI~U id 1> L,~ 1 ti 11' N .. e~"!1: "¡J! I ~1S IJ pi .. 'fH ~ ; ) nnli h¡ ! Hl • ~ ~ ~~ 279 drying patio is required to reach efficiency and economy; moreover, APPYs with land to spare can consider expansions and alternative processing technology in the future. APPYs buy fresh cassava from their members and other producers. They chip the fresh roets, dry them, sack the chips and sell them to UAPPY. The UAPPY general assembly, composed of the presidents of the APPYs, sets the price for !he chips. Some APPYs produce under this price and thereby make an additional profit over that !hey receíve afler UAPPY markets !he chips. Others have higher costs of processing and lose money on chips, profiting only from the final UAPPY sale. UAPPY has experienced a strong pattern of product diversification that is unlike any of the organizations involved in other CIAT-linked integrated projects. In 1987, two of!he APPYs experimented with preparing, packaging and selling fresh treatad cassava for the fresh consumption export market Although this activity expanded the following year, it was discontinued for two reasons: First of all, it was very risky. UAPPY was responsible for!he product all the way to the shipping docks. While!he channels from the APPYs to the docks were handled smoothly, there were often delays once the material arrived at . !he docks. The buyers would reject material afler it stayed toe long at the docks. resulting in losses for UAPPY. Secondly, to be successful in fresh expart production. UAPPY needed a year-round source of good-quality roots. Manabi Province, with its distinct wet-dry seasons. is not an ideal elimate for producing quality roots year-round. With both marketing and production constraints, the activity was dropped. Unfortunately, no other group wi!hin Ecuador's more humid zones--where the maJority of the fresh eassava for human consumption is produced--has tried the technology. In 1987-88 UAPPY added two new APPYs composed solely of women'who produced starch rather than dried chips (see Sect. 15.5). This represented an important expanSion into!he market of quality starch for human consumption, which was sold to various stores and individual food proeessing users. In 1989 UAPPY suecessfully sold industrial-quatity starch to a large Guayaquil cardboard factory as a substitute for maize starch used to make glue for eorrugated boxes. A 1989 study,··7 estimated the demand for industrial- quality starch by this industry to be about 3,500 t/yr. Glven the potential demand and the less-demandlng processing for industrial-quality starch. there was a decrease in produetion of starch for human consumption. In fact the factory demandad more starch than UAPPY could produce. So as not to lose this market, UAPPY contracted to purchase starch from prívate starch producers in the area. While !his move was strategicafly correct from a marketing perspective, it placad UAPPY in the pasitlon of being a middleman--a role that it was not intended to have. The shrimp faed industry. unlike other lívestock feed industries, demanded cassava flour, not chips; thus UAPPY had to develop mifling capacity and management as wall. The , .. , CENOES. 1989. la industrialización de la yuca. 280 UAPPY, in close collaboration with a local NGO metalworking training and research facility - FACE - developed portable hammer milis to grind cassava chips into flour. An APPY member to be in charge of the mili operation and maintenance. This sparked the idea of developing a UAPPY Oemonstration Center, where new technology could be designed and tested, and UAPPY members eould learn and experiment with the teehnology through training and demonstration events. With PL-48Q funds, land was purchased 17 km from Portoviejo (UAPPY HQ), and a structure for housing the processing equipment was begun in 1987. Although PL-48Q decided not to finance the termination of the structure, it was eventually completed with other USAIO monies through the FUNOAGRO-supported cassava programo UAPPY shifted the milling task from the APPYs to the Center and also used the structure as a temporary warehouse and depot to assemble treight loads for Guayaquil, Quito and elsewhere. The Center also has a drying patio and standard chipping and starch-making equipment. Although its productíon is not large, the Center does produce regular quantities to offset sorne 01 its operating costs. In 1988 several of the shrimp feed factories complained that the cassava flour made 01 whole unpeeled roots contained too mueh ash (due to the peel) and requested a peeled producto UAPPY complied, at a higher price far the peeted product, and began selling both whole and white (peeled root) industrial cassava flour, depending on the clients' needs. Initially, APPY members resisted peeling not only because 01 the additionallabor involved but also baceuse many people had no skill in peeling outside of the traditional areas where cassava starch is produced trom peeled roots. This activity soon baceme a valuable source of additional ¡ncome for both member and nonmember families who could earn as much as a month's mínimum wage in two weeks time peeling cassava in the late afternoons and evenings at the local APPY. In the 1990-91 processing year, UAPPY generated more than 8/.15,000,000 (US$16,OOO) in wages paid to peel cassava. Mast of these wages were paid to poor women, children and the elderly who had no other souree of income during this period of the year. The importance of the income generated for poor households has made it diffieult to consider introducing machinery to mechanize peeling. Oespite these early mCives toward diversification, the bulk 01 the UAPPY product was still destined to one single market--the shrimp industry--untiI1989, when two things happened to change this. First of all, 1988-89 had been a very profitable year so UAPPY and APPYs had money available to expand their facilities in arder to increase productíon significantly. With an assured promise of a large new secure revolving fund for operatíng credit trom PL-480, both UAPPY and the APPYs invested their profits in expanding their drying capacity. As the drying season approached, however. the funds were not released despite constant pressure from UAPPY and the FUNOAGRO coordinator. Wlth no other recourse available to them, sorne APPYs borrowed money from private lenders at rates up lo 12O%jyear; others bought and processed cassava on credit trom members; the rest simply could not process. Then, UAPPY received another blow: the shrimp industry slumped. Competition trom Asian producers and a shortage of larvae to stock ponds causad a cutback in production and a simultaneous halt to cassava purchases. UAPPY's 281 response was to tighten their administrative belt by releasing halt the staff. halving the salaries of the remaining staff, canceling all nonessential training courses and social gatherings, and putting the UAPPY vehicle in a garage. They then launched an all-out campaign to sen cassava flour to other industries. When the PL-480 funds were finally released at the clase of the procsssing saascn, they were used to pay back loans. UAPPY managed to find new buyers and the shrimp industry came alive again; by April 1990 all the stored production had besn sold. The economie balance far the year was poor, but a valuable lesson had been learned."'" UAPPY, in coordination with its committee of supporting institutions, set as its first priority, diversification of the markets lor its existing products and expansion of its produetion base to ¡nelude new products. Two examples of this poliey are already in place. In 1990 UAPPY began refining the whole-root cassava flour by passing it through a mechanical vibrating sifter (developed again in collaboration with FACE) to yield a flour of the sama granular size as wheat. This was sold to a factory that uses it as the filler for the resins usecl to maka plywood. Currently the three largest plywood factories purchase refinecl whole cassava flour from UAPPY as a substitute for 20-42% of the wheat. Bran, the by-product from sifting, is sold as a scurca of fiber to the livestock feed industries in the highlands. UAPPY uses the same mechanical sifter to produce a white flour for human consumption (noodles). While the total amount of flour scld for human consumption is small, UAPPY is currentiy testing better processing methods in arder to expand production. Table 15.2 summarizes the types of markets UAPPY is currently dealing with. the products sold to each market, and tha changes in the amoun15 sold to these markets between 1989-91. UAPPY has decreased i15 dependency on the shrimp market by 30%. Table 15.3 summarizes the total volume of production by produet and the total sales for 1990-91. UAPPY total sales for this year were US$352.61O. Figure 15.2 shows the avg income earned by UAPPY members and nonmembers over this same periodo Income to members is the sum of fresh eassava sales to the APPY, wages earned working at the APPY, money eamed from peeling, and mambar shares of profits from sales. For nonmambars, income is mostiy from sales of fresh cassava and paeling; however, a few APPYs employ soma nonmembers as wage laborers. At one APPY, Miguelillo, all manual labor in processing is done by hirad nonmember labor. Income dacreased in 1989-90 due to the aforemantioned problems. The further decline for 1990-91 is due to increasing costs, inflation and failure of prices tor cassava products ",. Although UAPPY vowad thallhls would never happen agaln, a very similar situatlon has developad this year with a promisad granl Irom the Central Bank Ihrough FODERUMA lo be usad lor operatlng cradit and construction. Because of bureaucratlc delays and demands, Ihe money wHl likely arrive in March 1992, 8 mo late and lOo lale lo be usad In this processing season. 282 Table 1/U. Market sil ....... 01 annual """",H 1000 mm annually; the other, 500-1000 mm. Ten APPYs are within the drier zone; 5 are well within the wetter zone; and 3 (Tablones, El 286 Algodon and Las Piedras) are located on the transition line, but as their area tends to be well watered, they are grouped in the higher rainfall zone. Those APPYs in the wetter zone have a longer rainy season and fewer total no. of sunny days and thus a shorter drying periodo But because of the rain, they have more abundant cassava, better quality fresh material and lower priees so their production costs are lower. In the dríer zone, the APPYs have the conditions for processing (sunlight), but no cassava Is avaitable for lack of rainfall. Because of greater competitíon for fresh cassava given their proximity to many private starch processors, production costs are higher. APPYs also differ in their empresarial nature. A few víew themselves as a business venture, and the bottom line is max. processing for profits. Others are more cooperative in style, operating more for the welfare of their members and demanding a level of individual contribution in terms of kind or labor. One of the suppositions underlying the concept of the integrated cassava projects was that the processing aetivity would encourage members to plant more casaava. In, Manabl, however, thia does not seem to be the case. More cassava in the region Is being processed, but the produetion increases at each plant are not a result of increasing amounts of roots aupplied by members. Most members have very little land and canoat increase the area under cassava and still produce other foodstuffs for on-farm consumption and the market. As processing plants increase in number and slze and demand more fresh caasava, membera cannot respond, so an increasing arnount of the fresh cassava processed is sold to the plant by nonmembers. For APPYs in the dríer 3Dfle, purchasing cassava from wetter regions ia a viable atrategy, rather than a negative illdicator. Examined from this perspective, the impaet of UAPPY in central Manabl has been the creatían of 18 new local purchasing sites for cassava, Where vírtually no market existed. The beneficiaries of the project are far more than UAPPY members. UAPPY's role in the future will be to concentrate on identifying new markets, new products and aven new crops to process. Staying ahead of the market and being responsive 10 shifts in demands will be the key to long-term víability for UAPPY and the cassava prograrn as a whole. Training and research in the project today refleet the recognition of these needs. 15.6 Tralnlng Wrthin the project training is carried out at several levels: international courses and seminars, training courses within participating institutions for staft and members, and courses·given by ane institution for another. Rather than listing all of the training avents at each level, sorne examples from each will demonstrate how training is organizad and how it benefits the projeet. In the past, participatian in international training courses and seminars for researchers was primarily by INIAP rasearchers though on two early occasions, MAG o#icials 287 participated in CIAT cassava activities. In 1990, when the interinstitutional committee of the cassava project took on responsibílity for planning project activíties, it recommended that tunds for training of researchers be extended to university taculty and students conducting thesis research within the program, and to members of UAPPY. This decísion recognized that INIAP was not and could not be the only entity conducting research in support of the project. University students conduct thesis research with UAPPY. UAPPY members regularly conduct adaptive research on processing and production technology for cassava as well as other crops in the production system. International training is now used as a mechanism tor promoting integration among these institutions and to develop new research activities. Two examples demonstrate this change. In September 1990, the project sent an INIAP agronomist, a UTM thesis student and the UAPPY training coordinator to the CIAT cassava seed course. Upon their return, they designed a cassava seed multiplication project that was funded and implemented in 1991. In February 1991 the UAPPY special projects manager and 4 other members were sent to CIAT tor a month to receive individualizad training in new processing technology for starch and flour. Three of the tour are now working with the special projects manager as UAPPY paratechnicians, testing new processing technology under Manabi and UAPPY conditions. Thefourth person is working as a promotor for a new women's APPY. UAPPY has always played a key role in training its members. Training courses on a variety of technical, organizational and management tapies have been held at UAPPY HO with appropriate APPY representatives attending. The assumption was that the representative participating in the course al HO would inform/train those members who were not participants. Last year:APPY presidents complained that they were nO! able to communicate the "Iearnings" from the training courses to their members because they lacked training skills themselves. They requested either to have training in how to train or to carry the training program out to the larger UAPPY membership. The training coordinator sent out a questionnaire to all APPYs to find out whether they wanted their own training courses and what kind of training was desired. They wanted training courses at the APPYs in addition to specific technical training courses normally organized centrally. This demand shaped the planning for the 1991 training programo Additiona/ funds were solicited from FUNDAGRO to expand the number of courses given. Four themes were selected for the year. The director identified MAG extension staff as trainers fer the courses, and together they programmed 4 short courses for each of the 18 APPYs over the year. UAPPY solicited support from the MAG's World Bank-supported extension project, PROTECA, for vehicles to transport the trainers to the courses. Although PROTECA promised support, they rarely carne through; and UAPPY usually supplied the transportation. UAPPY controlled all funds for the training program, including the per diems for the MAG trainers. If the trainers canceled courses or cut back on the no. of days, then UAPPY reduced the per diem accordingly. The UAPPY training program has been quite successful thus faro APPYs have kept attendance high, often with self-imposed fines for missing cOurses. To date, 80% of the 288 program is complete. FUNDAGRO funds have covered 76% of the training costs and UAPPY / APPY funds (a pereent of sales goe5 into an education fund) have covered the resto Evaluations show that 30% of the UAPPY members are now; capable of leading and managing their APPY5, and 45% of the APPYs are capable of functioning alone without direct UAPPY supervision. A mixed training plan of APPY and UAPPY-Ievel eourses will likely be implemented in 1992. Wlth respect to institutional training, INIAP normally runs a cassava production course each year far extension agents and participants from other institutions. The INIAP Tropical Roots and Tubers. Program also holds field days and tarmer demonstrations or competitions for UAPPY cassava farmers in conjunction with its on-farm variety and production trials. This year, due to scheduling difficulties, INIAP decided to pastpone its production course. The UTM had wanted to inelude faeulty members in the course and appealed to the FUNDAGRO eoordinator far help in running the course at the UTM. The coordinator agread to the course but propased that UA~PY organize it, and the UTM agreed. With leadership provided by the UAPPY seed project agronomist and collaboration from the sead project paratechnician, a UTM student working on artifieial drying technology. the UAPPYadminístrator, project manager and selected farmers, the course was given to a group of 16 UTM faculty. The outcome was very positive and may set the stage for future training by UAPPY for others. 15.7 Integrated Cassava Project Research: 1990-91 Hlghllghts As a result of the interinstitutional committee's recognition that research can and should be done by institutions in addition to INIAP. FUNDAGRO approved tor the 1991 cassava program work plan several research activities under UAPPY management. Although most of thase deal with processing problems, research in other areas sueh as seed multiplication and marketing is also managad by UAPPY. Sorne of the advantages of conducting researeh within the UAPPY vs. a ministerial context is greater flexibility te shift directions in response to preliminary results and a less bureaucratie financial system, both of which are crucial to adaptive research aetivities. Research on cassava in Ecuador can be divided into production and postharvest categoríes. The majority of the production researeh is conducted by INIAP at its experiment stations in Portoviejo and Santo Domingo, and in farmers' fields in Manabl and Esmeraldas. Postharvest researeh 15 concentrated in Manabl. 15.7.1 Productlon Research 15.7.1.1 Varíeta! testing. Since the early 19705, INIAP has been testing cassava materials from CIAT and others collected nationally. A collection of the national materials is maintained at Portoviejo, but a complete characterization and elimination of duplicates remains to be done. Evaluation of local cassava varieties in Esmeraldas and Santo Domingo is ongoing. INIAP will release a new variety in 1992 called ·Portoviejo 650," 289 ¡ f which is M Col 2215, widely grown in Venezuela and Colombia. After its initial introduction to Ecuador in 1987, it was submitted to several years of multilocational testing before release. Although it dces not always outperform other local varieties in yield in all ecological conditions, in drier regions it yields better than others and has a higher DM content. A recent harvesting and processing test at 8 mo showed a DM of nearly 40% and processing conversion rates at nearly 2.5: 1. This feature alone makes it quite desirable for UAPPY farmers. UAPPY's urgent demand fer new varieties with higher DM content, improved drought tolerance and earliness has stimulated the cassava project to support UAPPY in developíng a seeo multiplication activity. 15.7.1.2 UAPPY Seed Multiplication Project. A recently graduatad Agr. Eng. who had done his thesis research on rapid multiplicatíon techniques with UAPPY is in charge of the seed multiplication proJect. He is backstopped byan INIAP agronomist and assisted by a UAPPY paratechnician who attended a 6-wk small farmar seed course at CIAT. The objective is to create a seed multiplication capacity within the UAPPY to ensure the supply of good-quality planting material to UAPPY members and to earn money from the sale of materials to nonmembers. At present UAPPY has 3 seed multiplication lots (3500 plants) for Portovíejo 650: two in Bíjahual and one in Jaboncillo. Wlth these, UAPPY hopas to have 75,000 plants at the end of the year ín order to plant 8-10 ha of the naw variety by mid-1992. Material trom these will be usad to plant about 90 ha of the naw variety ín 1993. 15.7.1.3 Agronomy and production systems. Most of the agronomy research has been directed toward spacing, planting density, stake treatment and storage, and preemergent weed control. Nearly all of the station and on-farm agronomy trials have been done in monocultur!3, des pite the fact that more than 70% of the cassava in Manabl is grQwn in mixed cropping, primarily in association with maize. Gassava is also grown in association with at least 10 other crops, including peanuts, cotton and cowpeas. Work by INIAP soíl scientists has examined fertilizer-soi! relationships under maize-cassava systems; an agronomy thesis examined yields under different spaciog arrangements in maize-cassa.va production systems. Little work has been done on other intercrops with cassava nor has explicit attention been givan to the impact of cassava cultivation on hillsides where the majority of the Manabi cassava is grown. . Duríng the past.and present production seasons, UAPPY leaders have argued for IN!AP production research to be directed to mlxed cropping. hillside systems and agronomíc measures to conserve soil moísture. Some UAPPY members are conducting their own fíeld tríals on several of these concems. UAPPY is also requestíng recommendatíons on the management of the new varo PortovieJo 650 under the prevailíng mixed crop system as the architecture of the new variety (short and bushy) is quite different trom the traditional varieties (tall with little branching). UAPPY pressuré during planning meetings last year resulted in INIAP agronomy maize-cassava tríals this year with the new variety. In addition. an MS degree student in agronomy from the U. of Florida plans to conduct maize-cassava tríals during 1992 in collaboration with INIAP and the UAPPY. A UTM 290 student will conduct theses research on the response of Portoviejo 650 to K fertilizer, and another UTM student (son of an APPY member) intends to conduct thesis research on eassava-peanut systems under irrigation. 15.7.1.4 Pests and diseases. Conventional wisdom eoncerning cassava cultivation in Manabi maintains that cassava suffers little from pest problems and that aside trom preemergenee herbicides, few ehemicals are used on the crop. However, changing cultivation patterns, increasing cultivation in dry zones, and new varieties may change this. INIAP is conducting research this year on biological control of mites and cultural controls for Phyllophaga. A multidisciplinary survey of cassava pests and diseases is being conducted in all cassava-growing regions 01 Ecuador to provide the diagnostic information to guide future research in this area. 15.7.2 Postharvest research 15.7.2.1 Fresh cassava handling. Uttle attention has besn given to this area. An economics thesis"" on the marketing of fresh cassava to Guayaquil conciuded that there was a 20% loss due to detérioration during commercialization. Despite this, cassava is considered as an indispensable foodstuff far 81% of the lower income population surveyed. Annual per capita consumption of cassava by this group is 34 kg, and frequency of consumption is 2.9 times/wk. 111e majority of the fresh cassava comes to Guayaquil from La Mana area, near Quevedo. 111e study recommended that INIAP initiata production researeh in the zone as it is the most important fresh cassava production area. INIAP has not yet followed up on this recommendation; however, it is conducting trials this year on the use of perforated plastic bags and chemicals for preserving fresh cassava far marketing. 15.7.2.2 Uvestock feed. Given the potential importance of dried cassava as a livestock feed, it is surprising that so Httle research has been done on this topie. In 1990 INIAP proposed tour experiments with dairy cattle and swine to determine the feasibility of feeding cassava silage or dried cassava in mixed rations, but none of the trials was implemented. Some preliminary work has been done on cassava feed for highland dairy cattle at the INIAP Sta. Gatalina station. 111e U. Politécnica in Guayaquil was contracted by FUNDAGRO in 1989 to study the use and quality of shrimp leed pellets made with varying percentages 01 cassava flour. Resuits were inconclusive as apparentiy some 01 the data were mixed up and resulting values were not at all probable. Sorne of the balanced feed factoríes that use cassava flour in their shrimp feed pellets claim to have conducted research on the quaUty of pellets mada with cassava; however, these results are not readily available. "'" MosqUOO! larrea, Anlbal H. 1989. Caracterlstlcas de la comercialización de yuca fresca en la ciudad de Guayaquil. Economlcs thesls, U. Central del Ecuador, Facultad de CIencias Económicas, Escuela de Economra. 291 15.7.2.3 Demonstration canter. The UAPPY demonstration center is a prime site for conducting postharvest and procassing research. Given Ihe absence of any other institution with procassing research capability on cassava, the center should playa key role in tha research program for the futura. It is also an ideallocation for linking farmers and professional rasearchers in collaborative problem-solving research. Current research activities include Ihe evaluation of an artificial dryer for cassava chips at differant temp and capacitias; tha testing and modification of starch processing equipment introduced from Colombia; the deslgn of an artifiCial starch dryer; construction and testing of a centrifugal flour siftar; and constructíon of a stainless steel cassava chipper with a receptacle for producing quality flour for human consumption. 15.7.2.4 Flour processing. In addition to the work at Ihe demonstration canter, INIAP is currently conducting an experimant to determine the teehnical, economic and nutritional feasibility of produclng flour using Ihe grater commonly used to grind cassava roots to extract starch. Traditional starch processors often use Iha grater in this way when prices for starch are too low or when thara is intermittent rainfall and starch drying is not possible. Positive results could reduca the cost of processing flour and provide alternative markets for traditional producars. UAPPY is running a pilot project in Ihe APPY Siete de Mayo to test improvaments in the procassing of flour for human consumption. The new cassava chipper will be tested at the APPY, along with diffarant management techníques designad to improve quality and raduce contamination. 15.7.2.5 Starch procassing. Specific attention to starch processing increasad in 1990 as a result of the demand for industrial cassava starch from Ihe cardboard box industries. A CEEMAT·CIAT starch specialist provided critical technical assistanca to UAPPY and made recommendations tor areas in traditional starch-making technology that needet;! rasearch or modifications. A survey of nearly the entire universe of traditional cassava starch producars (208) was completed in 1990. It was found that two processing technologies exist: one is completely manual; the other, semimechanized. Producers procass from 50 kg to 2 t of starch/wk depending on the size of the plant, no. and size of sedimentation tanks, and drying space. About 3,500 t of starch, including both industrial and food consumption qualities, are produced yearly. The procassing activities generate a total of 2500 daily jobs during the approx. 6 mo of annual processing. There is no technical assistance directad toward this industry so the majority of Ihe processors rely on their own resources to finance their operations, informal sector cradit, or financing by intermediaries. The survey also revealed serious problems with contamination of starch that is intended for human consumption, adulteration by intermediaries of industrial quality starch, and a total lack of technology for handling waste water from the plants. The survey data are 292 being used by two UTM economics students doing a thesis proJect with UAPPY on the economics of starch production in Ecuador. As mentioned earller, UAPPY has built a pilot starch processing plant on the new site for the San Vicente women's APPY with funding from FUNDAGRO. The plant is nearly complete and tria! processing runs have produced starch products superior to what is currently produced by prívate processors or UAPPY. The plant will enter full production in míd-Nov. A UAPPY paratechnician is training the women to operate the plant and is monitoring the assessment of its operational and economic efficiency. The design of the plant retains the manual labor of the women members, but has made the labor more efficient and less tiresorne. Rather than liftíng buckets of water to wash the cessava mass, a plumbing system delivers it directly to each washer. A water filter removes sand from the water supply. A pre-washing tank cleans the roots more completely 'and with less effort. "the plant is successful, UAPPY plans to modify other APPY starch plants and experiment witlJ lower cost construction materials. In addítion to substantial market represented by the cardboard box industry, thera is also a strong interest in starch froro Latinreco, 'tha research division far Nastla in Latin America, which operates near Quito. Latinreco researchers are collaborating with UAPPY in characterizing the quality of the UAPPY products and for the past 6 mo have been running trials in their pilot processing plant using UAPPY starch. Results thus far have been quite good and they plan to increase their experimentation in hopes of creating commercial products using UAPPY starch. Entry into the Nestle market would open other important doors to commercial faod processors and expand the demand for UAPPY starch. 15.7.2.6 Marketing. The search for new markets for UAPPY products and understanding the demand characteristics for potential new markets has been a priority for UAPPY since 1989; however, the UAPPY commercialization committees have pursued new markets in a rathar ad hoc fashion. While they have achieved success in some areas, a more organized approach to marketing will likely increase the chance of successful market antry. Based on a market study of cassava food consumption flour in Colombia, the cassava project embarked on the design of a similar study for Ecuador. The study invoives a broad range of institutions and has brought new ones into more direct involvement with the project, such as Latinreco and the U. Politécnica in Quito. Different < from the Colombia study, the survey will examine the potential markets tor 9 UAPPY flour, starch and by-products, cutting across all potential industrial and cottage-industry sectors. An inventory of all potential users has been made, and a selection for the survey in 5 locations is complete. Due to delays, the first survey will be ¡nitiated in 1992. In addition to the new products survey, a Dutch MS student--in collaboration with FUNDAGRO, CIAT and UAPPY--is doing his thesis research on the macroeconomic issues and future for cessava products, particularly the market potential offered by the plywood and resins industries. 293 15.7.2.7 Characterization 01 UAPPY products and by-products. In conjunction with the new markets survey, the cassava project is characterizing all the flours, starches and by- products of the UAPPY according to variety used and date of harvest. Samples of 3 varieties (2 traditional and M Col 2215) at 4 harvest dates (8, 10, 12 and 18 mol are being analyzed for their nutritional, functional and microbiological quality. Results will be used to provide information to potential users and to identify needs for Improvements in quality. Analyses are being done at the U. Politécnica of Quito, Latinreco and CIAT. UAPPYalso had the help of a food technology student from L'Ecole Superieure O'Agronomie Tropicale, Montpellíer, France In organizing and definíng the methodology for the study, as well as handling the first se! of analyses. 15.8 Future Oirections It seems clear that the future for UAPPY wHl be directed to further expanding its markets, Improving the quality of its products, and exploring other agroindustrial avenues. The results of UAPPY's efforts to understand and define the parameters of the national market for cassava products will also determine whether the program should expand further into new areas and create new UAPPYs. Production research for the future should concentrate on varieties suitable for dry zones, mixed cropping on hillsldes, with high OM suitable for processing. UAPPY should enhance the demand for higher OM varietles by implementing variable pricing for fresh cassava according to OM content al the APPYs. Research should also direct attention to the longer range sustainability of cassava cultivation on dry, fragile hillsides and should begin to study technologies that will conserve the natural resources that support cassava production. Research should also move into other cassava-growing areas. Oiagnostic surveys to determine production problems and opportunities should be conducted in the other zones, especially in Esmeraldas. Processing research should focus on product quality, method~ of quality'control and market demando In addition, there is a critical need to address the contamination produced by waste water from starch processing and design systems to elimínate waste before the water returns to the streams and river systems. UAPPY should consider otlier kinds of processing technology compatible with their existing organizational structure and physical ínfrastructure. UAPPY is already considering a proposal to develop a maize starch extraction plant within the demonstration center as a way of industrializing the other major cash crop of the UAPPY farmers and take advantage of the drying space when cassava is not being processed. The UAPPY experience has validated the three principies that guided the project from the start. It is evident that the project is now fully embraced by the participants, who becoming more and more in control of the decisions that guide the activities of the programo It is likely that FUNOAGRO funding will decrease as the initial project goals 294 have been mel The challenge for UAPPY and its collaborating institutions will be to carry the effort into its fully commercial and diversifiad future. Huffstutlar''''o warned that the third wave of farmers would be different from those of the first and second wavas. Tha faet that larger, industrial interests are now inquiring about how thay might "ereata a UAPPY" to supply them with cassava products for various industrial purposas should not than be surprising. The next staga of the cassava projeet in Ecuador should be well worth watching and deserving of continued CIAT involvement. H\.tO , Huflstutlar, S. 1989. UAPPY evaluatlon and recommendatlons. Costa Rica: Agricultura! Cooperative Development International. 295 16. BRAZIL Brazil is the world's second largest producer 01 cassava (nearly 16% 01 the world total) and the largest cassava producer in Latin America, accounting for nearly 75% of the tata! production in the region.'8.' The CIAT Cassava Program's relationships with national research and extension services in Brazil have been channeled through the Brazilian Agricultural Research Institute (EMBRAPA), the Brazilian Institute of Technical Assistance and Rural Extension (EMBRATER), and the National Center tar Research on Gassava and Frults (CNPMF). CNPMF; in turn, implements Hs working strategy through the National Research Program for Cassava (PNP-Mandioca), which provides financia! and technical support to research and extension agencies at the state level. This fruitful collaboration has resulted in a number of projects. COllaboration at the disciplinary level (breeding, pathology, entomology, utilization, etc.) has been reportad elsewhere: this chapter facuses on the significant activHies that have bean undertakan within the framework of the Caará Integratad Gassava Development Project. 16.1 Actlvltles Leadlng up to ProJed Executlon 16.1.1 1987 Responding to a request from CNPMF, the CIAT Cassava Program organizad a workshop in Novamber 1987 for a group of 21 researchers and extension personnel from various states of NE Brazil, including six members of the CNPMF cassava team. The objectives of this workshop were threefold: ,. Update the knowledge of the participants in relation to advances made by the CIAT Cassava Program in the last years. ,. Evaluate their current activities in the light of this new knowledge . .. Exercise a criticar analysis of the national research program in cassava (PNp· Mandioca) with the aim of integrating research and extension aetivities and the overall deveJopment of the cassava crop in the region. The participants made a diagnosis of the situation of cassava in Brazil, particularly in the NE. The potential for establishing Cassava Integrated Development Projects was 18., FAO. Perspectivas alimentarias. Marzo 1990. 297 · , identified as ene of the principal alternatives fer overcoming the limitations faced for expanding the crop in the region. The workshop also included, tor the tirst time in the history of Cassava Program training events, a group 015 policymakers from key institutions in the region. This group attended the final part of the meeting, participating actively in the evaluation of the diagnosis and the discussion about future work plans. 16.1.2 1988 CNPMF underwent serious administrative problems, and its participation was practically frozen. Given the excellent results obtained in the 1987 workshop, CIAT was abre to cooperate directly with state agencies across Brazil, aspacially in the NE. . The state of Ceará with 6 participants in the workshop (2. policymakers and 4 technicians) formad the Ceará Cassava Committee (CCC), responsible for coordinating all activities related to the crop in the state. The CCC collaborated closely with CIAT in developing a project proposal submitted to the W.K. Keilogg Foundation (KF), requesting financial support for implementing an Integrated Cassava Development Project in the state. One of the CCC's main achievements during 1988 was the installation of 8 cassava drying plants with funds from various sources, among which the CIAT Cassava Program collaborated with US$ 20,000. The Program also cooperated with state agencies in carrying out cassava training events in Porto Alegre, Río Grande do Su! (EMATER-RS). Fortaleza, Ceará (EMATERCE), and Recite, Pernambuco (lPA). 16.1.3 1989 Based on the positiva results of Its tirst training axperíence with policymakers, the Cassava Program decíded to organize a study tour for a group of high-Ievel authoríties from 4 states of NE Brazil (Ceará, Pernambuco, Paraiba and Bahia), including the Secretary 01 Agriculture and the Presidents of the agricultural research and rural extension agencies. Representatives from EMBRAPA, funding agencies at national and regional levels, and farmer organizations were also invited. These 4 states plant 612,637 ha of cassava (1987), accounting for 36% of the total area planted in Brazil and 51% ofthe total area planted in the NE. '8" The study-tour included visits to ongoing Integrated Cassava Development Projects in Ecuador and Colombia, as well as to CIAT Ha. The foundations were laid 10r subsequent and successful developments in the history of the Cassava Program's collaboration wíth the overall development of the crop at regional and national levels in Brazil. '6" Fundacáo Instituto 8rasileiro de Geografla e Estatlstica (IBGE). Anuario Estatistico do SrasR, 1989. 298 In May 1989, after considerable planning and negotiations, the proposal for the Ceará Integrated Cassava Development Project was finally approved by the KF, and the 3-yr project got under way. 16.2 The Ceará Integrated Cassava Development ProJect The process of developing alternative markets for cassava producers raquires key institutional interventions aimed at overcoming the inherent limitations caused by the lack of diversification in cassava markets. The CIAT Cassava Program has besn involved in implementing this type of institutional interventions (Integrated Cassava Development Projects) in key target cassava-producing areas of Latin Amenea during the last 10 yr. The Ceará Integrated Cassava Development Project is aimed at establishing the production of dry cassava chips for animal feed as a viable agroindustry among small farmers across the main cassava-production areas of the state. 16.3 Progress Toward Outcomes 16.3.1 Development of the pllot project Project implementation has been strong/y influenced and benefrtted by prior actívities earried out by agricultural researeh and extension agencias in the state in relation to small-scale cassava farming and processing. The experience accumulatad ovar these past years has served as a foundation far the project and its organizational infrastructura. At the onset of the project, there were already 12 cassava drying plants organized (1986- 89); but they were not functioning successfully. These groups were reactivated and another 47 farmer groups established, for total of 59 dry cassava agroindustrles operating in 1991. Tha selection process for new groups includes a formal request by tha community. Then the Regional cassava Committees (RCCs) evaluate the new group's potential for dry cassava processing activities. Baseline data on rural communitias with potential to join the project are maintained at the centrallevel; and depending upon the availability of funding, the processing facilities are thén established. During 1989 a total of 11 new groups ware organizad; in 1990, ·16; and in 1991 another 16 new drying plants were established (Table 16.1). 16.3.2 Identlflcation oflocallnstltutional support and financial resources The incipient state-Ievel CCC, created in 1988, gained general recognition as the coordlnating body for all the activities related to the promotion and development of the cassava crop in the stata of caará. During these 2 yr, the cec has sought to decentraliza all project actívitias through the formatlon of the RCCs in the principal areas of project Influence. To date, the 5 RCCs established have besn accepted by the two main counterpart agencies--EMATERCE and EPACE-as part 01 the prevalling institutional 299 , Tabl. 16.1. Dry caosava agroindustrie. ""tabllshed in Ceará, 1986-91. Region 8$-87 88-89 89-90 90-91 91-92 Total Financial $ou,cos, Ubajara 2 2 7 1 (2);2(3): 3(2); Sobra! 2 3 3 4 13 1(5);2(5); 3(3);5(1); ~apipooa 3 2 5 6 17 1 (8);2(7); 4(1);6(1); ca.;ri 2 3 2 8 1(4);2(1); 3(1);7(2); Umoeiro 1 3 4 1(1);2(3); Crateus 2 2(2); Fortaleza 3 3 1 (3); RIJosas 4 4 2(1);7(3); Satuti1e 1(1); Total 4 8 11 18 18 59 1(24);2(22); 3(6);4{5); 5(2) , 1 • IBRD; 2 • PAPP; 3 = !!NB; 4 • SIC; 5 -. OIhor. landscape, thereby facilitating the performance of project personnel in their activíties as advisors to the farmer groups. Linkages between the cee and the Secretariat of Agriculture (SEARA) and Land Reform have been strengthened significantJy during this periad; and it is expected that by the end of the project, the cee will be functioning as a technical body annexed to SEARA. Some 01 the programs in which the cee is receiving direct support from SEARA are as follows: • Funding (US$ 15,000) of a special project--conductOO jointly by eNPMF, elAT and eee--aimed at controlJing asevere outbreak of wítches-broom disease. • Funding (US$2Q,OOQ) of specific raquests from farmer groups to install dry cassava agroindustríes in several cassava farming communities. • A project (US$50,OOQ) presented to the NE Bank of Brazil (BNB), has been approved; and in coordination with SEARA, 50 dry cassava agroindustríes wíll receive financlng to purchase oxen-pulled wagons for transporting cassava roots to the drying plants. This crOOIt lína is of special significance as this is the first time that a loan-type credit program basad on price variations of dry cassava chips has been acceptOO by the funding agency. The functioning of the credit program wíll be carefully monítored and 300 evaluated as it could beco me the basis for formulating future credit programs for investment, operatíon and production. • EMATERCE and EPACE created a revolving fund (US$10,OOO/yr) out of the resourees allocated to them by the projeet, whieh during the first 2 yr has be en used as working capital for the dry eassava agroindustries: 23 farmer groups in 1990 and 49 in 1991. It is expected that by the end of the project, the administration of this revolving fund will be directly assumed by the farmer groups through some form of second-order organization representing all the cassava farmers partieipating in the project. • The insta/lation of dry cassava agroindustries is being carried out with active participation of the main local executing agency, EMATERCE, which maintains direct relations with local, national and international programs that are potential sources of financial support. Thus far the project has been functioning within its goaIs using these type of grant-type financial resources (Table 16.2). 16.3.3 Produdion technology One of the more important components within the strategy of Cassava Integrated Development Projects is production technology. The Ceará Project uses a methodology that ¡neludes two components: pre-production trials and seed production plots. • pre-production trials. Aimed at increasing farmer adoption of improved cassava production technology, these trials are being installed in the vlcinity of the drying plants to allow maximum participation of farmer. groups. The planning, establishment and evaluation of these trials involve active participation of technieians frorn the two main counterparting agencies, EPACE and EMATERCE. Table 16.2. Eslímaltld financia! support 01 Btazilian development programslO 1M Ceatá ca ... "" Project. 19f16.91. No. Farroer Tofa! Value Sources 01 Funding , Períod Groups US$ 1%) 1986-87 4 16.025 1 l_ a 24,425 ';2;3;5 1_90 11 46,715 1;2:3 199Q.91 18 158,082 1¡2;4 1991·91 18 101,801 1¡2:5;4 Tota! 59 347,046 1 (36,8%);2(49.3%) 3(6.9%);4(5.5%); 5(1.5%) 1 1 # IBRO, 2 = PAPP, 3 = BNB, 4 = SIC, 5 = OOer •. 301 In 1990 a total of 15 pre-productíon plots (PPPs) were installed and harvested after 15 mo. Results show that productivity levels in the state can be increased significantly through improved production technology. Avg RY was 22.4 t/ha, 59% more than avg RYs obtained in farmers' plots (Table 16.3). In 1991 the no. of PPP was increased to 44, covering the main areas of project influence. • Seed production plots. loitial results have not been as promising as expected because the larmers have placed more importance on the seed plots as a source of roots for the drying plants than as a source of planting material, preferring to get low-quality seed from other sources and leaving the seed plots unpruned at planting time. Additional factors affecting the success 01 this actívity have been the farmers' lack 01 experience in the practice of pruning plants to obtain planting material and the fact that the period between planting and harvestirig represents a lag period 01 at at least 6 mo, making the storing 01 stakes for planting material unfeasible. Most 01 the 15 cassava seed plots planted in 1990 were harvested at 15 mo, thus becoming another source 01 benefits for the farmers as the roots obtained from these plots were donated to the drylng plants. In 1991 the no. 01 cassava seed plots ¡nstalled was increased to 41, covering the principal cassava-producing areas in the state. T_ 16.3. Avg yleldo of pppl In C..rá, 199Q.91. Regloo Commuoity Catlrl Se"a Santana ltapipoca Solidac Sobral Fclha LArg. Umoeiro Atacall Ubajar. Ju. des Vieiras Avg , Plaolíog dao.íly of 10,000 plfha. plet area 2,5QOm·. 2 FF • Farmors' Heldo Fresh AY ~/hal ppp 36.70 30.25 12.70 17.20 15.38 22.44 302 FF' 22.30 22.30 11.90 7.60 6.47 14.11 Variety Jcao Grande F,agosa Guarani Fragosa Cruvefa Aas1eíra 16.3.4 Processing technology. A very important factor in implemenling cassava-based rural development projects is the farmers' adaptation to the new processing technology, where they are supposed lo function as suppliers of raw material to markets tha! present specific characteristics of demand, quantity, quality and frequency. Experiences in similar projects in Colombia and Ecuador indicate that this adoption and adaptation process requires an adjustment period of 2 to 3 yr before the groups are able to administer their processing units efficiently. To assess this, a parameter called "Processing Efficiency" was used, which compares the real annual output of the drying plants with a theoretical output estimated according to climatic and management factors. These two factors combined determine the efficiency and Ihe level of profitability with which Ihe farmer groups operate their enterprises. In the case of Ceará, it was assumed th~t tor a 6-mo period (Aug.-Dec.), the climatic conditions are suitable for solar drying of cassava chips and that a loading rate of 10 kg/m' takes 2 days to reach desired MC levels (10-13% wet basis). Management factors inelude the no. of times Ihat Ihe drying plant operates during ttíe year and Ihe amount of chips processed each time. Table 16.4 compares Ihe processing efficiency of 10 agroindustries Ihat operated in 1989· 90. It can be observed that during the second processing season the total annual output ot Ihese plants increased by 191% and that the avg level of Processing Efficiency was 511.8% as comparad with the fjrs! year (20.2%). In general it can be concluded that after 2 lo 3 yr of operation. the efficiency levels of the drying plants must be between 6O-70"k to be profitable. 16.3.5 Commerclallzation • Marketing channels. Different from what had been expected. Ihe product is not being sold to large-scale consumers such as the animal feed industries as was the case in Colombia and Ecuador. In the case of Ceará, 70% of Ihe total production in 1989 was purchaSed by low- to medium-scale consumers « 10 tfyr) and 30% was sold to consumers with purchases over 10 t/yr (only 5% of the total no. of consumers). In 1990 Ihere were some changes, with the larga-vol. consumers purchasing a slgnificant 62"k of total production but they continued to represent only 5% of the total population (Fig. 16.1) . • Dry cassava and cassava flour pricas, 1990-91. It has been estímated that thare are more than 14,000 cassava flour procassing units ('casas de farinhaU ) operated under communal-type arrangements and located throughout the rural areas of the state. These farinha agroindustries process an estimated 200,000 tjyr. The commercialization system within which these farmer groups operate usually forces them to sel! at low prices. Factors such as the poor quality of the farinha, the small 303 Table 16.4. Efficiency of p,ocessing 01 ca"""", dryíng planto Cea,. ?rojee!. 1989-90. Annual CMput t/yr Drying Are. A.:tual F'armer Groups (m') Theor. 1 1989 1990 Poco d.os Gavatos 600 360 108.3 .289.3 Jua dos Vieiras 600 360 00.9 238.8 1..>908 Grande 434 2EIO 12().Q 311.5 Fellla larga 700 42() 92.6 248.9 SoIidao 450 270 118.0 232.2 Dourado 700 42() 3.9 132.3 Lagoa do Mato 700 42() 47.5 166.5 Cechoeira do Boi Morto 360 228 17.3 102.2 _ro 600 360 4l.9 108.3 Sena Santan. 400 240 5.0 86.4 Total 3.338 E51.4 1.896.0 Avefage locrement (%) 191 , Éstlmated .ccordiog 10 the following assumption: loading rata of la kg/m". 2Iots/wl<. 2() wk/yr. 2 Processlng Effieiency caleulated as \he f«Hatlonship betweet'l actual and theoreticsJ annuat output. Effieiency of Prccessing 2 1989 1990 ;30.1 74.8 26.9 66.3 46.1 119.8 22.0 59.2 43.7 66.0 1.0 31.5 11.3 39.6 7.6 44.8 11.6 ;30.1 2.1 38.0 2().2 56.8 191 scala of operation, and the rudimentary type of infrastructure utilized prevent the farmers from obtaining better incomes from this processing activity. The KF-funded project offers a solution to this situation by creating a marketing alternative for the cassava farmers through the production of dry cassava chips for animal feed, where the farmers can gel better prices and increase their net incomes. Far the last 2 yr the cee, in collaboration with the RCCs, has been monitoring the evolution of prices for cassava products and by-products in the areas of project influence; and it has been proved that during the processing period (July to Dee.), the produetion of dry cassava chips has been a more profitable activity for the farmer groups than proeessing the roots in the form of farinha. In 1990 the production of dry cassava chips gave the farmers a net protit of Cr$ 10.14 for every kg produeed vs. a netloss of Cr$ -10.29 tor farinha. In 1991 data for the first month of processing are showing the same tendeney, with the dry cassava giving the farmers a net profit of Cr$ 4.17 vs. a net loss of Cr$ -8.33 for each kg of farínha produced (Fíg. 16.2). 16.3.6 Organization • Institutions. The CCC is now fully established as coordinating body for statewide project activities, and the 5 RCCs formed in the main areas of project influenee are playing an important role in the deeentralization of the planning and execution of activities that the CCC has been promoting. The RCCs are gradually assuming 304 .- ... _----------------, ¡ ~ r::::J 1989 season (N'125) lB 1990 season (N'410) ! 400 ;..:N~O~.~O~f~c~o~n~su=m~e~rs~------------------------------------! 300 200 ~ ·1 100 r , O~· --'o_- 6 19 16 3 e 5 tly 5-10 l/y • 10 l/y Volume 01 annual purchases 01989 season (N-28S) _ 1990 season (N-1431) .. of tolal production 70~------------------------------------------- 60.2 62.1 80 30.4 e 5 tly 5-10 l/y • 10 tl'l Volume of annual purchases Figure 16.1 Dry cassava commerclalizatíon channels In Ceará. 305 a Raw material _ Selllng price Precessing Seasen 1990 m Flxed costa [¡;;llil Prom. Plocess,ng costa Cr$/kg (real prices) 1991 • 100 80~-~~---~~----~~------------~~---------, '00 % 20 -20 ~---------~-,----------------~-----------~---------~ Ory Cassava Chips 1:1 Raw materIal lIIIIIIlII Selllng prlce Processing Season 1991 _ FI •• d costa _ Prom. Farjnha Processlng costa Cr$/kg (real prices) Aug. 1991 • 100 80r-~~--~-----~---------- - -------~- '00.0 '!lo 60 40 20 -20'-------------------------------------'------------~ Dry Cassava Chips Farinha Figura 16.2 ProflIability of dry cassava and farinha in Ceará. 306 responsibility for coordínating regional-Ievel activities in areas such as selecting and c.ganizing new farmer groups, traíning, technical assistance and identifying new sources of funding tor expanding the project to other areas. The CCC offers logística! support to the RCCs and maintains overalf coordination of the project at the state level. The incorporation of these commíttees into the prevailing institutional landscape and their acceptance by their two main counterparts--EMATERCE and EPACE-as valid and tirlctional organizational forms is an important achievement and has facilitated the performance of projact personnel in the!r activities as advisors lo tha termar groups. • Earmer groups. The organization of tarmer groups for installing and operating the cassava-based agroindustries has been a very important activity. Oefining the type of organlzational structure that Is most appropiate for small farmer groups is not an easy task, especially in the case of the Ceará project, where the majority of the groups were iritlalfy based on larger communal-type farmer organizations (casas de farlnha). The cec and the RCCs are maklng seríous efforts to support already eslablished groups (20-30 socially homogeneous families) in order to capitalize on the degrea of social oohesion which they had attained. Initíal efforts are basad on tha specific task of processing dry cassava fer animal feading, which damand sophisticated new skills 01 . management and organization that may be beyond the groups' current capacity. To date, the project has been able to achieve a reasonable degree of success in establishing small farmer cassava-based agroindustries. In addition to reactivating the initial 12 groups in 1989, 47 new groups were organizad in 1990, fer a total of 59 farmer groups. ltis expected that by the and of the project (Mar. 199~), there Will be 80 groups in all. Thare are 3 fundamental elements critical to this process of expanding the social basis upon which the KF-funded project has been built: .. The farmer groups are being formed on the basis of a collective action utilizing a single investment. The function of the group is to gain access to a grant-type of financial resource for a specific purpose decided upon by the group-in thís case the installatíon of a dry cassava processing agroindustry. .. Membership of the group is initially decided upon among the potentially interested members, and all future decisions on new membership have to be agreed by all members. .. The role of the cec and the RCCs has been crucial in the task of approaching different goverment agencies and development programs on behalf of the farmer groups to obtain grants. At the same time, both these committees have acceSJ to KF project funds tor assisting and supporting tarmer activities. 307 Cne of the principal proJect strategies in the area of farmer organization is the creation of second-order groups at regional and state levels in order to improve the participation and the bargaining power of the farmers in planning and implementing the project. Efforts are being made to stimulate the formation of this type of organizational structures in the main areas of project influence. 16.3.7 Training The project's training strategy includes tour types of activities (Table 16.5): coursas, seminars, field trips and special days. The training courses are being conducted at state and regional levels, with the RCCs assuming greater responsibility for their implementation. especially in the case of the courses held at the regional level. The seminars are aimed at encouraging and facilitating tarmer participation in the planning and evaluation of proJect activities at both state and regional levels. The field trips a1low beginning farmer groups to become familiar with the principal technical, administrative and organizational aspects needed to operate the cassava-based agroindustries efficiently. The special days are held for purposes of promotion and divulgation of information. Some of the farmers act as trainers, not only allowing significant cost reduction but also stimulating farmer participation in proJect activities. 16.3.8 'M&E The M&E system ot the Ceará project is aimed at tracking progress toward achieving speclfic project activities. Technicians and farmer-managers of the dry cassava agroindustries are now more familiarized wlth the methodologies and the forms being used to callect and process information on day-to-day project activities. Additionally, the exchange of informanon among farmers, technicians, RCCs, counterpart agencies and the CCC, as well as the feedback of the processed information to the technicians and its delivery to the farmer groups, has reached the point where results are being used to guide planning and evaluation activities. Areas in which the M&E system is currently providing information are as follows: • Land tenure. Currently there are 59 tarmer groups participating in the project for a total ot 1380 direct beneficiaries. The land tenure system under which these farmers operate their holdings ¡nelude 3 forms distributed as follows: owners (59%), sharecroppers (13.8%) and renters (27.2%). • Age of the participants. Most of the farmers participatíng in the project are between the ages ot 30 to 60 (70%), 17.4% are younger than 3D, and only 12.6% are over 60. • Production. A central assumption to test the success of the proJect in relation to eassava productíon ís that the opening of an a1temative, more profitable outlet tor cassava will motívate the farmers to expand theír cassava plantings. Monitoring is 308 Table 16.5. Training evento 10r \he period 1989-90. 1!!!ll1 1!!!iQ ll!!il No. 01 Trainees' No. of Trainees No. of Trainee. Training evento evento T F Events T F Evenl$ T F LCout .... A. S!ate Leve' ProduCllon 38 2 30 24 Organizatlon 01 tarmer grouP'l 30 20 30 Prooeuíng 15 OIagnoslic okll'" 24 e. !lI!Igional L.ve' Prooeosing 5 21 8 20 55 11. Semlna,. A. Stale Level Planning 1 26 16 Evalualloll 1 14 17 • 33 Organlzatlon 1 '35 B. !lI!IglonalLevel Plannlng 5 47 54 119 Evaluatlon 5 25 35 111. fIeId Tripa 15 30 123 11 12 82 7 12 38 120 1 15 120 IV. Specla' Day. 2 15 VI 0Iher. Counse 10r Tachnicians from Paralba & Pamambuco SIudy lOur 10 Ecuador 20 6 4 & Colombia T_ 22 186 276 33 227 370 25 176 215 1 T,. Technrcians,. F = Farmers. Total no. of training events : 80 Techniclans - 591 Farmers ~ 551 being done for two parameters: the size of the cassava plots and the relationship between cassava plot size and the farmers' land tenura system. In relation to plot size, the information available shows that the cassava areas have remained fairly constant, with near 80% of the farmers involved in the project over the last 3 yr planting no more than 2 ha. Farmers planting larger cassava plots ( > 5.0 ha) do not represent > 5% of the total group. 309 Cl1988 (N·68S) _ 1989 (N·946) O 1990 (N·1257) No. of farmers 700TI-------------------------------------------,l" I 600 i I 500 ~ , . i 400i 300 200 ,lOO O 0.0-1.0 1.1-2.0 2.1-S.0 3.1-5.0 , 5.0 Cassava Plot Size, 1989-90 (hd) Figure 16.3. Siza 01 cassava plots, 1988·90. CI 1988 (N·68S) _ 1989 (N·94S) 1990 (N·1253) Mean Cassava Area Planted (ha), 1988-90. r 5,-----------------------------------------------, 4 s 1.95 1.117 OWNER8 aHARECROPPERa RENTERa EN TIRE POPUlATION Land Tenure System Figure 16.4 Cassava plol siza va. land tanura, 1988·90. '310 . Information coUected on the relationship between cassava plot size and the farmers' land tenura system (Figs. 16.3-16.4) indicates that the smallholders have generallly been planting larger cassava areas than sharecroppers and renters. The avg size of cassava plantings for the entire population in 1990 was 10% smaller than in 1988, which could have been caused by a significant decline in avg cassva plantings by the sharecroppers who now plant cassava plots 15% smaller than 2 yr ago. Monitoring of these factors will continue to assess changas that could ba relatad to projact impacto 16.3_9 Resulta of the processing season, 1989-91 • ~. The first year of the project, 53% of the roo15 processed were coming from nonmembers and 47% from tha farmer groups. In 1990 the vol. of roots sold to tha dry cassava agroindustries represented only 28.6% of the total vol., nonmembers selling 71.4%. After the first 2 mo 01 processing in 1991, it was found that members had sold 53.1% of the total vol. of roo15 processed by the agroindustries. Of the total population of members, 46% sold roots to the drying plants in 1989; 33.6% in 1990; and 29.6% only 2 mo into the processing season of 1991 (Table 16.6). This situation needs. to be monitorad and analyzad carefully as it could become a limiting factor in overall group performance . • Total annual incomes. 1989-91. Tha total annual benefits received by the mambers include cessava sales, processing wages and the sharing among members of annual profrts. Distribution of the total annual incomes according to tha size ofthe cassava plots planted indicates than the greatest part of the banefits go to the tarmers with fewest resources. In 1989, 78.2% of the total annual incomes went to farmers with < 2 ha cassava; in 1990, 77.9%. Data for the first 2 mo of 1991 indicete that only 30.3% of the total annual incomes was gained by those farmers who had > 2 ha plantad in 1990. Comparing total annual incomes during 1989-91, it can be concJuded that the benafits generated by the project thus far have incraasad considerably and that the main beneficiaries continue to be farmers with fewer rescurces (Fig. 16.5). Table 16.6. Cassava _s, 1989-91. No. 01 farmor groups """", .. Iog dry cassava No. 01 mombeJa "" membeJa 88lllng rcota "" membeJa eamlng processing wago. r_ vol ~) of cassava roota processod "" soId by membors "" ooid by nonmembors , Oniy 2 me of processing season. 311 19119 12 211 40.3 15.1 265 53 47 1990 1991' 33 36 !156 84Q 33.11 29.6 12.0 10.7 3,802 1,461 29 53 71 47 _ 1989 (246 t) _ 1990 (1431 tl D 1991 (533 t) US $('OOOs) 25. 20 15 10 5 O ( 1.0 1.0-1.5 1.6-2.0 2.1-3.0 3.1-4.0 • 4.0 1990 Cassava Plot Size (ha) Figure. 16.5. Total annuallncomes, 1989-91. 16.3.10 Future Plans The Ceará Integrated Cassava Project ends in March 1992. By that time, the project is expected to have accomplished its main objectives in relation to agroindustrial development, institutional building and improved welfare for its beneficiaries. SEARA has set up a working group that is formulating a proposal for a second phase of this project, which will be presented to the KF. The proposal includes a continued support role for the CIAT Cassava Program although it is expected to decline over time. At the same time, EMBRAPA (HQ) and the CNPMF have been working on the formulation of a project to expand the successful experience of the Ceará project to other states in NE Brazil. To date, this initiative has not defined clearly the role that CIAT will play in this process, but demand for increased assistance in the planning and organization of the projects is likely, 312 17. PARAGUAY Paraguay is the largest per cap ita producer of fresh cassava roots in the world. The most important producer of fresh cassava is located in the periphery of the capital Asunción. Soil degradation in this area has pushed production towards Caaguazú, some 300 km away. This area now supplies cassava for most of the country. The native forest recently cleared in Caaguazú opened up new land for agriculture. Cassava production líke most other crops depends largely on the natural fertility of the soils; however, soil fertilíty is rapidly declining in Caaguazú and crop productivity will probably decrease soon if new improved technology is not applied. Most available technology to improve cassava production has been developed far tropical conditions. The subtropics have received relatively less attention from rasearch. Consequently, developing better technology for the specific environmental conditíons of ' Paraguay implies relatively more technology adaptation activities in addition to strategic research. Research institutions in Paraguay are relatively weak and staffed malnly with young, inexperienced, but highly motivated professionals. The Agricultural and Cattle Extenslon Service (SEAG) is probably the most active institution, with offices and pel'$Oflnel stationed throughout the country. CIAT coordinates activities mainly wíth SEAG although support for the rasearch branch of the Ministry of Agricultura 15 aiso provided. IORe- Canada funds most of the activities in Paraguay. The French Technical Mission a1so previdas support, mainly far postharvest activitíes. The Cassava Project in Paraguay, like many others in Latín America, simultaneously considers improvement of production, utilization and marketing of the product although initíally more attention has been dedicated to solving production problems. Project activities are concentrated in the states of Paraguarf and Caaguazú a'though multilocation tests are frequently conducted throughout the country. Contrary to Caaguazú, recently open to agriculture, Paraguarf has been intensive!Y cultivated tar many years due to its proximíty to Asunción. Climatic conditions and general soíl characterlstics are very similar far the two localities; however, avg soíl OM and P content are significantly higher in Caaguazú. . 17.1 Productlon A rapid Initial survey in the 2 ProJect areas complemented the already existing information about cassava in tha country. At the sama time, an agroacological study was conducted, native varieties collected, and the most commonly cultivated variaties characterizad agronomically. Sugarcane, cotton and cassava ara tha most important crop componants of tha typical farm in Caaguazú. Cassava is cultivatad on mora than 90% of tha farms; 313 actual plot size for fue different crops depends on market conditions and they are rotated on a 2- to 3-yr basis. A survey of more fuan 300 farmers revealed the great importance of cassava as an on- farm animal feed, mainly for poultry and swine that are consumed locally. Approx. 50% of the cassava production is sold tor fresh consumption or transformed into starch; fue rest is consumed on farm (humans and animals). Large farms tend to consume proportionally more cassava tor animal production and at the same time inelude more cattle in the animal subsystem of the farm. Manual labor in cassava is intensive, particularly for land preparation and weeding. Harvest date depends mainly upon the type of variety planted and market conditions.The most interesting feature of cassava production in the subtropics is managing the erop to cope with low temp during the growing season. At the onset of the low-temp period, farmers cut the cassava stems at ground level and store them as planting material. As soon as the temp rises again, maize is planted in the interrows of the cassava stubs and the two crops develop simultaneously. Farmers developed this technology by trial and error. Planting new cassava fields also takes place as soon as the temp rises again; and both old and new plants coexist on the same farm. If prices for roots are low, the cassava remains in the field. On one farm it is possible to find 1-, 2- and 3-yr-old cassava fields. Farmers use oxen-driven implements for land preparation and the first weeding of cassava and other crops. Farmers identified the following as important production constraints: the sporadic incidence, but high severity of CSB, storage of planting material, weed control, marketing and declining soil fertility. Scientists idantified soil erosion and germplasm diversity as important production topics affecting the sustainability of the system. The Ministry of Agriculture presently keeps a germplssm collection with > 150 accessions. CIAT previdas technical assistsnt to characterize this collection, which is duplicated in the country and kept in vitro. Most of the Paraguayan accessions are now part of CIAT's germplasm bank in Colombia. An agronomic collection including the most commonly cultivated varieties as well as other promising material is continuously tested across different localities. In terms of fresh root production, significant differences exist among fue cassava varo adapted to the ecological conditions of Paraguarr. At both project sites, varo Meza-I and Tacuara say-yu performed best. At Caaguazú these 2 varo have continuously outperformed most other varo tested (Table 17.1). There are differences between the two localities not onlY in yields but also in terms of serial biomass production and harvesting time. 314 Table 17.1. Avg yield el mari<81ablo roots (t/ha) al dllferenl eaS1llMl var. I ... tod In Paraguarf and C8aguazú. Loeality Paraguari caaguazú Varlety 1989 1990 1989 1990 T apoyo 8 moroti 22.2 18.4 28.7 . , Poi 15.6 22.8 23.7 Coron61 23.7 24.7 28.3 Pyta~ 16.5 20.5 19.5 CabalIe~ 16.9 19.8 20.9 21.8 Cano-! 20.5 20.0 25.5 28.0 Can6 22.5 24.8 17.3 Tacuara say-yu 27.9 35.1 27.6 25.9 Meza-I 28.5 36.9 29.2 33.2 Pomberi 20.9 25.3 23.4 22.0 , NoIloatod that yr. The varo Meza-l. which consistently outyielded other varo in Caaguazú. a1so has good DM and starch content. It is not highly resistant to CSS but performs acceptably weU under field conditions and hig~ incidence of the pathogen. This varo is now cultivated by several farmers but stiU needs more díffusion in areas where it could coexist with other varo Earliness is a characteristic that most farmers consider of vital importance in the selection of an improved cassava varo Table 17.2 summarizes earliness of Meza-r and other varo commonly cultivated in the country. Table 17.2. Total roo! _ wt (t/hal of 5 cassava varo hatvestod al3 dlfferen! dales. Mo after PIonting 6 9 15 Meza-! 27 34 54 Cassava Varieties Tacuara say~yu 22 36 33 315 Caballero 25 25 35 Cano-! 28 28 49 Pambari 23 30 32 Storage of planting material under low temp occurs only in the subtropics. Farmers frequently report losses in both quantity and quality of planting material stored more than 3 mo, the length of time stakes have to remaín stored before the temp ríses again and planting can be done. Traditional storage teehnology consists in eutting, leaving the freshly cut stems in the open for 3-5 days, piling them horizontally and covering them with plant debris. A set of experiments eondueted wíth farmers provided ínformatíon on the effeet of vertical and horizontal storage of stems. Vertical storagE:: resulted in signifieantly less wt loss (fable 17.3). Vertical positioning for storage is now being recommended to f3rmers by the SEAG. The slow initial growth of cassava keeps soil unproteeted until the eanopy closes. Several basie strategies to cope with this problem have been suggested: (a) intereropping eassava with speeies that grow more rapidly; (b) inereasing eassava planting density; (e) speeding up the growth of cassava with fertilizers; and (d) planting early-maturing, more vígorous varo Intereropping and fertilization tríals are being eondueted with fsrmers in Caaguazú and Paraguarl to assess their economíe viability andtest theír possible advantage in redueing soll erosiono The most eommon praetiee of associating cassava with other annual erops is the simultaneous planting of tha 2 species. At both projeet sites, however, the first weeding of the interrow spaees using an ox-driven hoe takes place when eassava is still relatively smal!. Consequently, the erop that is planted in alternate rows with eassava can only be planted immediately after this first waeding. This delayed planting inereases tha competition with cassava and has a negativa effeet on the yield of the intarerop. Simultaneous planting of eassava and cowpeas in alternate rows results in better yields; however, alternate row planting can be practicad in Paraguay only by farmers who cultivate very small plots, where weeding is mainly done by hand. Experimentally, one and 2 rows of cowpeas planted in the interrows of cassava did not result in significant differences in yields between these 2 spatial patterns. Cassava AYs were not affected by cowpeas although cowpea yields were redueed more than 50% due to competition with eassava (fable 17.4). In other environments and cropping pattems, cowpea yields are reduced by no more than 20% when intercropped with cassava. Table 11.3. WlIOII$ ('lb) aft!l, $IOring eassava planting material fo, dilf ..... nt periodo; avg 01 2 varo Posi!ion for SlOrage Weignt lass After' 5 Days 90 Days Vertícal 5& 26& HOfizontal 5& 19b 1 Pereento followed by the same latter are not slalislically differen! (P~ 0.05). 316 100 Days 31 • 25 b Table 17.4. Y",lds cf ca ....... (all9 of 3 var.) Intereropped wi!h cowpeas In IWO opallal arrangements end sol. -pI· Spatial Cassava eowpea$ Arrengement (l/ha) (t/ha) LER 1 row of cowpeas 17 0.46 1.3 2 ro"", cf cowpeas 15 0.46 1.2 Sole CfOp 16 1.7 1.0 Maize has been traditionally intercropped with cassava in Paraguay. Farmers intercrop these species in different spatíal patterns, depending on thair relativa prices in tha market. Intercropping msize with cassava in Paraguay is siso affacted by the fact that maiza planting has to be delayed until after the first weeding of cassava. Different cassava var. were intercropped withmaize varo Suwan-B047 and also planted in monocrop as a check. In the intercrop most of the cassava var. performed similarly in terms of fresh RY, particularly marketable roots; in monocrop, however, performance was significantly different. Cassava var. Cano performed equally well in sole crop and in association; whereas other varo such as Raí yielded significantly more in sole crop. Maiza intercropped in ¡;lternate rows performed significantly better than maize intercropped in the row (Table 17.5). The performance of the cassava/maize intercrop depends, among other factors, on the types of varo that are usad. Short-stature new maize varo usually compete less with cassava than traditional tall var. The majoríty of the newly released maize varo are short types with high Hls that perform better at high plant densities, both in monocrop and in association with cassava. Among the most important maize varo released recently in Paraguay is Suwan-8047. As densíty of this varo increases from 7000 to 28,000 pi/ha, RYs ofcassava var. Píta-i tend tó decrease (Ag. 17.1). Cassava'reduced its yield by 60% when intercropped with msize at high plant densities. The relationship of these two well- Table 17.5. VI.lda cf ca ........ (alÍg 3 var.) Inlar_pped wllh mai •• (var. Suwarn-8027) In alternale ro"", and In lhe"me row.' Cropplng P_m Malze/cassava, ah:erna.te rOW$ Malze/ca_ aamo row Solo CfOp Caaaava TotalRoots 13.;; a 13.6 a 20.4 b 1 Ylelds foIlowod by !he _ letter aro nol statlotically dilferenl (P. 0.05). 317 MIIrkatablo Roots 9.7 a 10.1. 15.4 b 2.78 1.9b 3.08 Maiz" YI"I". /l/ha) Ca •• ava RO"I YI"ld. (¡/ha) 3.5 ';---~~~~~~~~~~~~~---, 21 r: 3 2.5 I I 1 0.5 OL_ .... J.~ ____ -L __ ~~ ____ ~ ___ ;_ 7142 14285 20000 28571 Malza Populatlon (pI/ha) C .... va Rool Ylal". -j- Malza YI"I". Sele Maize 111 Sote CassaV8 18 15 12 9 6 3 O Figure 17.1. Yields 01 cassava intercropped with maize p1ants al different denslties. adaptad cassava and maize varo provides a valuable tool for predicting cassava RY at different maize planting densities. Intercropping maize' and cassava in the same row permits the simultaneous planting of the 2 species and also weeding the interrows with the oxen-driven hoe. Cassava varo recommended tor both intercropping and monocrop are now available tor both project areas. A set of experiments aimed at reducing soil erosion by increasing cassava population in Paraguarr indicated that the arrangement in actual use by tarmers (1 x 1.2 m) is the only economically teasible alternative and pr.obably the only one acceptable to farmers. Highar plant densities resulted in en excess of smatl noncommercial types of roots with no industrial use as 01 yet. 318 Organic fertilizers are used by some farmers due to the local availability of manure, but their use is not yet widespread. Amounts ranging from 5 to 40 tfha of manure applied before planting and combined or not with chemical fertilízers have been testad in Paraguarl. Twenty tfha of manure has resulted in better yields and handling possibilities for farmers. The use of chemical fertilizers for cassava production, combined or not with the use 01 manure, increases production costs beyond the avg farmer's economic possibílities. Nevertheless, the use of chemical 1ertilizers is one of the 1ew possibilities for increasing cassava production in Paraguar!, at least in the short termo Similar to other areas of the wOrld, land preparation, weed control and harvesting in the subtropics are cultural practices that demand high inputs in terms of labor. Paraguayan farmers use preemergence herbicidas only for cotton production, given the favorable prices for the producto The most commonly available preemergence herbicíde is Codal-400 (alachlor and prometryn). This herbicide Is selective also for cassava; and given its availability to farmers, it was tested for cassava monocrop production. In preliminary tests 5 It/ha plus one hand weeding 60 DAP resulted the most effective for weed control and high cassava RY. Production costs were too high given cassava prices. In another set of experiments Codal-400 applied in bands only to the sites where cassava stakes were planted reduced doses and costs by 50% with good control of weeds. The use of this pre-emergence herbicide should be complemented with at least one hand weeding to obtain good yields. 60th the total and the localized application of Codal-400 increased yields by 17 to 40% in comparison with the check that was only hand weeded (Table 17.6). The development of technological components to improve cassava production in Caaguazú and Paraguarr has loo to the formulation of 2 complete technical recommendations for farmers, consisting of a mix of improved components as well as traditional technology that is constantly being evaluated in farmers' fields. The recommendation for Paraguarl includes improved component for varo (Meza-O; better selection 01 stakes before planting; planting stakes of approx. 20 cm; the use of 20 tfha 01 manure before planting plus band application of 300 kg/ha 01 12-12-17-2 (NPK and Mg). The rest of the technology (Iand preparation, weed and pest control) remains exactly as the farmer's traditional practices. Fer Caaguazú the complete package consists Table 17.6. Eflect of pr .... me'g~n'"' herblcide appIIed al1wO doses and han Poor record-keeping capacity and lack of short- and long-term production plans are also important problems. Raw material costs account for approx. 70% of the total production costs of cassava starch; labor, 14%; and drying 5%. These costs are very similar for small cottage-type and large, more mechanized starch factories. The differenca between these two types of factories lies in their transformation efficiency and profltability, which are higher in the latter. The majority of small factories sell starch to intermediaries while larger factones obtain better prices by dealing directly with the end user. Unlike Other countnes, by-products of starch factories are seldom usad for animal production or other purposes. A pilot (demonstration) plant will be built in the near future by a coop, using the improved small-scale technology developed in Colombia: grating with a system of cutting blades. starch refining using a vibratory sifter, separation by sedimentation in settling channels. and solar drying the final producto 17.2.2 Fresh root conservation The technology to keep chemically treated fresh cassava roots in plastic bags for long periods to facilitate commercialization, was developed by CIAT. This technology is in stage of either testing, adaptation or adoption in different countries. Ex-ante analyses in 321 Asunción and Caaguazú assessed the feasibility of this taehnology for marketing cassava roots. Several sueeessful experimental trials were conducted with the participation of farmers, transporters and merehants. At the producer level most of the experimental actMties were developed at Juan Manuel Frutos (Caaguazú). Trained farmers carried out the harvasting, salaction of roots, applicatíon of Martect (thiabendazole), paeking, weighing and transport of the treated roots to the Asunci6n market in small trucks. Treated roots (approx. 30 t) wera sold directly from trucks, in small- and medium-sized shops, and in supermarkets. Potential siloppers were stratified by avg family ineome for acceptability studies. Most consumers informad that aftar a ona-wk storage period, treated roots had the same quality as fresh cassava. Slight deterioration occurred in treated roots after 2 wk. Consumers from al/ ineóme strata reported that quálity of the treated roots was either 9000 or very goOO. Restaurant customers did not report significant differences between the stored and fresh product. Close coordination between teams of packers and distributors of treated roots is recommended. Transporting and distributing no more than 12 tjday in 5-kg plastic bags 5eems most conveníent. In the near future, it is hoped to expand this project component te a semicommercial level. 17.2.3 Commercialization The commercialization studies in Paraguay bagan with an ¡nitial survey to identify the origin (farmers); most common form of transport; wholesalers; distributors; stores and supermarkets; and finally different types of consumers of fresh eassava. An estimate of the no. of persons involved in the different stages of the commereialization proeess was used to stratify the population and carry out an in-depth survey of a representative sample. Tha Municipal Market (DAMA), located in the heart of Asunción, is the place where most 01 the fresh cassava arrives from the production regions and from where distribution begins. There are three different types of cassava dealers: (a) the truck owner who purchases eassava at farm gate and transports it to the DAMA (nearly 75% 01 the cassava sold in Paraguay); (b) the truck owner who only transports cassava; and (e) the farmer who also owns a truek and transports his product (14% of the total cassava that reaehes the DAMA). As Caagúazú is 300 km away from Asunción, cost of transport is the highest ítem in the total cost of commercializing cassava. The farmer who transports and sells his produet directly profíts more than the rest of the dealers. Fram DAMA 63% of the cassava roots are dístributed to supermarkets and small retail stores; the rest goes to State-run mobile markets that sell cassava and other products in different parts of the city. Approx. 26% of the cassava that wholesalers' purchase deteriorates befare further distribution, but only 13% of the roots deteriorates at the retail stores. This provides a strong justificatian for the fresh root conservation component of this project. 322 During the high-temp period (Dec. to Feb.), the total vol. of cassava passing through the DAMA decreases significantly, more being so Id in the relatively cooler period (April-Sept.). From approx. Oct. to Feb., sweet potatoes compete with cassava, both in the DAMA market and in consumer preferences. 17.2.3 Cried cassava for animal feed A pilot natural drying plant has been built near Coronel Oviedo as a collaborative venture between the SEAG cassava project and a farmer coop. lhis will test the technical feasibility of drying cassava chips naturally in Paraguay, where the subtropical, seasonal climate presents different problems from those of the tropics. lhe coop will use the cassava to produce its own balanced feed rations. 323 18. SEED SUPPLV SVSTEMS Improved seeds constitute the biological input through which biogenetic innovations are incorporated into agricultural production systems. A lack of adequate seed supply systems (SSS) prevents farmers from benefiting from efficient seed-embodied technologies. Consequently, R&D activities on SSSs has become strategic in modern agriculture. Organized SSSs have shown multiple beneficial effects· by accelerating the flow of improved varieties from the research to adoption phase, expanding the magnitude of adoption, and prolonging the productive life of varieties. More specifically, improved seeds benefit the farmer through higher yield, production stability, lower costs, prevention of disease spread, reduced use of toxic agrochemicals, etc., all of which are associated with the judicious use of quality seeds of improved varieties. However, when superior clones are released and recommended, only a few make it to farmers' fields. The development of SSSs, especially under atomized and risky market conditions, has been under-researched. There is a general lack of a knowledge base for organizing SSSs under small farmer market conditions. Given this growing concern, CIAT initiated a R&D thrust in 1988 focusing on seed systems. This activity is c?rried out by the Seed Unit as a multi-institutional eftort under the leadership of ICA and CIAT. Colombia has become an ideal country for this type of research given the achievements in market expansion for cassava end products, the release of superior varieties, and the interest of ICA's Seed Division and a hast of other local organizations in cassava seed supply. The objective of this thrust is to develop a working model(s) in Colombia. It is expected that the experience accrued, including socio-organizational and production technologies, will enable CIATto understand the constraints limiting the development of SSSs and to gain a;knowledge base that can be transferred to other countries through CIAT's wide range of international activities. 18.1 Problem Definltion It has already been widely documented that in Colombia cassava is produced by the small farm sector (see Chapo 14). This sector is becoming integrated into the market through cassava drying coops, which creates a need for improved seeds. The drying coops have been successful, to the point where they- are expanding spontaneously. In addition to ICA, CIAT, DRI and CORFAS, there is a good tradition of participation of a host of local organizations in cassava R&D. These factors, coupled with the growing interest in new cassava varieties, create a favorable environment for developing a cassava SSS. Differences in market, farmer organizations and agroecological conditions must be taken into account when developing an SSS. Excellent sites for studying alternative SSSs include the North Coast, the Piedmont, the Coftee Belt and northern Cauca State, which differ in both market and agroecological conditions. 325 At the same time, the cassava plant has other specífic characteristics that have impUcatíons for the SSS: 18.1.1 The market Cassava growers have limited resources. Traditionally, the source of seeds are the stakes saved by the farmer; e.g., a survey on the North Coast showed that tarmers leave a portion of the lot to be harvested before next planting or keep the branches to assure availability of planting materials. They become interested in seeds trom outside sources, mainly when changing varieties or when their own planting material is inadequate. These conditions make it nearly impossible to estimate the real demand and plan ahead oftime the quantities of seeds to be produced. At the other end of the spectrum, farmers in Ouindro (Coftee Belt), have a tradition of buying and selling stakes. Among these contrasting situations, a gradient of market conditions exist, which clearly indicate the nead of relevant SSSs. Genotype-agroclimatic interactions have imposad a need for specific varieties for given r¡:¡gions. Thesa conditions are reflectad in small markets scatterad in different regions¡communities throughout tha country. This ImpUes that a few larga saad emarprises supplying genotypes to larga and uniform markets is unlikely in the foreseeable future. A logical alternativa would be to develop SSSs closer to the farmers, in !he communities where seeds of specific varieties' are needad. Furthermore, this alternative has the potential for strengthening genetic diversity. 18.1.2 The product Biologically, cassava is a slow, low-vol. seed producer. It takes from 9-12 mo to obtain good root yield and mature stakes. At the end of this prolonged period, each plant will have produced only 5-10 stakes. This contrasts sharply with the multiplication rates found in true-seaded crops. The seeds (stakes) are bulky, heavy, high in water content, and perishable, thereby posing innumerable economic, technical and logistical problems in. seed management. Thus the SSS cannot depend on storage for long penads or transportation over long distances. This aga!n points in the direction of local seed production. Another important product feature is genotype replicability as the crop is propagated through clones. This means that once a variety is acquired, the farmer would not need to buy sseds of that same variety. Although this can not be generalized for all market situations, there is a clear indication that replacement of planting materials will depend on availability of seeds of new varieties. 326 18.1.3 Institutlonal panorama In the best of cases, the cassava SSSs receive support from research through germplasm development. Basic seed supply, quality assurance services, credit and marketing technologies are scarce. In most countries, institutions have not yet focused on the development of cassava SSSs--in great contrast with the institutional support systems developed around hybrids fer maize, rice, sorghum, and others with more advanced SSSs. Vis-d-vis the limitatíons, some strengths existo lhere ls an expanding market for cassava, which encourages the search for improved seeds as an input capabla of increasing yield and produet quality. Varieties with these features haya recently been released for key regions of the country. lhe ICA Sead Division has created a new saetion that focuses on developing SSSs for unaddrassed markets. It is anticipated that this organization will capitalize on local organization to bring about the institutional support for cassava SSSs. Coops and private organizations are emerging in an autonomous fashion for the purpose of producing, processing and marketing cassava produets. lhese conditions are ereating an overall favorable environment fer developing a cassava SSSs. lhe conditions of the market, the biological nature 01 the commodity, and the instítutional conditions cleariy point to the need of alternative schemes rather than one monolithic approach aeross the land. lhe system, including its socio-organization and production technologies, needs to be conducive to the Inítiation 01 a seed supply process. Furthermore, the overall national system and the specific local cases will need to respond to the conditions of the market, the erop and the agroecological conditions where it operates. 18.2 ActIvlties lhe focus· has been on developing pilot sead multiplication units with the active participatíon 01 private farmers and cooperativa organizations. To support thís objeetive, the followíng specific areas were addressed: training, basic seed supply, sead produetion methods, and norms and standard s for quality control. 18.2.1 Tralning lhe first course ever in cassava seed produetion (1988) was spolJsored by ICA, with CIAT playing a supportive role. lhere were 14 participants from ICA's Seed Division and CRECED regions, as well as representatives of prívate seed enterprise (MAIZENA and PROACOL). lhese professionals have become key promoters in seed produetion in their specific regions. lhe insights gained from the tirst experiences in Sahagún, Córdoba (CRECED-Sinú) and PROACOL, for example, were presented to 29 participants in the second course in 1990. 327 18.2.2 Basle Seed Supply Pure and healthy basic seed of improved varieties is of catalytic importance far developing an organizad seed multiplication and distribution system. As ICA had not inmated this activity, the CIAT Seed Unit undertook basic seed production with the purpose of being a role model, on the understanding that this would be strictly to support the evolution of the system on a temporary basis. ICA has now begun organizing this function in VilIavicencio (Meta) and in El Carmen de Bollvar. 18.2.3 Seed productlon methodologles There are two basic methods of cassava propagation: One is basad on lab and greenhouse techniques, where rapid increase and freedom from diseasas are the main objectives; the other is the traditional method of propagation by stakes. The latter method was ahosen as the most likely te set up the process in motion; in vitro methods, when commercially viable, would be useful to perfect the system in the future. The rationale for beginning with stakes is that this technology a!réady exists in cassava production systems so there is enough improvad technology that can be usad right away. Moreover, the simplicity of the method would make it possible for local institutions and farmers to participate in the process. Most component technologies were available at CIAT, which were rapidly assembled in a stepwise fashion at the' Seed Unit. The improved-stake selection method is based on good agronomy, selection of mature branches of healthy and high-yielding plants at harvest, and cutting and treating prior to utilization. To expedite the process on a more commercial scale, technology innovations in cutting, treatment and packaglng were incorporated. Currently, there is a prototype at CIAT, where this process cen be demonstratad. 18.2.4 Norms and standards AA organized SSS requires norms and standards to ensUre the quality of seeds being marketed. In latin Americe, with the exceptionof Cuba, there has been no experiehce in developing these standards, much less in implementing them. Neverthelass, basad on tMe experience of the Cassava Program staff and sead scientists from ICA and CIAT, a draft proposal entitled "Minimum specitic requirements for production of certified and selected cassava stakes' was prepared for aventuaL submission to the Ministry of Agriculture. As the ICA Seed Division considers cassava seed supply a nascent actMty, it has opted not to implement these measures. Recently, enterprises such as COAGROARAUCA and ASOQUINDIA--anticipating potentia! externa! markets-have requestad certification of their seads; thus there appears to be a need for implamenting certification on a selective basis. 328 18.3 Challenges far the Futura The proces$ has been set in motion, but a great deal awaits further actions. The objectively verifiable outputs of tI1is R&D tI1rust are that important functions such as basic seed supply and commereial seed supply--tI1e two misslng links of the chain at tI1e outset- -are evolving. As a result increasing amounts of seeds of traditional and improved germplasm are becoming available. 18.3.1 Basle seed 6upply Basie seed supply, whích has been the key for the massive utilization of varieties in other crops, is by no means a solidly institutad function. CIAT assumed this function because it was the key barrier that had to be overcome to set the process in motion (Table 18.1); however tI1is needs to be strengtl1ened by ICA in Villavicencio (Piedmont) and Carmen de Bollvar. Furthermore, good coordination between variety release and basie seed production is needad to prevant situations where a new variety is released (e.g., ICA- Costeña), but tI1ere is no seed available for the farmers. To have a graater multiplying effect and impact, basie seed may also need to be directed purposefully to committed seed multipliers. Table 18,1. Availebllity of basic cassava _ a1 CIAT Ha. AugullllGG1. T .... dl!!on.l "",leila. Chirozl ""nezolana Regional Amari". Role.sed .arletl!. Pll P12 P13 ICA-Calumare fCA.Costella P .... ""'!Hd cIonu ... For Ibe NorIh CoIoot CM 33Q6..4 CM355!Hl CM33Q6..4 b. For eaotem Plalnl ro 165-7 CM 2165-6 TOTAL CIon"/Vari"ty , Estímate of _ ... !o be halVested. 329 No,o! Stake.' Hatvest Dale 10.000 April92 15.000 April92 4.000 Sept, 91 110.000 April92 100.000 April92 40.000 April92 150,000 April92 50,000 April92 50.000 Sept. 91 40.000 Sep!. 91 100.000 ApriI92 6,000 Sep!,91 20,000 Sep!. 91 595,000 18.3.2 Commercial seed production The central challenge has been and will be for some time in the Mure how ro institute commercial seed production in specialized organizations. Training, basic seed supply and production teehnologies have all centered around this objective. Many organizations have participated in commercial seed production (Table 18.2). The partieipating organizations ¡nelude cassava drying coops, individual private entrepreneurial farmers and large prívate enterprises. In some cases the seeds are for intema! consumption of the organization such in the case 01 integral coops. In other cases (e.g., COAGROARAUCA and ASOQUINDIA) the seeds are finding internal use as well as surrounding markets. Experience has shown that entry into commercial seed production has been dynamic although there are a no. of cases of enterprises movlng out of cassava seed production. Although a tria! perlod is loglcal before producing seed on a permanent basis, this "entry- exir' phenomenon needs eareful analysis, for axpansion of the system will depend upon the institutionalization of cassava seed supply in suppliers with fue knowledge and skills to' assura avaílabílity of quality stakes at a low cosí. This will require continuity and specialization. The rapid entry of new venturas can be tracéd back to the growing interest in new varieties. Availability of foundation seeds supplied by CIAT and ICA, accompanied by training on seed production technologies, has besn the key for materializing that interest in concrete actions. It should be noted that insufficient quantities 01 basic seed has been a limiting factor which has resulted in reduction 01 area planted for seed production in many cases. The abandoning of eassava seed production activities has been frequent as well. The major factor mentioned by those who have stopped this activity has been the high risk due to price fluctuations for the roots, which has had a negative effect on the demand far stakes. When analyzing the cases, a trend can be faund: Those more directly interested in cassava roots and end products seem to stay in the business. This had besn anticipated from the market analysis, and the results tend to confirm this hypothesis. This suggests a possíbility that seed supply can be 'piggy backed' onto those organizations that have a more direct intere:;;t in cassava production. It is almost impossible to predict accurately which organizations will endure in seed production. Some groups may engage in this activity with a one-time objective--as a mechanism to obtain basic seed of new varieties. Others may havesome comparative advantages and new business opportunities so they can produce seeds in a continuous fashion. Thus participation of a wide ranga of organizations in commarcial sead production sesms to be necsssary step in the induction phases. 330 Table 18.2:, Organizations that haya or are currenlly par1icipating in cornmerciaJ production of cassava seeds in Colombia, 1991, Organizalion Sitos Varíeties Area (ha) Ob_tions Norlh Coas! AGROESTACION SuOf. JCA.Catumare 0.5 Discontinued, unstabte market COAGRO-AL8AN1A SuOfe Venezolana 1.0 Discontinued. unsteble m8lk&! MAlZENA-INYUCAI. Barranquilla v..nezo!anaP-12 2.0 Oisoontinued, unstable market APROSOCORRO El Socorro, &Jefe Venezolana, 10.0 Discontinued. chango al plan. Costeña COPROALGA Algarrobo. Córdoba Venezolana. P-I2, 1.0 In operation 1 Costeña COPROTUCHIN Tueh!n.Córdoba eos;eña 0.3 Soaree basic aeed cauaed reduc:tion In area planted COPROSAN San Andrés. Costoña 0.3 $caree basic _ cauaed Córdoba reduetion in area pIanted ICA¡DRf SuOfo Venezolana 19.0 Promoting produetion In 19 organizatkms Pledmont COAGROARAUCA Tame, Arauoa Catumare. 15.0 In expansion Cebu.án COAGROCASIBARE Casibare. Meta Catumare 2.1> In operation. ClAT-PNR Agreement Ceuca fUNDAEC Santander P-ll. P-13 2.0 NGO promoting aeed production among farmet'$. in expanslon CETEC Santander ICA-Catumare 0.5 In expansion Celdas ASOOUINDIA Armenia. OUind!o Chiroza 5.0 _ .. tion al Agronomlsts. in expansion Others PROACOL Palmira, Valle P-". P-l2, P-13 3.0 Discontinued, unstable market Alvaro Aya Palmira. Valle P-ll 1.0 Oiscontinued. unstable market Jaime Sandi Jamund!. Valle P-ll. P-13. 10.0 Discontinuad. unstable market lCA-Catumare Samlllas del Tolima Ibagué P-I1. P-12. P-13 3.0 Oiscontinued. unstable m8lket RamIro Restrepo Honda. T olima P-12 6.0 Oiscontinued. unsteble market , Because of difficulties in the ooop, ooe 01 the members decided to -ce c:- Cl ¡,ro 12 ¿ 110 .- ''1:1 ,lJ---ry/ 1 .- AII-enlry mean 100 8 (Ieft 5cale) 90 \, 80 I j\AII-enlry mean 70 60 (relativa lo Rayong 1, righl scala) I I I I I I I I I 1982- 83 84 85 86 87 ·88 89 90 Yaar 01 planting Figure 20.2. Change In mean dry RY of yield trlal entrias In Thailand; a11-entry mean ls the mean of·a11 reg!onal trIaI entries (8-10 clones) at aJI the regional trIaI sitas (6-8 locatlons). 120~ ~ ~ O'> c: o ~ ce 1""1 _0---0 _----0----0-- , /o--__ ---==:I!:::::::=:j!t::::::~ -::::_:::ii-1J"7~ "O ~ ~ c: m E c:- e ID • :( 60 Fresh RY \ ¡ I I I I ¡ j I 1982 83 84 85 86 87 88 89 90 Year 01 planting Figura 20.3. Changa In mean Irash RY and DM content 01 yield trlal entries (all-entry mean) In Thailand. 345 ~ ~ O> c: o ~ ce "O ~ ~ >-ce i::' 't:! 11 i ce 110 ~ ~ O'l lO o 100 "" ro /"-\ ~if \ ~" n ...o~ \ /,----.a / \ ". , .. ..1""'1 /' D"" a-- -................. /" Harvest index "O" Il: 'O ~ ~ 90 - c: ro '" E <=- 80 e ID ...l. « 70 -1-9'-82~~á3--á4--8rI5~-8T'6---r87----'Ba~~89r--9''0- Year af planting Figure 2!t.4. Change in mean HI and biomass of yleld trlal entrles (all-entry mean) In ThaHand. .. Rayong, the Thaí-CIAT populatíon was far superior to the CIAT HO populatíon for all yield traits; but the difference was not conclusive at Lampung (Table 20.1). The result at Rayong was somewhat predictable because Ihe Thai-CIAT population carne from cross parents selected at Rayong (local varieties and clones of CIAT HO origin). This suggests that the Thai-CIAT breedíng population has added adaptation to dríer elimates and that for ímmedíate varíetal selection for the semiarid lowland tropics of Asia, Thai-CIAT materials may offer a batter ehance, while bolh are a good source far expanding the germplasm horizon. Thís shows once again Ihe basics of germplasm varianon; i.e., germplasm from the center of origin and díversifieation offers abundan! genetie variation, but the germplasm of each loeation contains genes for local adaptation in a mueh higher frequency· 20.2 Oistribution 01 Breedlng Materlals to Aslan Programs A large number of clones are produeed and distributed annually to national programs. These materials offer not only genes for general high-yielding capacity but also genes for adaptability and tolerance to many adversa factors--biotic and abiotic--as a result of 346 ~ ....¡ Table 20. ,. ComparlllOO between CIAT Ha and Thal-CIAT populatlons In yleld characte", al Rayong, Thalland and Lampung, Indonasia.' Rayong, Thalland Thal-CIAT Populallon CharaC1ef CIAT Ha Population (% o! CIAT Ha) DlyAY 100 138*'* F, •• h AY 100 132-- Total pianl wt 100 119" HI 100 111** Root OM conlonl 100 104"'* Plant type ,aling 100 119" Germination/sul\!1val of planting stakes 100 160""* Lampung, Indonasla CIAT Ha Populatlon 100 100 100 .100 100 100 Thal·CIAT PopulatlOfl (% 01 CIAT Ha) S7" se" S7" 99 101 121*1r 1 Dal.lrom a síngl •• row tríal al Rayong 1981 (No. antri ... , 755 CIAT Ha 01008$ and 1228 Thal-CIAT clonas) and a repllca18d yleld trlalal Lampung 1968 (No. antri •• , 30 CIAT HO clono. and 41 Thal-CIAT clon •• ) ••• copllar gormlnation/suMval data whlch camo lrom 1989 rapllcatad ylold tríal. repeated evaluations in harsh environments such as Carimagua (Colombia) or Rayong (fhailand) . Distribution of breeding materials to Asian programs was as follows: • Sexual seeds from CIAT Ha. A total of 274,196 sexual seeds from some 5,400 crosses have been sent to 9 countries since 1975 (fable 20.2). This is still a major source of breeding material for many national programs in Asia. • Sexual seeds from the Thai-CIAT proaram. A total of 47,224 sexual seeds from so me 960 crosses have been sent to 11 countries since 1985 (fable 20.3). The importance of this source is aainina significance in many national proarams. • Advanced clones from CIAT Ha. Some 186 clones have been transferred to 5 couQtries since 1975. This was an important source of varietal selection in the early years. • Advanced clones from the Thai-CIAT programo Some 215 clones.have be en transferred to 11 countries since 1988 (fable 20.4). This is a highly significant source for immediate varietal selection by young national programs, su eh as Vietnam. 20.3 Varletal Selection 20.3.1 Thalland The Thai-CIAT cassava breeding program may be the largest national cassava breeding program (see Chapo 19). From 1984-93, 7 clones with different adaptive niches have been or will be released. The earlier varieties tended to be adapted more to the relatively fertile soils of better-off farmers. Vigorous efforts are being made to select new varieties adapted to less-fertile soils of smaller farmers. 20.3.2 Indonesia Selection for high yield and starch content from local, CIAT Ha and ThaJ-CIAT materials has been highly successful (fable 20.5). Promising results in drought tolerance and broad adaptability are also emerging. • 20.3.3 China Some promising selections have been made from the CIAT Ha materials; yet, the best selection appears to come from crosses between local clones and locally selected CIAT clones (fable 20.6). Selection from Thai-CIAT populations has also begun. 348 • rabie 20,2, ea ...... F, .... ual _. 'rom cIAr Ha distrlbutod lo AsIan programo, CountJy 1975 1977 1978 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Total Thalland 900 6170 7720 3050 1400 7450 7900 8000 9300 8000 11800 14200 9400 14121 9583 118994 indonesia 900 700 4800 3050 4980 9800 8000 4219 33019 China 2300 6100 3500 1500 2100 4400 8399 7400 5511 41510 Vietnam 1900 4250 7700 3750 17800 W Phllippin •• .¡,. 900 980 5100 4700 27.6b 29.4 ab 5.8:3 reduced erosion compared with no-ridging, and live barriers of elephant and setaría grass were more effective in reducing runoff and erosion than a barrier of peanuts or hedgerows of Leucaena leucocepha/a or Glirickiia sepium. Cassava and maize yields were highest in the presence of elephant grass barriers. Unlike most other trials with grass barrier strips. there was little competition between cassava and the elephant grass and no significant reduction in cassava RY; moreover. farmers can cut the elephant grass to feed their cattle, making this a highly attractive management practice. In Thailand a total of 9 erosion control trials were conducted at 3 locations for 3-4 yr in collaboration with the Dept. of Agriculture and KU. Table 21.5 shows the effect of various crop and soil management treatments on erosion and yield at Sriracha Station. Contour ridging and fertilizer application consistently resulted in high yields and low levéis of erosiono Intercropping with grain legumes such as peanllts was very 360 ro[ A 40 (ij' 20; .;;: :;:, -fJ) fJ) ..Q .¡¡¡ 2 ¡:; "ti "ti SOr' B Q) 'iá :; E :l ~ 40 20 o 2 - 400LC E §. :lii c: .¡¡¡ 200-~ ~ -c: o ;:¡; O D J .. • .. • • • ~-o ... cassava 273><60.60 cm • cassava ; 200x50 cm • eassava ; l00xl 00 cm a peanU1S TI rice 4 maiZe , , I 4 S 8 • cassava; 1 00x1 00 cm v cassava + maize + narpeanuts o C$$sav8 + peanuls - mung bean. 4 <:assa'llS + maize - peanuts 4 6 8 Months afler planting ~ .. F M A M J J A I 10 10 FIgure 21.3. Accumulated dry soilloss by eresion in various monocrop (A) and intercropping (8) systems durlng a 9-mo cropplng cycle on 5% slope in Tamanbogo. Lampung, Indonesia In 1989/90. Rainfall distrlbution is shown in C. 361 1600 ca ~ 1200 a: § -~ 800 (ij > o- e o 400 ~ A Cassava spacing (m) 1.Ox1.0 2.OxO.5 2.73xO.6xO.6 Figure 21.4. Effect of cassava planting arrangement on (A) total crop value and (B) soiI lossas du& to ereslon tor elther monocropped or Intercropped cassava. grown In Tamanbogo. lampung. Indonesia. Data are avg of trials conducted lo 1987-89. Intercropped cassava la avg of 4 cropplog systems (C+M+R-P; C+M-P; C+R-P; and C+P-Mu. where C=cassava. M=malze. R=upland rice and Mu=muog bean . .. 362 T able 21.4. Effect of vatfous systems on cassava and mm. yields, as well as on runoff and erosiono on an 8% slope at Jatikerto Experiment Station in Malang¡ indonesia in 198&-89. Yleld Il/ha) Runoff Ory SoU (% \.0$0 Treatments Ca""",. Mai.a Rain/all) (t/ha) ConIDur ~dge., no r .... barríor. 28.35 3.00 13.90 44.8 ConIDur ridgo., e!ephant gr ... barrier. 25.90 3.64 9.82 20.2 Contour rldges, setaria grasa barriera 19.90 3.53 10.90 34.2 Contour ridgos, peaIIul _m 18.05 3.25 12.17 37.4 ConlOur ~.s. G. SlJP/um barriero 17.90 3.20 12.48 39.0 Contour ridges, L t.ucocephala barríO,. 24.55 3.15 13.10 40.8 No ridges, p88IIu1 barri.m 25.10 2.83 16.64 64.5 No ridge8, _ gr~ barrio,. 25·59 2.91 15.52 49.2 Fallow plo! 119.0 effective In reducing erosion, but usually reduced cassava RY slightly. Intercropping with pigeon paas markedly reduced yields, making this practice unacceptable to #armers. Reduced tillaga, subsoiling or zara tillage were practlces that generally reduced erasion and somatimes (but not always) Increased ylelds. Uve barriers of elephant or king grass were less atrectlve in controlling eraslon and ustJally reduced cassava RY. Table 21.6 ranks !hase practlces in terms of erosion control, yleld and costa. Complete land praparation followed by contour ridglng was very effactive in reducing erosion and incraasing yield but was relativaly cosUy. In contrast, complete land preparation without ridging and no fertilization was the least costly. but also the least etractive In reduclng eroslon or increasing yield. An intermediate practice of no-tillage, closer plant spacing and fertilizer application was quite etrective in reducing erosion and produced reasonably good ylelds. making this an attractlve practlce ter farmers with limited resources. However, adequate crop resldue management ::md weed control are crucial for thls practice to be successful. After several years of conducting erosion-control trials at varlous experlment statlons In Asia, some soil and crop management practices have been identified that soom promising in terms of erosion control and yleld enhancement The next step will be to test some of thase practices on farmers' fields In collaboration with the extension seNices in order to obtain tha opinion of farmers and extension agents on which practices may /may not be adoptable and why. In Thailand 4 such trials have been established on farmers' fields in Rayong Province In collaboration with the Depts. of Agricultura, Lana Development and Agricultural Extension. Depending on the outcome of thasa trials and 1he Interast shown by farmers and axtension agents, these trials may be expanded in the future into other cassava-growing arezs of the country. Field days and the publicetion of simple pamphlets or booklets on cassava production and soil conservation are planned for the future. Similar extension activities are also planned in other countries. 363 Table 21.5. Bfect 018011 .... 0 9'0p management on ooillo •• (t/ha) and """ .. va AY (1/ha) In 4 •• "",imenta al Srkaolla RosaatOh Slation, Chonburi, Thalland. 1i§7 12M2 1989" 19902 Sl' Vlold Sl Vllid Sl VI.1d SL 'Vilid Soil ano Crop Managemont No IIIlag. 93.8 30.7 5.7 26.2 11.6 11.4 7.4 18.2 2 plowlngo followed by 2 disklnga, no r1dglng 30.5 25.8 11.1 31.1 6.0 12.0 10.4 16.0 2 plowlng ..... 0 2 dlsklngs lollow.d by contour ridglng 11.1 35.5 5.1 29.6 4.1 11.4 3.9 18.1 Subooiling a! 4O-cm deplh, 8().om row wldlh, ",,_va al 80 • 125 cm 22.2 34.3 4.6 22.5 6.9 11.7 6.4 17.3 Tr •• I. 2 but wllh conlour bank ovel)' 10 m grown with L loucoc"phata Strip proparalion wllh 4-disk plow, proparlng 2-m oontour atrips alt.rnated wilh l·m unprepared . 17.7 27.4 3.9 12.4 4.5 11.0 strip; ca ... va a! 100 x 67 om ú) Plowlng only once wHh 4-dlsk plow 42.1 26.7 al 2 plowlngs and disldn". loUowad by ul>-.nd-down rldgl09 37.5 22.3 .¡:. T ... a!. 2 wHh "" ..... in double rOW8 anu ím.,oroppeo wllh 2 row$ 01 mal •• 35.4 25.6 11.8 33.2 Tr.a!. 2 wHh ....... v.ln doubla rowe and Inl8rClOpped wittl2 rowa 01 peanuta 44.9 22.5 Tre.1. 2 and Inta""opped wittl 1 row 01 Mehl. plntol 17.2 22.8 4.7 26.4 Tre.1. 2 and inte ... opped wHh 1 row 01 plgean pea. 5.4 11.9 Uve barrior of Blachlarla gr ••• , 2-m wlds conlOur slrip ovel)' 10 m 7.1 2.4 Uva "a"ler of elephant gra .. , 2·m wlde conlOu( strip evary 10 m 52.7 22.6 Uva barrler of king gr ..... , !-m wio. con1Out strip evary 10 m 35.5 23.3 5.1 12.0 10.6 11.7 T .... I. :¿ bu! wiIIIout !ertOízer appUcation 9.4 11.0 12.0 12.8 Troa!. 2 bu! planted al 8OK8O cm (15,625 pi/ha) 64.9 • 18.8 16.7 24.1 Tra.!. 2 bu!dlemicai weed control 24.4 25.4 LSD (P < 0.10) 26.7 30.9 2.7 5.3 3.9 NS 1.6 1.2 , Sl = SoIIIoss duo 10 8f08ion 2 Avg of 2 rapa. Table 21.6. Ranking 1 of casava cu!tural practlces in term$ of eresion control and yiéld, based en 9" 8rosion control mala conducted al 3 Io<:ations in Thailand. Erosion Control Y1eId Cast Plowing, dlsklng. centour rklging 1 1 8 PIowing, disklng. peanut in\ercropping 2- 7 5 PIowing, disking, planting cassava at elo .... spacing 3 5 4 No IItlaga 4 4 2 Plowing, disklng, elephant graso barrier 5 6 6 Plowing, disking, no rklging 6 3 3 PIowing, dlsking. up.ancklown rldging 7 .2 7 PIowing, dl8klng. no ferlilaer appllcation 8 8 • Tho lowéI' 1M ocore the better, • 21.3 Soll Fertillty Malntenance Most cassava soils in Asia are characterized by low soíl fertility (Table 21,1); moreover, cassava is aften grown on severely eroded slopes with very low levels of OM and plant nutrients. Many short-term. fertilizer trials conducted by national programs have shown that cassava responds mainly to the application of N; significant responses to P and K are much less common. Long-term fertilizer trials conducted in lhailand (3 locations), India and Malaysia, however, indicate that after several years of continuous cassava production, K deficiency became the main IImiting factor ane! high cassava RVs could be maintained only through adequate K fertilization. lhis Is because cassava extracts relatively large amounts of K from the soil, most (60-70%) of which is removed from the fiald in the root harvest As most tropical soils contain only limitad reserves of K-bearing minerals, the K-supplying power of these soils will eventually become exhausted. In order to sustain high yields of cassava, it is essantial to maintain tha productivity of the soil by applying chemical fertilizers or organic manures, by enhancing N inputs through biological N fixation, and by recycling leached nutrients. Research was conducted to determine the optimum rates and methods of fertilizer-appiication and to develop economically feasible practices of green manuring, intercropping, covar cropping and alley cropping. ' 21.3.1 Fertilization To determine both the short-term fertilizer response as well as the long-term fertilizer requirements, simple long-term NPK trials have been established at 13 locations in 5 countries. lhese trials generally have 12 treatments with various combinations of 4 levels of NPK, in such a way that the response to each efement can be determined while the other two efements are applied at near-optimum levels. 365 Figure 21.5 shows the results of the second-year trial at 2 sites in China. There was a significant response to the applícation of N only al levels of 50 or 100 kg/ha. The combinad application of high levels of NPK more !han doubled yields in Guangzhou and significantly increased yields (up to 100-50-100 kg/ha of N-P.O.-K"O) in Nanning. In China no significant responses to K have yet been observed. In contrast, in Bac Thai Province of northem Vietnam, K applied to these extremely poor and eroded soils quadrupled cassava RYs (from ca. 8 to 30 tfha) !he first year; there was also a significant but less dramatic response to N and P. Despite intermediate levels (1-3%) of soil OM, there were significant first-year responses to N application in 7 out of 10 locations; while significant responses to P and K were observed only in 3 and 2 locations, resp. It is expected that K responses will become more pronou"nced after more year=5 of continuous cassava produetion. Despite low levels of available P in most cassava soils in Asia, the response to P was seldom significant, indicating the presence of a very effeetive natural mycorrhizal association. Although significant responses to some nutrients have been observad in most locations, cassava tarmers seldom have the resoOrces to buy chemical fertilizers. In India and Indonesia this is partially overcome by the application of larmyard manure and/or wood ash. While cassava responds favorably to these natural sources of plant nutrlents, the level of application is seldom high enough to sustain high yields. Other practicas that may supply additional N through biologicai N fixation, or !hat reaycle lsachad-out K, -ea and Mg, have been investigated through collaboratíve mals, mainly in Thailand. These mals evaluated native or introduced legume species and testad sorne al !hese species as green manures, cover craps and intercrops, or for alley cropping systems. , 21.3.2 Intercropping Intercropping cassava with '\farious grain legumes usually decreases the yields of the component crops but increases the total productivity of the land and farmers'· incomes while reducing risks. At Hung Loe Station in southern Vietnam, intercropping consistently reduced cassava yields; grdwing cassava in single rows (1.0 x 1.0 m) produced better· yields and net income than planting eassava in double rows (2.0 x 0.8 x 0.71 m) (Fig. 21.6). Peanuts, winged beans and mung beans were the least competive intercrops, which also produced the highest net income. Cassava monocropping, however, gaya higher economíc returns than any of the intercropping systems. Table 21.7 shows !he effeet of cassava planting arrangement on cassava and grain legume yíelds in 3 tríals condueted in Thailand from 1988-90. Planting cassava in either single (1.8 x 0.55 m) or double (3.0 x 1.0 x 0.55 m) rows had no consistent effect on cassava or intercrop yields. Thé intercrop yíelds were significantly higher in the single-row system in 2 of!he 3 yr; however, the double-row system allows the planting of a second intercrop after the first crop has been harvested, thereby increasing the farmer's total income. In general intercropping reduced cassava RY, especially in the case of the 3B6 ~ ~ ;g .!t ª ~ tll ., ~ A SOr Guangzhou 1990 20, ¿;::.O- -=8 .0--- 0-""" / / 101- (f' ~/. ..0---- . 0-- " , c:t 0\ l. l. B 30r Nanning 1990 20r- p-__ (.1',/ J""L ---o ~ ¿ 10 o 200 o 25 50 100 o-.o~ 0-- ..0---0...---__ 0 ~I)~, / ' d :-0-_----0 O 50 100 200 / / o/ [J SC 201 o SC205 ,,_-----0 / / ./~ rr o-o-o 100-50-100 200-100-200 Figure 21.5. Effect of NPK application on cassava yields during the second year planting in Guangznou, Guangdong (A) and in Nanning, Guangxi (8) China in 1990. A 20 - 15 - '" g >-ce: '" ~ '" e> • :JI c: 'c: '" ~ :> 01 ~ () 01 ~ ! ~ 1II '" ~ ~ E • ~ • '" ~ . o. .s '" Figure 21.6. Effect of various Intercropping systems on cassava yield (A) and net Income (8) when cassava was planted in single or double rows in Hung Loe Center of Oong Na! Province In South Vietnam. 1990-91. 368 ,,",~,,,,,,, , Tabl. 21.7. Etf8 O> c: 111 J: ~ LU 7 6 5 o no green manure, no lert. o no gresn manure, with lert. 4 • gres n manure (110), no lert. B t 12 0.8 .t 0.4 O 100 e t 80 60 • 40 20- ¡ = cassava planting o 6-888-88 6-898-89 6-908-90 6-91 8-91 MonlhlYear Figure 21.7. Elfect 01 3 yr 01 cassava cropplng, wlth and wlthout green manures (avg of 10 spp.) and with and wlthout fertlizers on (Al sol pH, (8) OM and (e) exchangeable K content In Pluak Oaeng, Thanand. 372 Table 21.9. Effect oIlntereropping ca,,,,~a wilh I"guminou. ""ve, ClopS on caasava AY in th,ee mal. al Pluak Daeng. Thailand,' Cassava AY ~fhal 1_ 1989-90 j990-91 No CO\I8I' crop 11.68 a 7.79 a 19.62 a Covor aop 01 S!yIosanthes hamata 10,2] ab 3,91 e 4,45 de Cavo, ClCp 01 1Wch1. plnfoi 8,46 be 6.56 ab 9.71 cd Caver ClCp 01 Centmsema aculifolium 7,66 be 6.68 ab 15.33 ab eover crop of Centmsema pubescens 7,51 be 5.60 be 6.17 d Cavar crop of Mimosa envisa 7,49 be 6.48 ab 13.33 be Covor ClOp 01 o..smoálum ovaJifolium 7,26 be 6.78 ab 13.46 be Cavar ClOP 01 Macroptilllum a/ropcI s.m-a-...e c:J~T"""",," 0 ........ -mi'l~~~c fEZ1~Hc:!ó __ B ~Bnmi.M--''' mm~~-­mm l..-ds..-.d _ _ ~s-IJI4""""'''''''' Figure 23.1. ellmalle homologues for cassava belween (A) Brazíl and (B) Arrice. 380 • Transplanting healthy seedlings to the field and inspecting them closely under field conditions in the first month after planting, followed by monthly evaluations for the main pests occurring in each ecosystam Sy following this scheme the probabilities of introducing any exotic diseasas into Africa are minimal, and the opportunity to eliminate diseased individuals occurs at various stages of the process. The selection of the evaluation sites in Nigeria was made on the basis of the maps of the climatic homologues (Fig. 23.1) according to the. cassava elimates defined by Carter (1987)."" The 4 evaluatíon sites in Nigeria are: Ibadan (Lowland Semihot Isothermie); On':1e (Lowland Tropical); Kano (Lowland Semiarid Isothermic); and Jos (Mid-alt.). Evaluation of the progenies introduced in each of thase ecologias was based on the reaetíon of families and individuals within families to the main diseases and pests of cassava in Africa (ACMV, cee, anthracnose, CGM and cassava mealybug), plant growth and habit, and qualitative and quantitative traits measured at harvast. ~ eharacteristics being evaluated and the frequency of evaluations are shown in Table 23.1. Monthly measurements of individual plant ht ¡:¡nd no. of leaves/pl on 10 pljfamily were used as estimates of plant growth. These measurements are useful in identifying genotypic differenees in plant vigor. Evaluations for host plant resistance wÉlre made on individual seedlings, 1hereby obtaining data on incidence and severity within and across families. Evaluation for leaf retentton Q.e., the portion of the plant with leaves at harvest or the thickness of the canopy) provides an indication of the plant's ability to retain its leaves at the end of the growth periodo This ability ls expected to be (educed by environmental factors such as drought and biotie constraints. The no. of families introduced at Ibadan, Cnne and Kano were 149, 78 and 62, resp. The total no. 01 seedlings transplanted and harvested at each localion in 1990 were, resp., 5400 and 2993 in lbadan; 3699 and 3010 in Onne; and 3288 and 2458 in Kano. Differences between no. of seedlings transplanted and evaluated atharvest were due to Iosses durlng establishment and, in Ibadan, by severe incidence of ces immediately after transplanting. Rainfall distribution at the 3 Iocations Is presented in Figure 23.2. Seedlings from seeds introduced to the 3 locatíons and at Jos (mid-alt.) in 1991 are still under evaluation . ... , Carter, S.E. 1987. CoIlecting and organizing data on the agro·soclo-economic envlronment of!he cassava crop: Case study 01 a method. In: Buntlng, A.H., (oo.) Proc. Workshop on agro-ecological characterizatlon, c1assificatlon and mapplng. FAO. Rome. 1986. CASI. U-K pp. 11-29. 381 Table ;23.1. Evaluation parameters determined in cassava germplasm of Latin American origino Parameter ReactiOn to pasts Plan! h! Ht of 1$\ branoh ara"eI! level. LDdging Stak9S{pl Lesf retention Root skin ~O( Roo! tl •• h color Root cortex color Root eonstrlctions Roo! hesh wt -{pl HCN oontsnt (rooto) Meallness 700 600 ~ 500 ~ 400 -j I 3001 200 ~ I 100 Jan Rainfal! (mm) Jun ------~-~~~ 1990 ---------- Monlh Frequency ¡nme ot Evaluadon Monthly Monthly At harvest A! harvesl At harvest Al harvelll Al he"",st Al harvest Al harvest Alharvest Al ha"","1 Al hatvest Al harvest At haNest Al harvelt D ONNE _ IBADAN e KANO Figure 23.2. Distribution 01 rainfail ¡rom Jan. 1990 lo June 1991 at 3 localions in Nlgerla. 382 23.2 Reaction to Pests Data on incidenca and severity of ACMV, CSS and CGM at various stages of the growth perlad are presentad. Incidence is represented by the no. of seedlings showing symptoms in relation to the total no. of individuals evaluated. Severity is the degree of susceptibility of plants showing symptoms measured on a scale of 1 (no symptoms) to 5. 23.2.1 ACMV Under conditions of high pressure from ACMV, the Latin American germplasm showed a high degree of susceptibility, particularly at Ibadan. Seedlings started showing symptoms of mosaic as early as 2 wk after transplanting. Progenies from crosses invoMng TMS 30001 and TMS 30572 showed a higher degree of resistance to ACMV. Severity of ACMV infection at Ibadan 8nd Kano is shown in Tables 23.2 and 23.3. As the disease pressure increased, there was a tal']dency for fewer seedlings in Class 1, but more in classes 4 and 5. At the end of the rainy season (mo 5), at Ibadan, only 0.65% of the seedlings evaluated were in Class 1. During the dry season an increase in the no. of seedlings with no or few symptoms was observed as a result of the recovery of seedlings previously showing severe symptoms. At the beginning of the second rainy season, the level of symptoms increased again, suggestíng a relationship between symptom expression and high rainfall. Table 23.2. $everlly 01 ACMV aod C88 al lbedan: % individual. in each cia •• of Ih. evaluation seal •. Evaluation $cale 1 2 ;! 4 ~ MAr' ACMV cee ACMV cse ACMV Ce8 ACMV ces IC.MV cee 1 34.1 23.5 Z7.7 14.3 0.3 2 2.2 40.2 1.2 22.1 7.8 3.6 40.0 121.9 48.7 21.2 3 1.5 63.3 4.0 15.6 25.2 4.7 43.4 8.3 25.9 8.0 4 1.1 30.8 2.4 57.8 8.8 8.1 25.6 2.7 61.1 0.5 5 0.6 53.1 3.8 32.1 26.0 11.1 42.5 2.3 27.0 1.3 6 1.3 64.1 5.2 31.1 35.3 3.7 47.7 . 1.1 10.4 0.3 7 5.8 58.9 15.9 32.7 64.2 7.6 13.5 0.6 0.6 0.2 8 6.3 62.6 11.0 29.9 SO. 1 6.4 30.0 0.7 2.1 0.3 g2 3.1 82.5 22.3 13.8 51.8 3.2 19.6 0.4 3.2 0.1 10 0.4 14.2 SO.3 23.9 l.2 1 Monlh alter ttanoplanting. 2 Evaluation for CBS was nol done alter Ihe 8th MA T. 383 Tabl.23.3. MAT 2 3 4 5 6 7 8 9 10 Severity of ACM'! and eBa at Onne: % indlviduals in 8ach class of the evalúation acare. Evaruation Seale 2 3 ACW CBB ACW CBB ACMV CBB ACW 96.1 93.9 2.4 5.4 0.8 o.s 0.2 94.1 96.2 0.9 3.5 1.6 0.3 1.4 87.7 94.5 2.1 5.3 3.0 0.2 3.3 16.9 94.3 3.9 4.1 8.3 0.4 7.2 87.3 88.5 9.5 9.9 19.8 0.9 10.5 34.1 96.8 6.2 1.8 42.8 1.0 15.1 38.7 55.2 27.8 37.8 29.8 5.4 6.0 69.6 79.4 9.6 16.3 20.8 3.5 4.4 3.0 11.$ 60.4 39.6 31.4 27.2 q¡, Seedlings Showing Symptoms 100 ~ e IBAOAN _ ONNE 80 60 40 20 o L-"--_"""',-"" o 1 2 3 4 5 6 7 8 Months After Transplanting 4 CBB 0.1 0.3 0.4 0.3 1.5 0.3 4.2 5 ACW CBB 2.0 0.1 4.0 3.7 0.1 2.8 0.3 1.2 0.1 0.2 0.5 11.0 0.9 10 Figure 23.3. Ineidenee of ACMV on cassava seedllngs growlng al lbadan,and Onne. 1990-91. 384 1 Differences across locations for severity and incidence of ACMV infection were great. The level of symptom expression was higher at Ibadan than at Dnne (fable 23.3), where a mal<. incidence (65%) was attained at 9 MAT (Fig. 23.3). These differences were probably caused by a higher concn. of inoculum in the surrounding cassava fields. It may be that the high temp observed at Kano during the dry season inhibited symptom expression of ACMV. Given that the period of highest ACMV pressure was batwaan July and Nov. 1991 (2-6 MAT), comparison of the families' reaetíon to the disease was made considering the avg no. of individuals showing no or light symptoms (classes 1 and 2 of the evaluation seale) per family (fable 23.4). Families with more than 10",4, of the individuals showing few or no ACMV symptoms were predominantly progenies derived from clones TMS 30001 and TMS 30572. . . 23.2.2 ess The incidence of CBB was also higher at Ibadan than at Dnne (fables 23.2 & 23.3). Aftar asevere outbreak of CBB in the first MAT at lbadan, which contributed to the death of approx. 20% of the seedlings, levels of incidence were reducad. Contrary to what happened in relation to ACMV, the levels of resistanee to CBB were maintained as a result of previous selection of parents from Latín American and Atrican origin for resistance to thls dlsease. As shown in Figure 23.4, CBB incidenee was always higher at Ibadan until the 8th MAT (Jan. 1991). At the onset of the 1991 rainy season at Dnne, T abls 23.4. Comparison aeroS$ famili •• showing lTlOfe than 10% al !he individual. with few Or no ACMV symplom. during the perlad 01 max. dlllase pressurs al Ibadan. Parants Family Female Mala SM 1275 TMS 3001 (lITA) CM 7499 TM33001 (lITA) CM 2909-36 (2) CM 79656 M Col 638 (2) TM330001 (lITA) CM7965A TMS 30001 (lITA) M Col 638 (2) CM7966A M Ccl638 (2) TMS 30572 (liTA) SM 1276 TMS 30572 (lITA) CM 7558 M era 12 (1) TMS 30001 (lITA) $M 1179 TMS 30572 (lITA) CM7600S TMS 30001 (lITA) CM 2771·3 (2) CM 750QA TMS 30572 ~ITA) CM 2909-36 (2) CM 7446 CG489-34 (2) TMS 30001 (liTA) CM 7559 TMS 30001 (lIT Al M era 12 (1) CM 75008 TMS 30572 (lITA) CM 2909-36 (2) CM 6971 CM 5fJ1-37 (2) HMC-l Ovarall avg o! 140 famllies No. b._n par.n!h •••• rete .. 10 \h. ecz o! esen par.nt. 385 % lndlv. In 0 ....... 1-2 41.4 :29.0 25.0 24.4 23.6 21.0 17.4 17.2 16.6 16.6 13.4 12.8 10.4 10.2 3.9 % Plants Showing Symptoms 100 e IBAOAN _ ONNE 80 60 40 20 0'--"---- o 1 23456 7 a Months After lransplanting 10 Figure 23.4. Incidence of CSS in CBSsava seedlings growing at lbadan and Onne. 1990-91. the levels of cas ineidenee ¡ncreased dramatieally as a result of heavy rains and low night temp. As in the case of ACMV, CBB was insignificant at Kano during the 1990-91 growing season, whieh was to be expected given the elimatle eonditions typical of that lacation. As the maln cause for seedling death during the first 2 MAl was the severe attaek 01 CBa occurring in June 1990. The survival rate for eaeh family at 2 MAT gives an indication of the suseeptibility of the several familias evaluated for cee. Of the 31 families that showed a survival rate " 90% in relation to the total no. of seedlings transplanted (fable 23.5), 28 . are derived from erosses involvlng CIAT elite mataríais adapted to the aeíd-soils savannas of South Ameñca (ECZ 2), whare csa Is the major biotie eonstraint. Twelve families involving lITA elite materíals are also among the families with a high survival rate ,!Jnder severe CSS attaek. Table 23.5. Su,,",,'" rele DI .. Iacted famUles alter a severe attack 01 CSS al Ibadan. Parenta Family Femafe Male 'l(, Survival CM 70008 TMS 30IXIl (ITA) CM V72-3 (2) 100 CM 7859 TMS 30572 (ITA) M Pan 51 (2) 100 CM 7921 CM 24e6-5 (2) TMS 30572 (lITA) 100 CM 7968 M Col 2215 (1) TMS 30572 (lITA) 100 SM 1382 M Alg 13 (2-6) 100 SMI387 M Sra 715 (2-6) 100 SM 1399 CM 2909-36 (2-6) 100 SM 1401 M Sra 5 (U) 100 SM979 CM 2l!98-3 (2) 98 CM7W7 TM5 30572 OITA) CM 2952·1 (2) 95 SM 1231 CM V74-11 (2-3) 95 CM 7746 CM 3408-1 (1-4) SO 104-284 (1) 94 CM 7387 TMS 30572 ~ITA) CM 2177-:1 (2) 94 SMI225 CM 2452-5 (2..J) 93 SMI346 M Atg 13 (2-6) 93 CM7ge5A M 001638 (2) TMS 30IXIl (liTA) 92 SM909 CM 2087·101 (2) 92 SM 1147 CM 1999-5 (2) 92 SM 1177 SO 1)23.29 (2) 92 SM 1216 CG 165-7 (2-3) 92 SM 1219 CG 1450-4 (2-3) 92 SMI223 CM 2174-7 (2-3) 92 SM 1243 SG 104-284 (2..J) 92 SM 1245 SG 250-3 (2-3) 92 SM 1275 TM5 30572 OITA) 92 CM 75006 TMS 30572 QITA) CM 2909-36 (2) 90 CM7600A CM V72-3 (2) TMS 30IXIl OIT A) 90 CM7904A CM 507..J7 (2) TMS 30IXIl (lITA) 90 CM7904B TMS 30IXIl ~A) CM 507..J7 (2) 90 SM 1009 CM 2087.101 (2-5) 90 SM 1332 C0896-2 (2-3) 90 Overall a.g DI 116Iamill •• 79 No. _.on parenll¡osoa reler. 10 ECZ DI ......... parent. 23.2.3. CGM CGM was the pest that occurred at the highest level of attack at both locations, particularly during the dry season. Tabla 23.61ists tOO5e families with 10% or more of the indMduals within c/asses 1 and 2 of the evaluation scale and above, at Ibadan. Of th? 21 families Iisted, 9 are derived from parents adapted to the dry areas 01 Latin America (ECZ 1), where mites are the most serious casséwa arthropod pest. 23.3 Agronomic Characterlstlcs The agronomic characteristics measured at monthly ¡ntervals and al harvest are described in Table 23.1. With the information obtained at harvest, based on 538,474 and 311 387 Tabf& 23.6. Comparison aerolS families- showing more than 10% of the individuals wnh few or no symptOITl$ typical of a CGM attack during the period 01 max. pr&ssure at Ibadan. Parents Family FemaJe Mal. % Indiv. in CI .... es 1-2 CM 1968 M Col 2215 (1) TMS 30572 (liTA) 50.0 CM 6069 M era 12 (MQN) M Pan 51 (2-3) 50.0 CM 6815 CM 184~ (2) CM 2144-2 (2) 40.0 CM 6781 CM 2087-1 (2) CM 507-37 (2) 39.5 CM 7417 CG 5'-79 (MON) TMS 30001 ~IT Al 23.5 CMn45 CM 3408-1 (1-4) CM 3064-4 (2-3¡ 21.5 SM905 CM 996-6 (2) 20.0 SM 1150 CM 2563-5 (2) 20.0 CM 6871 CM 507-37 (2-3). HMC 1 (1-4) 17.0 CM 7221 CM 681-2 (1-4) CM 2174-7 (2·3) 16.5 CM 6822 CM 1851-4 (2) CM 2144-1 (2) 16.5 SM827 CM 966-2 (2) 16.5 SM 1009 CM 2087·101 (2·5) 15.5 3M 1141 CM 879-19 (2) 15.5 CM 7559 M era 12 (MONl TMS 30572 (lIT Al 15.5 CMn41 CM 3408-1 (1-4) CM 507-37 (2-3) 14.0 31.11363 SG 49$-19 (2-3) 14.0 CM 7460 TMS 30572 ~ITAl CM 523-7 (2) 12.5 CM 7976 TMS 30572 pITA) CM 2766-5 (2) 11.5 SM 1003 M Col 1522 (2·5) 19.5 CMn44 CM 3408-1 (1-4) CM 2n4-11 (2-3) 10.0 Overal! o.g 01 116 I.mili •• 5.5 No. betNeen parentheses refers 10 ECZ of eaeh par&llt; MON ~ CIAT cede 10r g.rmplasm resistant 10 CGM. seedUngs selected at Ibadan, Kano and Onne, resp., it was possible to draw examples of the typiesl seedlings salected at aaeh loestion basad on the means for eaeh charactaristie (Fig. 23.6). Seadlings harvested at Onne were taller and branched at a higher ht than those harvestad at Ibadan and Kano. An avg of 3 forking points was found in the seadlings selected at Ibadan and Onne, differing from thosa salected at Kano, with only 2 branehing levels. Avg no. of storage roots was 6 in Onne and Ibadan; 7 at Kano. Becausa af rains during the last 3 mo of the growth period in Kano, seedlings harvested at that loeation showed the highest leaf retention. A brief diseussion of eaeh eharactaristic measured at tha 3 sites follows. 23.3.1 Plant height, helght of flrst branch and branching levels Comparisons of the population's morphological charaeteristics take into consideration: .. The similarity of the populations introduced at Ibadan and Onne: 94% of the families ware common to the 2 locations_ 388 A % Selections 60 -11 i 'E:1 IBA DAN 60J m ONNE ,.j KANO 20 10 ( 100 B % Selections 60 50 100-149 150-1~9 200-249 Avg Plant Ht (cm) 250-299 '299 D IBADAN _ ONNE ( 50 50-99 100-149 160-199 200-249 300-349 > 349 Ht of 1st Branch Figure 23.5. Frequency dislrlbutlon of (Al plant lit and (B) ht of the flrst branch In cassava seedllngs seiected al 31ocations of Nlgerla, 1990-91. 389 .. A completely different population evaluated at Kano, obtained from parents adapted to the ecosystem represented by the Colombian North Coast (ECZ 1). As a result differences observed between Ibadan and Onna are strongly influenced by differences in the environmental conditions of the 2 locations. The frequency distributions shown in Rgures 23.5-23.6 reflect the differances in plant ht, ht at tirst branching and no. of branching levals/pl at harvast, resp. Plants grown at Kano were smaller than thosa evaluated at the other 2 locations. Avg plant ht at Kano was 145 cm, while those grown at Onne and Ibadan showed avg of 236 and 213 cm, resp. The higher frequencies of avg plant ht in the range of 100-140 cm at Kano contrast with the distribution observad at the other sites. % Selections 35~············ -------.-----------------.------------, t:l IBADAN 30 _ ONNE KANO 20 15 10 5 o o 2 3 5 Branching Levels/P I Figure 23.6. Frequency distríbution 01 no. 01 branchlng levels In cassava seedlings selected al 31ocatlons ot Nigeria. 1990-91. 390 Figure 23.7 shows the growth patterns of the seedlings planted at the 3 locations basad on monthly measurements of plant ht taken from the first 10 plants of each family. The values at the end of the cycle differ from those avg calculated from the selections (Fig. 23.5A) because not all the 10 plants measurad each month were selected at harvest. It is particularly interesting to note that the seedlings growing at Kano were able to grow (although at lower rates) even during the long dry season experienced at that Iocation, showing the adaptation of the populations to dry conditions. Avg values for ht at the first branch and no. of branching levels at Kano were 97 cm and 1.6 m, resp., compared with 141 cm and 3.5 m for Onne and 113 cm and 3.4 m at lbedan. In general plants harvested al Onne were taller than those growing at Ibadan and brS!'lched at a hlgher leve!. The no. of branching levels did not diffar at the 2 locations, both showing a high concn. of indMduals betwean 3 and 4 branching levals (Fig. 23.6). These differences can be partially explained by 2 factors: the belter distribution of rainfall at Onne during the growth perlad (Fig. 23.2), and the lowar pressure of pests at Onne. 80th factors can be considered as major components of the 2 environments. Plant Ht (cm) 200r - KANO -+- IBADAN 150 r -*- ONNE 100 , Months After Transplanting Figure 23.7. Avg plan! ht of cassava seedllngs growlng al 3 lecaUons of Nigerla on severa! occaslons during the growth perlod. 1990-91, 391 23.3.2 Lodging Figure 23.8A shows that at al! locations, plants were not severely affected by Iodging. The hlgher frequency of plants within classes 2 (intermediate lodging) and 3 (severe lodging) observed at Ibadan was due to strong winds during the raíny season. 23.3.3 Leaf Retention Leaf retention gives an indication of the plant's ability to keep functional photosynthetic area at the end of the growth cycle. Rgure 23.88 shows the frequency distribution of leat retention in the plants seJected at the 3 locations. As a result of the unusually high rainfalt that occurred during the last 3 mo of the growth cycle at Kano (Rg. 23.2), the seedlings se!ected there had a thicker canopy (57 cm) when compared with the data obtained at !badan (51 cm) and Onne (36 cm). These differences may also be due to the effects of pests and diseases. At Onne the 1991 rainy season started in March (before that in Ibadan), resulting in increased pressure from C8S and mosaic and lower leaf retention at harvest time. 23.3.4 Root Char'acteristics Root characteristics such as color of skin, flesh and cortex are not expected to change as a result of the environment. The data presented in Figures 23.9-23.10A show that at both Ibadan and Onne (where similar popuJations were evaluated), the majority 01 the selections had dark-skinned roots with white flesh and no or light pigmentation of the cortex. Selections from Kano, on the other hand, showed a tendency for having lighter skin, cream flesh and a high degree of pigmentation in the cortex. As a general rule, most 01 the selections showed none or few of the root constrictions that are detrimental tor the processing of cassava (Fig. 23.108) although there were a high no. of constrictions in plants evaluated at Kano. As the population introduced there Is different from those evaluated at Ibadan and Onne, it Is not possible to affirm whether the no. of constrictions was influenced by the environment or if it reflects the population evaluated at Kano. RY was measured by weighing roots from all selected plants at each site. Although not the sole factor, RY was also taken into consideration in the selections done at the 3 sites. Seedlings selected at Ibadan show.ed the highest avg root fresh wt (2.9 kg/pl), followed by the selections harvested at Kano (2.5 kg/pl) and Onne (2.2 k9/pl). The frequency distríbutions shown in Figure 23.11 A show that the majority of the inclividuals selected had avg root wts between 1 and 2 kg/pl although some outstanding values were recorded at the 3 locations, especially at Ibadan, where over 40% of the plants selected had ~ 3 kg roots/pl (Fig. 23.118). The lower RYs obtained at Onne, although under a situation of lower pressure from ACMV and C88 throughout the growth cycle, may be due to the 392 A B % Selections 100 60 40 20 o .L-__ _ 1 2 Lodglng El IBAOAN _ ONNE KANO % Selections 50~~~~-------------------------------¡ 40 30 20 10 o .20 20-39 40-59 60-79 80-99 Síze of Canopy (cm) D -O • 119 Figure 23.8. Frequency dlstrlbutlon 01 (A) IOOglng and (B) leal retentlon in cassava seedllngs grown ar 3 locations 01 Nlgeria, 1990-91. 393 A B % Seleclions 80{1 'o 70~ ~ i t= IBADAN ONNE KAN O .. ~~_.~ .. ~~~~ ._~~ ···~~~~···--I , I 60 11 ""."7, ..... UIl .. tCU.f, *_0\&"" allo ... 50j I ¡ i 4Q l i 30 i i 20 1 10 1 % Selectíons 80 60 40 20 O .JL.-"-_..r.l 1 2 3 Roo! Sk in Color I:l IBADAN _ ONNE O KANO 2 3 Roo! Flesh Color Figure 23.9. Frequency distribution of (A) rool skin and (B) rOO! flash color in cassava seedlings grown al 3 locations of Nígeria. 1990-91. 394 A % Seleclions 50 40 30 20 10 O E:I IBAOAN _ ONNE i 1 KANO ••• "" •••• u .... , 1.'., ••• 111'·.,,-00.". , ... ",. 1 2 3 Plgmenlalion 01 Root Corlex Figure 23.10. Frequency distribution of (A) pigmentation of roo! cortex and (B) root constriclions in cassava seedllngs grown at 3 locaIions of Nlgeria. 1990-91. 395 A % Selections 60~r-----------------------·--~------------------~ 50 40 30 ~ 20 '1 10 o • 1 1- 1.9 B % Selections 40] 35 ~ 30 ~ 10 5 o 2-3 4-5 2-2.9 3-3.9 4-4.9 Root Fresh Wt (kg/plant) 6-7 8-9 Roots/PI 5-5.9 » 5.9 e IBA DAN H ONNE o KANO 10-11 >11 Figure 23.11. Frequency distribution of (A) fresh roo wt and (8) no. of roots/pI in seedlings selected at 3 locations of Nigeria. 1990·91. 396 extremely vigorous vegetative organs observed in the plants selected at that location, which may cause a situation of competition for photosynthates between roots and tops. The avg no. of storage roots/pl in the selections made at Ibadan, Onne and Kano were 5.8, 5.9 and 6.9, resp. According to Figure 23.11 B, there was a tendency for a higher no. of plants with 4-7 roots both at Ibadan and Onne (as indicated, with 94% of familias common to the 2 sites); whereas in Kano a larger no. of plants with 8 or more roots were harvested. As the data do not refer to the no. of commercial roots but to the total no. of storage rocts/pl, this can be a result of new roots formed after the second rainy season in Kano. The results obtained thus far from this large-scale germplasm íntroduction support the feasíbility of such a program and suggest that a previous selection of parents according to agroecological adaptation is an adequate strategy for transferring genes from Latin America to Africa. The population evaluated at Kano, formed by seedlings resulting from crosses using cassava geimplasm adapted to.dry areas 6f Latín America, showed a remarkable ability to withstand a dry season of 6 mo established after 2 MAT. A total of 474 seedlings were cloned and plantad at Kano and Zaria (Nigería) for further evaluation under dry conditions. The use of liTA" elite clones as a source of resistan ce to ACMV has preved to be a good strategy as the progenies resulting from crosses involving liTA ACMV-resistant clones proved to have a higher level of resistance. Together with CIAT elite clones adapted to the aeíd-soil savannas of South America, liTA clones also proved to be effective in producing families with a higher degree of resistanee to CBB. 397 24. COLLABORATIVE STUDY OF CASSAVA IN AFRICA COSCA is an interinstitutional effort to provide basic information about cassava in Africa, to ¡ncrease the relevance and impact of research related to the crop, and to help increase income and foOO security tor people in Africa. The study began as a joínt collaborative project between CIAT and lITA (charged with the project's execution). A number of other institutions have collaborated to date, including the NAI (ODA, UK). The study involves the collection of data which will be utilized to characterize: .. the structure of cassava-based cropping systems in Africa .. the nature and distribution of various cassava processing techniques .. the markeling systems tor cassava ... present and future demand tor cassava in rural and urban areas .. the relationships between cassava consumption and nutrition Whenthe study began in 1989, it was originally to be conducted in 6 countries: Ghana, Ivory Coast, Nigeria, Tanzania, Uganda and Zaire. In 1990-91 it was extended to Liberia and Sierra Leone, Burundi, Cameroon, Congo, Kenya, Malawi and Zambia. Data are collected at two levels. In the first phase, generalized. mostly qualitative informanon was gatherad from group interviews at the village-Ievel. The second and third phases, which are currently being implementad, will generate informanon speciflC to the household, farm or processing unit. The CIAT ASU elaborated a sampling frame and survey method for the project's first phase. Unit members have subsequently participated in training exerclses tor phase one, and in some of the data analysis. A description of CIAT's contribution and a selection of results from this phase are presented below. 24.1 Slte Selectlon tor Phase One As the goal of the flrst phase was to characterize many aspects of cassava production and use across large geographic areas, a spatial sampling frame was required to ensure a representative picture tor every country. At the Third Planning Meeting for the project (Sept. 1988), participants from the countries ¡nvolved and members of the SC selected 4 factors with which to subdivide cassava-growing areas to be surveyed: Distribution of the crop, agroecological conditions, density of human population and accessibility. The COSCA project began at a time of important methodological developments in the ASU. The Unít had begun to implement its first microcomputer system tor handling geographic information, both for digitizing paper maps and tor simple geographic 399 analyses. These new packages were quickly ínterfaced with CIArs mainframe climatic database and mapping facilities. Powerful new techniques tor managing large datasets therefore influenced the methodology tor site selection in COSCA, by facilitating the combination ofthe environmental and socioeconomic variables required to build a spatial sampling Irame. Thesa same techniques have provided the basis tor rapid mapping of the processed survey data using microcomputers. The ASU had developed a spatial survey method based on viflage-Ievel questionnaires to generate information for stratificetlons and tor research orientation. The method was directly compatible with the type of grid-based geographic informalion system required for Ihe COSCA Phase·One sampling frame. By geo-reterencing survey sites, results 01 Ihe Phase-One questionnaire cen be mapped for visual interpretalion or overlayed onto other maps for statistical comparisons. 24.1.1 Construction of the sampling frame The ASU produced a lT1ap of cassava distribution tor Africe (Fig. 24.1), using the mas! recen! census statistics and other information. This map served to identify the main cassava-producing areas of each 01 the project countries and to exclude areas where cassava was not grown. The climatic classification devisad for cassava in Latin America was used to subdivide the range of agroecological conditions in which cassava was found (see Cassava Program Annual Report, 1986, pp. 43-52). Climatic classes were mapped via a set of interpolated grid files created lrom CIArs mean monthly meteorological database. Human population density was considered to be the most important socioeconomic factor to take into account tor stratification in the Phase-One survey. Population density has a strong influence upon the intensity of land use, Ihough not in isolation. Secondary-Ievel administrative units were digitized for each of the 6 countries, and their respective areas calculated. The US Census Bureau provided population totals tor these units from recent censuses which were then projected forward to 1990 using avg annual population growth rates tor each country. Population densities were then calculated and mapped on a microcomputer. The final factor used to stratify the countries in !he survey was accessibility, which was considered as likely to be an important determinant of commercialization and migration. AII-weather roads, railways and navigable rivers were digitized. Places within 20 km of these were considered to have easy access; those further away, difficult. This crude measure proved more reliable, once tieldwork began, in countries with a well-developed network (e.g., Nigeria and Ivory Coast) than in countries with a sparse network (e.g., Tanzania). 400 I L--\ j \ • \ \ 2O"S • \ o • o o o S.OCO 10.000 SO,OOO 100,000 5<)ttOOO Ha. I 20~W rrw 2O"E 40'E Figure 24.1. Cassava dislribution in Africa. 1980. Maps of cassava distribution, infrastructure and administrative units were turned into grid cell maps (rasterized) to correspond exactly to the grid created for the climatic map (cells had dimensions 01 12 min lat. by 12 min long.). For any one cell, therefore, there existed a hectarage of cassava on one computer map, a population density on another, a c1imatic class on another and an accessibility factor on a fourth map. Population density and accessibility were simplffied into 2 classes each (Iowjhigh, poorjgood, rasp.). The elimate, population and access maps were !hen superimposed to create a series of homogeneous regions for éach country. Cassava area was !hen totaled for each region, and maps of these were produced from which survey locations could then be selected. 401 24.1.2 Identification of survey regions and villages The site-selection procedure began at the first COSCA training meeting, he Id in April 1989. Project management and national coardinators usad the informatian generated about the relative distribution af cassava in the different geographic areas 01 each country to identify the areas that they would survey. Throughout the first week 01 the meeting, national survey teams refined estimates 01 the no. of sites they could survey in the time available, based on budgeting and manpower limitations. seasonal weather conditions and the distances to be cavered. The total no. of sites was then apportioned among the different regions to be surveyed. A total of 294 sites were selected for each of the 6 coul']tries as 101l0ws: Ghana, 30; Ivory Coast, 40; Nigeria, 65; Tanzania, 42; Uganda, 37; and Zaire, 80. The 12-min grid used to construct the maps of homogeneous regions provided a frame from which a stratified random selection of general locations was made for administering Pha;se-One questionnaires. The method used ensured an even spread of survey points across each region. Each grid cell selected then had to be transferred to larger scale maps of the respective survey region from which individual villages could be identified. Although survey teams were able to practice this during the training week, they selected sites on retum to their respective countries .. Finally, various methods for selecting a survey viUage at random within each grid cel! were discussed. 24.2 Data Analysis Fieldwork for Phase One began in June 1989 and continued in some cases' until Feb. 1990. By March 1990 most of the questionnaires had been coded and entered on a microcomputer at liTA. A second training meeting was held that month to introduce the national survey teams to techniques of database management, statistical analysis and mapping of the survey data. Participants were required to present a preliminary report on the findings of the study in their respective countries, including a set of maps to represent geographically significant aspects of these results. The ASU provided relevant environmental and socioeconomic computer maps to test some of the hypotheses that Phase One was expected to answer and also helped participants map the data they chose to present. During the remainder of 1990, the Unit cooperated in data cleaning and analysis. Analysis facused on relationships between aspects of cassava production and environmental and socioeconomic conditions. Some of the results are outlined here. 24.2.1 Distribution of biHer and sweet varletles of cassava Little was known previously about the distribution of bitter and sweet var. of cassava in Africa, and the COSCA survey has provided some detailed information about this for the 402 countrias covarad. To distinguish bittar trom sweet var., thosa that can be eaten raw without causing intoxication were defined as sweet. In 103 of the villages visited, both bitter and swaet varo wera cultivated. At 38 sites, only bitter varo wara recordad--Iargaly in Ivory Coast and southarn Nigeria, with a few sites in Zaire. At 83 sites, only sweet varo were recordad: a11 places visited in Ghana, most in Uganda and northarn Nigeria, and a few in Zaire and Tanzania. For tha data set as a whole, the total no. of bitter var. at a sita was found to be independent of dry-season length: whereas the total no. of sweat varo recordad was slightly greater in humid climates (0-3 dry mol. Factors that were positively relatad to the no. of sweat varo grown were total area in cassava and elevation. Sweet varo were found to be less common at sites with seasonal cUmates. The total no. of bitter varo was negatívely correlated with altitude. . With respect to the type of variety that occupíes the greatest area in each village surveyed, the relative area inbitter and sweat varo at a site was found to be independent of length of dry season; however, t~is masked different situations within each country. In Ivory Coast and Ghana, the predbminance of sweet varo was independent of climate. In Uganda, where sweet varo predominated, all sites had humid climates. In Tanzania and zaire bitter varo predominated, but climates were mostly seasonally dry and humid, resp.; and it was not possible to detect any interaction due to the small sample size in these countries. In the case of Nigeria, cUmate could not be discounted as a determíning factor, but it was not clear why it should only have an influence there, and not in other parts of West Atrica, for example. A fuller understanding of interactions between varietal type and environmental conditions required estimates of the actual area sown to sweet and bitter varo (data to be gatherad in household-Ievel surveys in the proJect's second phase). Sites with > 2 bitter varo tended to have acid solls; very few sites (5 of 58) with unrestricted soils had no sweet varo The probability of sites having no sweet varo was significantly higher for those with acid soils than for sites whose soils presented no restrictions for cassava. The overa" impression given by these data was one of diversity and complexity in the choice of cassava var. and the relative importance of the crop. 24.2.2 In11uence 01 blophyslcal and human environments upon age 01 cassava at harvest For many rural people flexibifity of harvest perlad is one 01 the most important qualities of cassava. The COSCA study identified a wide range of harvest periods (3 mo to ~ 3 yr) in the countries surveyed. 403 Information from the Africen geographic database permitted comparisons of the age-at- harvest data with environmental variables and population density. Age at harvest was positively correlated with the total area of cassava around each survey village (within approx. 10 km). Most of the late-harvested varo (24 mo or longer) were feund in areas of low population density «50 persons/km'). 24.2.3 Abandonment of cassava varietles The reasons for abandoning up to 4 cassava var. were elicited in the questionnalre. Cassava varo were abandoned most trequently because their growth cycle was too long. In humid climates, competition from weeds was the second most common reason. In dry areas, processing problems and pests were cited as trequently as weeds, as secondary reasons for abandonment Whan examinad in ralation to soils, . weed competition was citad as the majar reason for abandoning varieties at sites with acíd soils. Similarly, this reasan was citad most frequently in areas of low population density, probably where labor was scarcest. In aieas of high population density. by contrast, long growing cycle was most commonly cited, pr~sumably because land was scarcer. Having examined data on cassava varieties in the 6 COSCA countries, the most salient point appears to be the significant variation that exists in all the findings. There are faw elear, strong ralationships with environmental or population density that cut across all countries studied initially. This indicates a very complex interaction offactors determining cassava's status, uses and importance in different parts of different countries. It emphasizes the need for further studies to understand local variation within particular cultural, economic and environmental contexts, which can then be brought together to increase overall understanding of cassava in Atrica. Analysis of the Phase-One results from countries new to the study, in conjunction with the data from the first 6 countries, may identify relationships with broad geographic factors more clearly. The second phase will provide the details necessary for a deeper understanding 01 the local situation to Noradd depth to the results from the first phase. 24.3 Further Analysis of the African Geographlc Information System for Cassava Much of the work undertaken fer the COSCA project by the ASU was not limited to the 6 original countries. Cassava distribution, climate classification. administrative units and infrastructure were originally mapped far the whole continent. Together with digital soil and elevation maps acquired after the sampling trame had been designed for COSCA, relationships between the distribution of cassava, population and environment were explored tor all of Africa. The proportion of land area devoted to cassava, length of dry season (from the climatie classification). sojl restriction type, and population density were calculated. A new map of unique polygons was created in which the above factors were all homogeneous. The population and area in cassava was then calculated tor each polygon; and a regression 404 modal was constructed with area in cassava as the dependent variable, population density as the independent variable, and the two environmantal variables as factors of the independent variable. The model shows that the proportion of land devoted to cassava ¡ncreases with population density. The relationship is describad by a quadratic curve. The modal has served as a check for the area data used in the original cassava distríbution map (Fig. 24.1). It was used to extrapolate cassava distribution for the year 2000, using population projections. Spatial residuals--areas where the model does not fit well--have been ídentified using the model. These are eoncentrated around the Great Lakes, where the original distribution map showed the highest spatial concentrations of cassava. This finding suggests either that cassava area is overestim¡;¡ted, yields are significantly lower than elsewhere (necessitating a larger area to support a given population), or there are strong cultural reasons tor higher cassava production in this region. Other areas of unexpectedly high cassava production are southeastem Tanzania, NE Mozambique and Bas Zaire. Possible explanations in these cases are lower productivíty or cultural factors in the first case,. the war in Mozambique and the large market demand of Kinshasa. Significant differences were found in the amount of cassava grown between elimate' types. Humid elimates had the largest percentage areas, followed by seasonal climates. Dry cllmates had the lowest areas in cassava. These differences were largely a function of population distríbution and the agricultural systems upon which people depend for livelihood, in part determined by cllmate. Cassava introduced to Africa by the Portuguese was more adapted to humid and seasonal climates, and only recently has CIAT begun to take material from drier areas of NE Brazil to test in similar climates in Africa. The findings from this study have implications for the geographic focus of cassava research in Afriea. Areas of highest population density, around the Great Lakes and in SW Nigeria, for example, are predicted to see cassava area increase considerably this decade. This expansion will probably accompany the intensification of the areas' farming systems. In the Sahel belt and much of the Zaire bas!n, expansion of cassava over much broader areas is also likely. This may take place as an extensification of farming systems. CIAT will publish an atlas of cassava in Africa, featuring maps that made up the geographic database, a study of the diffusion of cassava in Africa, and a series of case studies of cassava-based farming systems in 3 producer countries. 405 25. AOOPTION ANO IMPACT STUOIES The role of socioeconomics in the Cassava Program is to collaborate in the development 01 cost-effective appropriate technology components and their efficient diffusion in arder to maximize impact ón preselected target areas and audiences. As such. its activities inelude identification and analysis 01 problems and opportunities, experimental design and data analysis, monitoring the release and adoption of technology components. impacto and prioríty setting. During the 1987-89 period the Economics Section allocated most 01 its resources to Jntegrated Cassava Projects R&D and monitoring activities. The Section collaborated íntensívely with the Utilization Section in alternative product and market assessment and validation in Colombia, Ecuador and Paraguay. In addition, socioeconomic ínputs have been given in the M&E 01 pre-production plots (PPPs), in collaboration with the Agronomy and Breeding sectíons. The majoríty of these studies were of an ex-ante nature. Some emphasis on ex-post adoptíon and impact was given to project monitoring. Especially in the pilot and semicommercial phases of the Integrated Cassava Projects, monitoring activities have proven to be crucial to ensure the feedback of processing plant performance data in arder to fine-tune short-term objectives. Monitoring also serves to check longer term objectives related to the distribution of benefits and overall adoption. Given a lack of quantitative information on past and current performance of cassava technology and the urgency to become more resource accountable, the E(!;()nomics Section reassessed its research priorities at the beginning of 1990. Thus in 1990-91 most of the research20.· has focused on ex-post technology adoption and impact assessment. These studies cover varietal adoption, production management and cassava processing technologies, both in Asia and Latin America. In addition, the studies have both a micro (Iarm household) and macro (aggregate) locus. Although several studies are still in progress, this report highlights some of the exciting· findings, showing a very significant adoption and impact 01 different cassava technologies in Colombia. 25.1 Integrated Cassava Projects The underlying philosophy behind the Integrated Cassava Projects first developed in Colombia was that declining traditional cassava markets did not offar incentives for cassava farmers to adopt technologies to increase production and that therefore the introduction of dried cassava chips could broaden demand and create adoption incentives at the farm leve!. 20.1 For a partiallisting ni the economics research issues, sea the Cassava Program Annual Report 1990. 407 After discussing the diffusion of the Integrated Cassava Projeet model, the relative performance of cassava processing plants is analyzed. Then the way in whieh on-farm cassava eonsumption and sales have ehanged as a result of the broader market is studied. This is followed by a seetion on the adoption of cassava produetion teehnologies, concluding by showing how Cassava Integrated Projeets have served as a vehicle for technology diffusion. 25.1.1 Adoption of drying technology In the previous ehapters on regional collaboration some information was presented on the rapid diffusion and adoption of cassava drying plants in Latin America (see also the Cassava Program Annual Report, 1990). In this seetion ít will be shown t'lat besides a rapid adoption in time and locatión. there has been a very dynamic adoptlon by different user groups. Also, the basic "eoop model" has been adapted by the commercial seetor to serve different objeetives and market struetures. Figure 25.1A shows the total no. of eassava processing plants in Latin America. The acceleration of plant adoption has been very elear during the last 2 yr. To a major extent this has been caused by the rapid adoptíon of coops by Brazil, and the acceleration of eommereial plants in Colombia (Fig. 25.1 B). The data for 1991 are estimates, however, as it has no longer been possible to keep an aeeurate eount through monitoring actívities because of the fast, widespread and diverse types of cassava drying adoption. During the initial phase of the project in Colombia, the eassava drying coops sold the dried chips to several large animal feed manufaeturers near large urban centers (Le., , Barranquilla, Bucaramanga and Medellln) and the North Coast of Colombia. During the last couple of years, however, several changes have been observed. . • The market for dried chips has broadened towards central and southern Colombia, including large feed companies in Bogotá, Buga and Cali. It is estimated that the majority ofthe chips for these latter markets are being supplied by plants on the North Coast. However, the share supplied by plants and prívate chippers in the southern part of the country has be en increasing significantly. This is another indication that the eassava processing technology is spreading into new areas . • While the buyers of chips were largely animal feed factories, the current user group is mueh less homogenous. There is a strong demand from (a) swine, broilers and egg producers, who mix their own feed rations; and (b) cattle operatians that need an energy source (on-farm) to reduce animal wt losses during the dry seasan, which coincides with the cassava proeessing season. In addition, a growing demand has become evident trom cassava starch producers, who have started buying cassava chips in addition lo fresh rools. 408 No. 01 Plants 300 250 200 150 100 50 o~~~~~~~~~~ 200 150 100 50 81-82 82-83 83-84 84-85 85-86 86-87 87-88 88-89 89-90 90-91 _ PARAGUAY _ BRAZIL-Ceara No, 01 Plants _ PANAMA e COLOMBIA ¡jigii! EOUADOR o~~~~~-L~~~~~~~~~ 81-82 82-8383-8484-8585-8686-8787-8888-8989-90 90-91 e Cooperatlves _ Private Figure 25.1. Adoption of cassava drylng plants in (A) Latin Amenea aOO (B) Colombia, 1981-91. 409 • The initial cassava processíng "model" as started in Sucre in 1982 was based on a small-scale cassava farmer coop concept. However, during tha last couple of years this modal has been modified because of different commarcial objactivas and markets. In 1990 only an estimated 60% of dry chips (sold to factories only) .... , were prOduced according to the basic model (Fig. 25.2); the rest was produced by a heterogenous group of processors based on a large variety of "models." For 1991 it is estimated that the Jatter group has surpassed production of the formar. Currently the following cassava drying "models' have been identified: • A small-scale cassava tarmar coop, with an avg of 20 (legal) mambers • A coop as abOye but with 20Q400 members • An association with 2-4 members who on the avg cultivate more land; the drying plant is on ona of the members' farms (typical for Santander) AutonomOU$ (manual. mechanlzed) proce88lng 24% Farmer coop.eyaUvea ---__ .~ 60% Semlprlvate planta Prlvat. planta 8% Figure 25.2. Adoptlon of cassava drylng lechnology In Colombia by user groups, 1990. (Total feed Industry usage, 16,500 lo). 25.2 Sales to feed faetones is the onJy reliable information that currently exists. It Is estimated that these sales comprlse "5Q..75% 01 tota' chip production. 410 .. 80th large and small prívate commercial plants. owned by one or more persons. purchasing all roots from local farmers .. Private entrepreneurs. who rent tha floor and chipper by unit of wt (or time) to local cassava farmers .. Large-scale drying plant, vertically integrated with an animal feed company. which is the owner and manager .. On-farm smaU and medium drying plants as part of a cattle operation .. Cassava starch processors (Cauca) who. when starch prices are low. produce dried cassava chips .. Individual small-scale cassava farmers, who manually chip and dry (on any kind of surface) cassava trom secondary quality roots, and/or when the cassava fresh price is too low, or becausa thay are isolated tram the fresh cassava market • Initial dry chip marketing was basically direct negotiating between chip-producing plants (or its marketing association) and teed factories. Given the increasing diversity of producers and "consumers," volumes traded and tha geographic dispersion of damand and supply zonas, intermediarias have entered the marketing channels, introducing one or more pricing points. Among other things. this has made it possibte far isolated small-scale, sporadic and low-vol. cassav8 chippers to sell the!r product. The introduction of intermediation has naturally increasad marketing margins. It is avidant that this cassava processing technology has been rapidly and widely adopted, in time and space, for different uses and user groups. Strong commercial Interest and adoption ís probably a fair indicator of the potential benef1ts to be gained from the technology. Moreover, the adaptation of the original model into a wide array of "applied models" increases the sustainability of tha technology, cetaris paribus. Although basic data on adoption has been shown. there still exists a lack of quantified impact data. both at the farm and the aggregate levels. A study has been started to analyze this impact. Given the evolution of the basic technology concept into various models, there is a need to assess the differences in structure, management, efficiency and especially the distribution of benefrts "across" these models. A study to assess this will be started next year. The sustainabiJity of the cassava drying technology depends to a large extent on the sustained demand for chips (and other cassava-based products), which in turn depend on relative price competitiveness. Hence a study was started to assess too Mure impact of ~ovarnmant policies such as the libaration of imports on the potantial of cassava vis-é- 411 vis competíng products. This is a kind of monitoring that is needed for timely decision- making regarding cassava product and market strategies. 25.1.2 Relative performance 01 drylng plants As part of long-term monitoring, the drying plants on the North Coast of Colombia were assessed for their individual performance. The objectiva was to idantify and analyze the most important factors that contribute to the success (or fanure) of a cassava drying plant. As such, in 1991, 29 cassava coops were surveyed. Critería for success were: dried cassava productíon, profits, capacity utilization, and proflts per ton of dried cassava. The general conclusion is that the prevailing factor .for a plant's success is adaquate root supply commensurate with its floor slze and sufficient working capital. With respect to output efficiency, it was concluded that drying plants situated in the savannas were more successful than plants in other areas (cassava is a more predominant crop in the savannas because of harsh agroclimatic conditions). Sorne correlatíon (positive) existed between dried cassava productíon and profits. Assured root supply is more important than processing efficiency, being directly relatad to tha importance of the plant's location. The degree of utilized capacity has sorne influence on success. It was found that mora than 60% of tha successful plants in the savanna utilizad at least 50% of thair capacity. When looking at profits par ton of dried cassava, the issue of plant siza becama apparant. It was found that a large drying floor and large productíon )Vare not necessaríly a key for success. Although in theory largar plants should enJoy some economias of scala, this was being offset by the major problem of coordinating a sufficient supply of fresh roots. Largar plants haya had problems with limitad transportation (no. of tractors) and hauling larger supplies over largar distances. When it was established that root supply and working capital ware the critical factors for plant success, both were analyzad further, across individual plants. Factors that directly influenced fresh rool supply wara: • Competitiveness of fresh root purchase price. During the processing season, drying plants compete for supplies with the demand for fresh roots from urban centers, and to a lesser extent, with the roots for starch processing. It was found that during the main harvest, root prices allowed healthy competition. During the dry season, the price offered al the plants justified the transport of roots from > 250 km. • Farm~size distribution near the drying plant. It was found that two types of farm-size distribution had a positive effect on assured root supplies: a large no. of small (1-2 ha) farms within a 1-5 km distance from the plant and a smaller (15-25) no. of medium (5- 12 ha)-sized farms within a 5-10 km radius. It was also demonstrated that a concentration of production in the vicinity of the plant ls of importance. 412 • Tractor availability is an important factor for land preparation as well as for the transportation of roots from farms to tha plant. Under existing conditions, farmers need tractors in arder to expand planted area. • Collective cassava fields. A drying plant with coUectiva cassava plantings increases its independence from competing markets. As such. it increases its self-produeed supplies and lowers the risk of undersupply. In addition. cooperative cassava plots are a way for farmers to inerease production in the absence of titled land far individuals. Especially in Suere, with the assistance of low-cost financing by local institutions. 25-60 ha eooperative plots are common. The majority of thase plots produce eassava with improved cassava teehnologies. Besides an assured root supply. suffieient working capital has proven to be a majar critical factor in the sueeess of a plant. When plants have insufficient funds, the purchasing of roots is hampered. which directly influences production effieiency. The principal issues behind this are: (a) the absence ot savings from the previous processing seasen; (b) delays in the release of loans to purehase roots; and (e) delays in payments for dried chip sales by feed factories or marketing organizations. The information that has been generated with this study Is of great value in the deeision- making process for expanding old plants and establishlng new ones. The feedback of monltoring information serves to increase the successful performance and continued adoption and impact of cassava drying plants. 25.1.3 Productlon technology adoptlon The formula 01 integrating cassava utilization. marketing and production aspects in this researeh has offered the opportunity to use the drying plants (tarmer associations) as a vehicle for developing, testing and diffusing improved cassava production and management technologies. One methodology for accomplishing this is with the PPPs, where the best available production recommendations are tested and valldated on-farm on commercial-size plots in areas where drying plants have been established (see Chapo 9). Another method has been to involva drying plants directly In the multiplicatlon and diffusion of impraved cassava (and maize) varieties. and in stake selection and treatment methodalogies (see Chapo 18). ICA, the Colombian natianal programo has for many years been developing and diffusing technology components /ncluding several improved varieties through its Technology Transfer O/visiono As a result. a wide array of cassava technologies have been adopted throughout Colombia, but especially in the predominant cassava-producing states of the Atlantic Coast. It has becoma impossible to distingu/sh one specific teehnology. in a defined area, introduced in a speeific year, from othars in other areas and years; rather there is a mixture of different technologies, with diffarant levels of diffusian, adoption and impacto 413 In assessing the adoption 01 cassava technology on the North Coast, the first step was to define obJectiv~s and specific hypotheses to be tested. The obJeetive of the study was to analyze the adoption of ímproved cassava produetion technologies and their effeet on the farm production, marketing and utilization system, given increased cassava demand after introducing the drying plants. Figure 25.3 presents a simplified scheme of how production and processing technologies and their adoption and impact are interrelated. At the adoption level there are two technologies: dried cassava and the cassava production package. The newly created demand will influence on-farm cassava consumption and sales. as well as stimulate cassava production at the farm level in two ways. In the short term, the farmer Is able to reaet by increasíng cassava area in the production system, using traditiOllal technology. In the intermed¡ate term, the tarmer will attempt te increase productivity, which ¡ncreases the demand for improved cassava production technology. In the intermediate to long runo both productivity and area expansion will increase aggregate supply. The following hypotheses were tested: " Increased cassava demand creates demand for improved produetion technology at the farm level. .. In the short term, farmers ínerease planted cassava area, deerease I'¡¡II0w perlod and area, and decrease nonharvested cassava area. .. In the Intermediate termo the adoption of technologies will increase cassava produetivity. .. Incraased expected retums on cassava and possibilities of obtainilig cOoperative credit will intensify cassava productiOll through increased input and tractor I,lsage. .. In areas with cassava drying plants, the rate and extent 01 production technology adoption is greater than in areas without drying plants. As can be seen from the hypotheses, the current study does not IImit itself to adoption only, but ineludes certain on-1arm production, utilization and marketing systems impacto The second step in organizing this study fer analyzing adoption 01 these technologies was to make an inventory of (a) baseline studies, (b) released technologies, (e) time-series data on cassava planted area and produetion, and (d) primary and secondary data on adoption levels and rates. In the case of the Atlantic Coast, it was found that many studíes and surveys had been conducted on cassava production, consumption and 414 .¡.. ..... (J1 e o ~ "C «: 'E_ ~s E~ '-I!.! ECI .E E .... . - .I! Oc .... 0 ~- Production tecnology package .ssava drylng technotogy New market prlce Iloor i Production increase I " • lA Income A Utilization Á Sales (1) short run; (2) interrnediale termo Á labor Á Household expendilures Á Diet Animal leed lactorles L Figure 25.3. Flow dlagram 01 the adoptlon and Impact of CSS$8va technologles regatdlng Integrated Ca_va Projects. Colombia. 1991. marketing aspects since 1974. However, a representative baselíne study did not existo Given this lack of essential data, the assumption was made that possibly a traditional cassava-growing area that had not experienced any (cassava) technological influance ovar the last decade, could sarve as an approximation of the "pre-technology influence" situanon of the beginning 1980s. In order to identify these areas, an inventory was made for the Coastal states of all cassava studies, experiments, presence and activities of ICA, CIAT. DRI, Caja Agraria, Secretariat of Agriculture, etc. over the last 10 yr. This information was analyzed and checked during a Rapid Rural Appraisal. Tha following "Ievals of technology influence" were classified (by municlpality): Level1: Municipalities with cassava drying plants, startad before 1988, with good instítutional presence, and that have had farmers participating in experiments, studies, etc. Level 2: Munícipalities with cassava drying plants buílt after 1988, with institutional presence through cassava drying activities Level 3: Municipalities wíthout drying plants and with low or nonexisting institutional presence With this prestratification, it is assumed that Level 1 has enjoyed the largest technology influence and that Level 3 has experienced the least. ConsequentJy, the latter will be assumed to approximate traditional cassava farming, and function as the baseline area Cfr O}, to which technology adoption data from Level 1 are compared. Level 2 is assumed to be an intermediate area that has received some direet technology ¡nfluence. This ex-ante classification and 1985 census data on cassava area by municipality formed the basa for a representative sample frame and size selection, covering 7 states in Northern Colombia. A farm household survey was developad and tested. Surveying of tha first 5 states took place in July-Aug. 1991. The prefiminary results .... that follow are from the states of Sucre, Bollvar and Córdoba only (N=299). 25.1.3.1 Qn-farm consumption and markating system change. The newly created demand for cassava dried chips will have an immediate effeet on on-farm cessava consumption and sales. Table 25.1 shows that farmers consumad 22% of their cassava produetion on-farm, mostly for human consumption. While on-farm cassava consumption by farm animals is virtually the same across levels. Level 1 typically retalns less cassava for on-farm household consumption and sells more to the market than Level 3. In other words, cassava farmers in high technology-influence areas have becoma more market oriented . ... 3 ft mus! be noted lha! as data Ior only 3 stat •• are avaiJoble. lh. ,.sullS havo boon adjust8d accmdingly. 416 Table 25.1. Comparison 01 cassava on-Iarm consumptlon and sal •• in .elactad state. 01 Colombia by leve! DI lechnology Influenc8, 1991. Technology Influence Level Avg 2 3 -- % Tolal Produotlon -- Oo-farm consumption: .Iannily 16 12 19 16 • animals 6 5 6 1 SublOlal 22 17 25 23 Sales lo: - Itesh market 55 42 67 56 • drying plants 22 40 8 19 .. other 1 2 Sublotal 78 83 75 71 Total Production 100 100 100 100 ~: Technoloey Adoption Survey, ClAT Cassava Economics. 1991. What ie even more striking is that currently 22% of cassava production is sold to drying plants; the remainder is sold to the freah consumption market. Given the criteria for selecting the different levels used in this study, it is consistent to observe that Level 1 farmers seU twice as mueh cassava to drying plants as Level 3 farmers. Subsequently, the latter SUU seU relatively more cassava to the fresh market. Given the strong influence of the drying technology and its demand across technology levels, ít would be useful to compare the few existing data over time. Table 25. '.l compares cassava consumption and sales data between 1982 and 1991. The differences are significant and consistent wíth the analysis aeross levels. It shows that the share of cassava production consumed on the farm (as pereent of total production) has been halved during this periodo While the production share sold to the fresh market in Córdoba and Bollvar virtuaJly remained the same, Sucre experienced a majar decline in this market, from 62% in 1982 to 35% in 1991. The majar and most obvious explanation for this ís the high concentration of drying plants in this state, which is also the reason tor 47% of the production being sold to drying plants there. While Córdoba shows a respectable 24%, faímers in Sollvar seJl only 2% tor chip production. Compared to 9 yr ago, this trend is evídence of the adoption of cassava drying technology. In addition, effects of this adoption have not been restricted to areas with a relatively higher level of technology influence and institutional presence, but also to areas with fewer or no processing plants or institutional presence. 417 Tabla 25.2. Comparlson el ""osava yield, eonsumption and sal.s In seleoled '_'01 Colombia, 1982 and 1991. Su",a C6rd0ba BoIl"", On-farm cassava consumpllon' ('X.lOtaI producüon) 1982 study' 34 38 :;r 1991 study" 17 18 213 Sala to Ir.oh mark.t ('X. total produollon) 1982 study' 62 53 60 1991 study' 35 " 69 Sales lo dOOng planto ('X. total producüon) 1982 studyl O O O 1991 study3 47 24 2 t Janssen, W., 1982. "Producción. m ...... d.o y al potencial Indus1rlal de la yuca an los dep_mentoo de AIlá_, Bollvar. Suero y Córdoba. CIAT. mlmaograph. 2 This ¡neludes both human and animal ensava oonsumption. 3 Prellminary dale, Cassava Adoption Survey, ClAT Cassava Economics, 1991. 25.1.3.2 Technoloay adoption rata and area. During the last decade a wide array of improved cassava production technology components have been introduced, either officially or unofficially through spontaneous farmer adoption (rabie 25.3). Tablé 25.3. Varlous cassava impr of Level 1 farmers are using tractors. as could be expected from this reasoning. Table 25.5. Adoplion 01 cassava productíon lecnnologie. in saleoled slalo. of Colombia. by level 01 .echnology inlluence. 1991. Technology Influence levol 1 2 3 Avg N·OO N-lOO N-l03 -- '1(, Respolldonlll- TractO( usa in land preperatlon 41 82 20 15 SI.k. trealmenl 10 19 5 5 Slak. storage 71 81 61 72 Herbicida use 53 68 50 39 ""',e.sed planUng densily 60 85 68 46 Technleal assistanco 40 61 45 12 ~: Preliminary dala. Ces .. v. Tochnology Adoption SuN.Y. CIAT Casoava Eoonomics. 1991, 420 The aforementioned technology components are all production focused. Further evidence of the adoption of these technology components is the change in yield and, of course, the change in cassava net revenues. Data, by level, showan overall cassava yield avg of 12.0 tlha (for cassava/maize intercrop). Given the fact that significant technology adoption has also taken place in Level 3, this will not serve as an approximation of a traditional (absolute) base. Hence, the data were analyzed by state and compared with results from a 1982 survey ..... As can be seen in Table 25.6, since 1982 cassava yields have increased by 52%, 56% and 76%, rasp., in Bolrvar, Sucre and Córdoba. In addition, comparing current yield averages with the national statistics (Table 25.7) does not show any significant differences. Nevertheless, the yield results of the studíes need to be interpreted with caution as they do not capture any fluctuations due to climatological factors. 25.1.3.3 Qs\ssava area increase in the short runo A fanmer can increase cassava production by expanding cassava area in several ways, assumíng that in the short run, total farm size remains the sama and that improved technology is not yet adopted. These options are to increase planted area by (a) reducing area in fallow, pasture or other crops; and (b) increasing area in cassava monoculture. In addition, traditíonally noncassava growing farmers start to plant cassava. Table 25.8 shows that total crop area both in absolute terms (ha) and as pareent of total farm area is greater at Level 1 than at Level 3. Hence cassava fanmers in areas of technology influence plant, on the avg, a larger area with crops; and the share of crops vs other farm activitles is larger than cassava farms in areas with low technology influence. Thís seems consistent wíth the observatíon that farmers at Level 3 dedicate relatively more of the!r farm area to pastures than Level1 (68% and 58%, resp.); and the absoluta siza of pasture land is more than twice as large as at Level 1. Table 25.6. Comparlsan ol ... ssava yíelds (cassavajmalzelntercrop) among •• lected Slates In Colombia, 1982 and 1991. Yleld ~/hal 1982 study' 1991 study' Su",. 7.0 10,7 Có'" semiarid ~nd .•. ·.···subtropiear . ...... ~ ··<;.r;]~,,~;);(I(~r ····;;..j.~¡~.¡;k;··:¡ :".;;./ ~s:~:?SoJr:~.?!~~'~~E~~~Ji~: ... ,;;~¡i~i.~,;¡,:~¡;~~":;~":. ......:.;;". ··.·····;·<:C. .• ;.f;:·notr'lant-usa ·.·efflciency,m¡ncassava ". and·· ; "",, - ; - . ~ ...• .•......... .... idéntifibatlori . ofplant CfíaraCtetlSUcif relátad to· '. ";;..' nutrient use that may be emplbyed assalectlof'l ". critEiria forcassava ¡mprClvam~nt .. DeVel0PÍl1ent . ............. : ...•.. ; .••.•.•........ '•.... :........ . ... of appropriate sOHfElrtllltytnaíntehanceand . • .... erosion control maaSCres: 'FóCüswillbÉJon hillsidé. sobhumid and semiarfdecosystems(5 '. yr): . . Research withllTA ón.theimPleméntatiC)!1,of ihtegrated 'management of thélPDM, inclociing diagnosis 01 fElrmers' '.' pesf control practicas, augmentation and conservatíon of natufal .. . anemies, classical biologicalCéintfol •. and ~ffeCt. of cropping systems on pest popufatlons' (5 YTl. . . CasS~¿á é¡~~bhhbíogy/ .. ' . Joint coordination with tha BRÜahdrese~rcih ·~etWÓfk(é§N~ . . . . . institution mambars of the caN ta ldentifYl'lew ~~i~+k~~ ... ;i:'=;:;~;;~ - ~-", '.'- , ':::::-.':, /'-)_::~ ~:-.:;, ;"- define an appropriate genetiq.· sth-Ícture.for . cassava propagatlOh from trtieseéd(5Yr}: .. .. Afri~ ·~··.Gérrtl~¡~fu~C~~} Jolnt·· project with .·IITA for th~ 'bitroductlon'arld ':,> ••.....•.•• evaluatíon of germplásm,adáptéd to mid"altifUde ····· .. · ... · ... ·.i···./and.· ... · .. ·saasonally· dry ..••. enWClnments. trom ""hbmOlogous areas .in theAl"rierict:ls (5 yr). 439 'BOX 26.1 COMPLEMENTARY ACTIVITIES (Cont.). , " " " . Asia· . SOclcl~nOm¡é reselÍrch.·· Establishment •. '01' .. á regional '.' maCl'OeCOnomic database fer bngbirig.evaliJatibn of!he dynamics . . U"'. " .. ' .. of da~sava develbpmenflrtAsiaand prometlón • .of .farm-Iével.researcfl. tp.detérmine the effectl\leriéss ofl1Elw' próduCtion teclíllofoQy' (3 . - '''' , -~: ':'-- - ,",'\" {:"¡',-"" .:,;',.,: " ". . . • yr);"., .. ..>.., ..... . , . -' ';', -~" ·.·.·~¡a·SUtiíii~·ahd~~tíiigjQ¡rif;ét:uVity with·théc~h@¡~X~OI1a/Cfeia . . ... .. '.. .......;..... ,(·p·i;l::1r'1P1··· tC1eStablfsffareiooiiflíífOrniatloh. ··F;;;J¡~¡~; :¡;;ex~~h~erietwórkaDddettfié~egfortáh,eséarCh .. 'oi,:,' ... prióriti8s.·· andoPf)Ortl..loities ror ". horlzohtál. .;;:: cooperation (3yt)¡ . "'lnte9ráf~'~~'sav~pro~~li1" Transferof knowl~dge to nationa/ programs bn ·ttopicarAm~ripa .•...• projeetconcQptlJalizatidh,. deSign, . executión and. ,';;:,.,-"" .'., ", • evalu~Óri;together with preparatión()ftr~l!1ing .' mateHals .' aria guideliries far' R&D "personnel ". . warkih9 ol1projects (ayr). '.