ANNUAL REPORT 2003 PROJECT IP-06 Tropical Fruits, a Delicious Way to Improve Well-being ntro lnlernaclonal de Agricultura lnternolional Cenler for Tropical Agrfcullure CONTENTS PROJECT IP-06 .................................................. ..... ...... ............ .. .. ..... ......... .... ... ......... ........... 3 Title: Tropical Fruits, A Delicious Way to Improve Well-being .... .... ........................ .. .......... 3 1 Researchers: .............................. ............................. .... ..... .. ... ...... ................ .. ..... ... .... .... ........ 3 2 Cooperators: .................................... .................... ............ ...... ........ ... ................... ............. ... 4 3 Highlights in 2003 .............. ........................... .. ................................................. .. ................. 5 3.1 Funding . ......................................... .. ................... ..... ... .. ........... .................................... 5 3.2 Strategies ....................... ............... ............ ....................................... ................... .... .. ... 6 3.3 What can be grown where ........................... ......... ................... ................. .......... ... ...... 7 3.4 Participative Selection of Perennials .... ................... ... ... .... .. ........ ... ...... .... ..... .............. 8 3.5 Thematic Tropical Fruits Network ............... ... ....................... .. ... .... ..... ... ....... ..... .. ...... 9 3.6 Fruit flies .... ... ........... .. ............................................. .............. ...... ............. ... ........... ..... 9 3.7 New Research Policies for Fruit Crop Research ...................... .... .. ....................... ... ... 9 4 Full Reports ...... ..... ............................... ......... ................... .. ................. ........... ..... .............. 1 O 4.1 Homologue development ...................... ..... ... ........... ...... .......... ...... ................... .... .. ... 1 O 4.1.1 Materials and Methods ........................................... .............................. ............. 1 O 4.1.1.1 Climate probability mapping ........... ........... .................. ............. .................... 1 O 4.1. 1.2 Soíl characteristics probability mapping ... .......... .......... .......... ... ....... ............ 13 4.1.2 U ser interface development. .................. .... ........ ....................... ...... ............. ...... 14 4.1.3 Results and Discussion ..... ....... ... .. .................................. ... ...... ... ....................... 14 4. 2 Participatory Selection ... ...................................... ......... ....... ..... ................................ 15 4.2.1 Selection of elite clones by famers usíng in vitro propagated plants of Solanum quitoense (lulo) ......................................................................... ....... .......... ........................ 15 4.2.1.1 Materials and Methods .................................................. ..... ........................... 15 4.2.1.2 Results and Discussion ............ ...... ........... .............................................. ....... 16 4.2.1.3 References ..................................... ................................... ............................. 19 4.2.2 Genetic variability ofthe Colombian collection of soursop (Annona muricata L.) and related Annonaceae species .. ................... .. ..... ...................................................... 20 4.2.2.1 Genetic variability ofthe different accessions of Annonaceae ..................... 21 4.2.2.2 Proposed taxonomic classification ofthe different accessions analyzed in this study 24 4.2.2.3 References: ............................................................ ...... .. ................................ 26 4.2.3 Anthracnose in Soursop (Annona muricata) in the Production Areas ofValle del Cauca, Colombia ...... ........... ............................................... .......... ............. .................. 27 4.2.3.1 Materials and Methods ....................................... ........................................... 27 4.2.3.2 Results and Discussion ............ ........................................ ... .... ........... ... ......... 36 4.2.3.3 Conclusions ................................................................. .......... , .................... .. . 51 4.2.3.4 References ........................... ..... .. .... .............................. , ............... ............. .... 53 4.3 Thematic Tropical Fruit Network ............. ............... ....... ... ................ ........................ 54 4.4 Fruit flies in two departments ofColombia .......................... .. ................................... 56 4.4.1 Literature Search ......... .............................................. ... .................. .................... 57 4.4.2 Results .............. .. ................................... .... ........................................................ 59 4.5 Towards New Research Policies for Fruit Crop Research .... .. ................................. . 63 4.5.1 Proposed Research System .......... .. ..... ............................................... ....... ... ...... 65 4.5.1.1 Existing Production Chains ..................... ............ .......................................... 66 4.5.1.2 New or Incipient Production Chains ................................................... .. ........ 70 4.5. 1.3 Technology Export ..................................... ................................................... 71 2 ANNUAL REPORT 2003 PROJECT IP-06 Title: Tropical Fruits, A Delicious Way to Improve Well-being 1 R esearchers: Major contributions to this years work in the Tropical Fruits Program were m ade by the following persons: Hom ologue Development: Peter Jones & William Diaz in the development of Homologue algorithms and software; James Cock, Simon Cook, Rache! O'Brien, Peter J ones & Thomas Oberthur in the area of conceptialization de,·elopment of new projects; Xav-ier Schelderman (IPGRI) and James Cock in the establishment of an inventory of neo tropical fruits. James Cock coordinated these activities. Participatory selection : Juan JairoRuiz, Vanessa Segovia, Edith Tabares, Fernando Hincapié, James Cock, Luis Alfredo Hernández, Carlos Quiroz, F. Parra, and Zaida Lentini all made major contributions to this cross disciplinary effort on lulo. Corporación Biotech, Alvaro Mejia and James Cock were active in the work on propagation and selection of sour sop; Elizabeth Alvarez and C.A. Ospina carried out the extensive work on Antthacnose. Thematic Tropical Fruits N etwork Liliana Rojas and J enny Correa Fruit Flies in two areas oí Colombia: Tony Bellotti, Maria del Pilar Hemández and Monica Lucía Mario. N ew Research Policies for Fruit Crop Research in Colombia: James Cock, Ximena Rueda (MADR), Juan Jaramillo (CORPOJCA) Many others made important contributions to discussions and prm·ided support and services to the program. Their anonymous contribution is hereby acknolwledged. A special mention is reserYed for Adriana Cardona the adminstati,·e assistant of the program. 3 2 Cooperators: Within CIAT: Integrated Pest and Disease Management (PE-01); Land Use in America (PE-04); Rural Agroenterprises SN-1; Participatory Research SN-3; use of Genetic Resources SB-2. Outside CIA T: Nacional Programs: Colombia (CORPOICA, SENA, Corporación Colombia Internacional); IPGRI, Missouri Botanical Garden, University of Queensland (Australia); PROTA-University of Wageningen (Holland); ICRAF; GF AR; University of California Berkeley, University of Florida 3 Highlights in 2003 The Centre's efforts in tropical fruir research and dev-elopment in 2003 concentrared on (a) consolidating the basic straregies dev-eloped during rhe preYious year and developing and implementing an operacional plan and (b) setting up a m ore solid and stable funding base. The program has suffered from extremely limired and uncertain funding during the year and future funding is extremely uncertain: this situation evidemly makes coherent long rerm planning difficult The assured resources of essentially one intemationally recruited staff member (75%) with limired support are being used to (i) to search for financing (ii) to iniciare research, based on the long term straregies, that will help attract funds and (iii) ro continue with research to resolve problems identified by the Ministry of Agriculture and Rural Development of Colombia and (iv) to continue with the Sour Sop project funded by Spai.n (conflitlled in the second semester of 2003). In this report we present the funding situation, modifications ro the O\erall strategies, and researcb and development activ-ities. 3.1 Funding. The program is at present finaneed principally by a small arnount of eore, pan of the Colombian contribution and the Spanish Gm·emment. The present finaneial status is not suffieient to implement the strategie plan, and hence an increased funding base is required. The possible sources of future funding are an inerease in the allocation of eore resources as seed capital ro ser up the programme, special projecr funding to implement specifie components of the strategic plan and contraer researeh . A t present the Centre has assigned less than 2% of its projected unrestricted core budget to the fruir Fruit Program and its sbare of administrative eosts for 2004. This not only places the Fruit Program in a difficult siruation due to the low leve! of core funds themselves, but it also severely restricts the programs capaciry to obtain special project or contraer researeb funds. Most CIA T special projects or contraer research projeets are beaYily subsidized by eore, normally in the form payment of a large proportion of the salaries and administrative costs directly from core. Those programs that have a small core assignment are not able to use core as leverage to subsidise special pro jeets and make them lower cost and henee more attraetive to donors. The fruit program lost the opportuniry ro carry out an interesting contraer researcb project on Integrated Pest Control of Cape Gooseberry (PfD•sa/is peruviana), that was highly rated from a technieal point of Y-iew, due ro the high cost of the project: the project was eventually assigned to a uni\ersity whieh placed Yery low charges for scientists time. The siruation of tbe Fruit Program is further complicated by the fact that many donors are looking for projects which will ha\e a direct impaet on poYerty in a short period of time: this is difficult to achieYe in a new program that is starting from an extremely narrow knowledge base with few research results in tbe pipeline. A eonsequence of this siruation is that the future directions of tbe program are likely to be more dependent on tl1e whims of specific donors than on the lines laid out in the strategic plan. 5 The program is ata crucial point where it requires seed capital to build up a reputation so that it can present itselfto donors as a viable unit that has an established research capacity. Once the program is functioning and has reached critica! mass it should obtain special projects that build on the demonstrated knowledge and expertise developed within the prograrn. The financia! dilemna of the Fruit Program is that without a reasonable assignation of seed capital to generare knowledge and expertise the program it will not be able to attract funding. The solutions to this conundrum are not clear at present. The possibilities, all of which are being actively sought, are: 1. To increase substantially the core budget assigned to the Fruits Program either directly or through collaboration with core financed units within CIA T 2. T o attract a major donor interested in providing long term support for the development of an Imemarional Fruit Program 3. Package the components of the proposed Fruit Program into individual projects and search for .financing of the individual packages. 4. Incorporporate components of the Fruit Program into project proposals jointly presented by the Fruit Program and other projects. During the year a large number of concept notes and project proposals were prepared, often in collaboracion with other encities within and outside of CIA T, for presentacion to several potencial donors under the strategies outlined above. These included: 7ñe development of Homologue and Cropldent to determine what to grow where; Better informed decisions through the integration of local and scientiiic knowledge; Selection and Propagation of Proven Perennial Fruit Trees; B~oing information symmetry in smaU holder coffee systems for inteDigent production and transparent marketing of high value products; Rapid selection of Sour Sop (.Annona muricata L) suitable for specifJc agro-ecological conditions (presented to Colciencias by Corporaración Biotech with participacion of CIA1); Integrated pest management for Cape Gooseberry (l!_hysalis pemvíana), presented with the Nacional University, Bogotá, Colombia; Genetic resources of native and exotic fmit species in the Brazilian semi-arid region: improved Jivelihoods through consolidated agribusine5s in the irrigated areas of the Sáo Francisco basin presented to the Water Challenge Program by EMBRAPA; Zonas Aptas Para Producción De Frutas Y Hortalizas Especiiicas; What to Grow Where; Technical Assistance to Strengthen the Production of Alternative Agricultural Commodities in Colombia and other Andean Countries; and Flowers in the Park. Colciencias approved the project on rapid seleccion of Soursop, with the financia! support descined to Corporación Biotec and CIAT providing counterpart support from the final year of the Sour Sop project fmanced by Spain. The project What to Grow Where was approved for funding, at a reduced level, by the USAID linkage grants scheme. 3.2 Strategies The mission of the Tropical Fruits program conrinues to be: " to use science, technology and modem information technology to provide information and support to partners in the public and private sector that prome te production, processing and marketing of tropical fruits by rural communities which leads to increased wealth and improved welfare for present and future generarions in the couotryside." The strategy conrinues to be to prov-ide support for the development of Tropical Fruits in general without concentrating on any one particular fruit species. At the same time we recognize that to 6 develop general principies we will need to carry out research on specific fruits. In general in these cases we try to ally ourselves with over agencies, such as the de,-elopment of techniques for rapid selection of perennial species in which Sour Sop is the model species and we 'vork closely with Corporación Biotech. The strategic research areas have been revie\.ved in the light of the possible donor interest and the e}.:pertise that we have in house. It is difficult to develop coherent project proposals in fields in which we do not haYe in house expertise. TI1e program is emphasing the following areas: );> Targeting which crops or cultivars will grow well in particular conditions );> Generic research of general applicability to many fruit crops );> Agro-enterprise development :>- Contract research Market intelligence, which was considered last year as one of the strategic iniciatives has been place on hold as we do not have a great deal of in house e>.:pertise in this area, and furthermore several other agencies are active in this particular fíeld. In the area of generic research of general applicability to many fruit crops we have not been able to obtain financia! support for the flowering and post harvest deterioriation iniciatives. On the other hand there has been limited donor response in the area of developing technologies for rapid selection of improved genotypes of perennial fruits and we have moved ahead in this area. 3.3 What can be grown where In last years annual report we noted that in agriculture one of the foremost questions is ''What can be grown in a specific site". Furthermore the developroent of specialized software (Cropldent™ & HomologueTM) was suggested as the fust step in the development of a system to target what crops or cultivars will grow well in particular conditions (see section Homologue). Cropident requires to be supported by databases on the individual plant species or cultivars, and work commenced on the establishment of a working inventory of Neo Tropical Fruits. We are also progressing on the development of the user interface for the software. In addition arrangements have been made with the University of California, Berkeley (USAID linkage funding) to develop methodologies for combining local knowledge, scientific knowledge and expert opinion related to which crops succeed under which biophysical, socio-economic and management conditions. These methodologies will then be incorporated into software modules to provide appropriate, user-friendly tools for each situation along with farmer or farmer group accessible outputs. If a particular site is georeferenced the climatic conditions of that site can be accurately estimated from existing climate databases. Soils on the other hand are spatially much more heterogeneous and methodologies for on site evaluation of soils are required for successful use of Homologue and Cropident. Work has commenced on the development of a Rapid Soil Appraisal (RSA) methodology. Revision of previous 'vork suggests that the RSA can be based on the work in l'.·fexico ofSiebe,Jahn and Stahr ofthe University ofHohenheim. 7 A first step in the setting up of Cropident is to have a full inventory of all the potencial crops. In the case of tropical fruits we have initiated the establishment of a readily accessible database of Neo Tropical Fruits. 11ús work is based on the ethnobotanic in\entory of neotropical frui ts currendy on the web which was developed by CIRAD-FLHOR and IPGRI, based on original work by Fouqué. This inventory was conceived as an electronic book, and its structure is not appropriate for eventual incorporation into Cropident and linkage to Homologue. In conjunction with IPGRI the information in the electronic book is being transferred to a structured database. During this process additional information is being added with emphasis on geographical distribution of the species, and the availability of germplasm. Linkages are being established with such organization as the Missouri Botanical Garden which have extensi\e databases on tropical plants with massive information on the geographical distribution of species. 3.4 Participative Selection of Perennials In spite of the attractive nature of new varieties of fruit trees as a means to increase rural income, d1ere are few examples of successful formal breeding efforts to improve perennial fruit crops. The lack of success in breeding fruit trees is related to few long-term efforts to improve fruit trees, particularly in tropical species, and the long generation interval and period of testing required in formal breeding programs. Furthermore, with so many different fruit crops it is most unlikelr that tradicional research organizations will dedicate the time and resources necessary to structure comprehensive breeding programs similar to those developed for short season annual crops or food staples. On the o ther hand, most fruit tree varieties favoured and planted by farmers c:hroughout the world are the result of selections made by farmers themselves, or researchers, from chance variation found in naturally occurring seedlings. The key to successful use of these selecti.ons is first of all to nurture them and recognise their value. Second is asexual propagati.on, which ensures that all the plants are geneti.cally identi.cal, but which is also carried out in such a manner that systemic diseases are not propagated. A major potencial lirnitati.on of participative selection of pereronial species is the location specificity of the elite lines, and hence the difficulty of transferring the technology. We are developing methodologies to determine the similarity of sites in different locati.ons (see above) and this will be used to determine in which areas the selected elite lines are likely to be successfuJ. We have taken two species, Lulo (Solanum quitoense) and Sour Sop (Annona muricata) and are using them as model species to develop methodologies for farmers and traders to select superior materials that are found spontaneously in the field, multiply them and plant them in areas where they will perform well (see sections on Lulo and Sour Sop for full details). Colombia is the center of origin for sour sop, and the crop suffers various problems caused by phytopathogens that have co-evolved with it, in particular, andtracnose. This fungal disease, caused by Co/letotrichum species, is spreading as the crop expands into areas that are environmentally favorable for the fungus. As a result, anth.racnose is now a major constraint to soursop producti.on in Colombia. To support our work on Sour sop the nature of the disease was studied (see secti.on on Sour sop). 3.5 Thematic Tropical Fruits Network This year, making use of REDECO's Internet experience ro generate a flow and exchange of perrineot, updated information, the Tropical Fruits Program and REDECO plan to create a themaric network of users interested in the theme of tropical fruits. The network will facilitare idenrificarion of the demand, capture, and d.iffusion of specialized information, and will promete dialogue between users interested in belonging to the network. From May 2003, the theorerical framework of the Tbematic Network on Tropical Fruits was developed, and a sun·ey prepared aimed at the REDECO community (see Thematic Tropical Fruit Nenvork). 3.6 Fruit ilies There are certain groups of pests that consist of a species complex that can attack and damage numerous fruit species. The fruit fly complex is an example of a pest that can damage numerous fruit species, and hence its study falls within our strategy of research on generic problems of fruits. See section on Fruit Flies for full details 3.7 New Research Policies for Fruit Crop Research Tbe Colornbian Ministry of Agricultu.re and Rural Developrnent (MADR) requested assistance from CIAT in designing research policies to support tropical fruit based enterprises. Working closely with the MADR we have developed guidelines for more effecrive use of research resources on Tropical Fruits .. -\Jthough the policies are directly related to the Colombian situation, many of the recommendations are highly relevant to many other developing countries. The development of strong nacional or local research capacity is a sine qua non for success in promoring tropical fruits and strengthening local capacit)• is an integral part of the program strategies (see section on Research Policies). 9 4 Full Reports 4.1 Homologue development TI1e tropical fruit program is deploying two rnajor software tools to assist in its goal of long-range transfer of tropical fruit germplasm. Cropldent will use a large database of inforrnation on crop characteristics to match fruit crops ro potencial environments. However this database does not exist at present and must be built up as new information becomes available. Homologue will provide a complementary function and assist in the production of the Cropldent database. The basic concept is that a farmer's field (target atea) will have homologues somewhere in the tropics. Once these are known the local crops in these areas can be inYestigated with the v"iew to introduction in the target area. In this process information will be gathered that will provide input to the Cropldent database. We also envisage Homologue being used to extrapolare from a small nurnber of characteristic sites where fruit crops are known to do well. By combining Homologue probability estimares from these sites into a 'cloud' of estimares we hope to overcome the mínimum accession set restrictions of FloraMap. 4.1.1 Materials and Methods 4.1.1.1 Climate probabili!J mapping We compiled a set Eigenvectors and Eigenvalues from 25 FloraMap analyses. We chose to estimare the first five that account for over 96% of the variance in almost all cases. We rnapped the probability surface from the components that explained 96% of the variance at levels over p> 0.3. We calculated the pixel coverage to get an estimare of the geographic area of adaptation . The variance of the accession points was also recorded. The adaptation area and variance defined the adaptation range. The studies were extended ro 136 FloraMap analyses for a wide range of species, and we attempted to fit regression models to the eigenvectors and encountered a number of problems. The first is that the sign of the eigenvector is ambiguous. This normally does not matter, but in this case we had to write a special program to determine the correct alignment by trying the fit with the sign in both directions. After various attempts to irnprove the model fits, including clustering the species by climate type, sorting by the nurnber of accessions and eliminating unusual eigenvectors, we decided that the task was beyond resolution. Even if a good fit could be achieved with the individual eigenvectors the complication of fitting all at once with restrictions to maintain orthogonality would have made the exercise too complicated. Eigenventors for a cluster of 12 species 0.3 . 0.1 - ¡.E ·> o -0.1 - -0.2 - -0.3 Figure 1. Eigenvectors from U species in a cluster grouped by climate type Figure 1. Shows a typical example of a set of species clustered by clima te cype. The black lines are the fitted values from the best model fit, the red lines the actual eigenvectors. Although the R squared value is oYer 70% the fit is nor good enough ro serre. We therefore decided to select a set of representative species and borrow the eigenvalues and eigenvectors as they stand. This had the advantage that, when they are applied to the acrual observed data the probabiliry model that results will be correctly orthogonal. To do this a simple climate classification was devised: this was based on the temperarure record dividing the observed climates into five groups by mean temperarure and five by the variance of remperature. Table 1 shows the classification along with the colour index for the maps of che classification in figure 2. There are only 16 colours available in the palette for these maps so the colours cross classes in the cable. 11 Climate Classes by Annual temperature characteristics Variance Mean °C <1 1-2 2-5 5-10 >10 Table 1. Climate classification and color legend for the maps in Figure 2 Figure 2. Clima te classes for development of the generic climate probability model The 136 FloraMap analyses u·ere classified using this classification and representatiYe species for each group selected: Species Mean Variance < 15.5 15.5-19 19-22 22-24 >24 < 1 Oxalis tuberosa Phaseoltt.r /eptostacheu.r Pha.reolu.r vu/;!,aris V~na dat')li V~na neroosa 1-2 S olanum nigrescens Phaseolu.r /epto.rtacheus Pha.reoltts vu/;!,aris V~na datyi V{~na nen,osa 2-5 v~~na parkeri V~na luteola Vigna ve ...... ?llata !Arachis burcharlii V1rachis duranensis 5-10 S~ylosanthes guianensis Sryiosanthes viscosa Bemisia tabaci ~rachis glabra/a 4rachis duranensis >10 V{gna ambacensis Vigna tJemtlosa Manihot aesculifolia l!lrachis cardenas V1rachis duranensis Table 3. Representative species in each climate class used for the generic model. Those in red are surrogates standing in where no species was yet available in that group. More data sets will be collected from FloraMap users o,·er the near future and the table will be extended. The eigenvectors of the selected species were used to represent the model for each climate. The eigenvalues were easier to model and we derived regression models to estimate them for each climate. Based on this classification we wrote a generic routine for climate probability comparison. We derived new climate files and indices and detennined that the compressed data can acrually be stored in RAM in a space of about 20Mb. This makes the algorithm much more efficient. Subroutines were written to load the data, produce the climate probabiliry shapefile and to return a climate record on reguest. 4.1.1.2 Soil characteristics probabiliry mapping This section of the work concentrated on working out how to provide and store the cumulatiYe probability curves for the soil factors. The WISE soils database raw data prm·ided enough individual profiles for many of the characteristics in a fair proportion of soils. From these we calculated variances for Depth, soil carbon, nitrogen, ph in water, KCI and CaC12, CEC, and texture as sand, silt and da y. There were a fair number of missing values in the resulting table, but most of these were filled with estimating regressions. This table of means and variances of the listed characteristics is now complete for all agricultura} soils in the FA O classification. The format for the soil characteristic integral takes 12 bytes. 2 bytes for the percentage of agricultura! soil per pixel, 2 for the OYerall soil coverage per pixel and 8 single byte scaled incremems. End points for the x-axis of each soil integral are held in a table separare from the pixel integral file. We have de,·eloped an interpolator function for rhis integral that can handle any shape o f monotonic probability integral. After a fust read through the Latin American files we determined that the data for soil carbon nitrogen and CEC need transformation to normalize the distributions. Carbon needs a naturallog transformation and both nitrogen and CEC need square root transformations. Other characteristics seem to be sufficiently normal to work without transformation. An encoder was 13 developed that takes all soil mapping uoits falling within a climate pixel, sorts out the agricultura.! soils and using the means and variances for each soil in each mapping unit compiles a probability integral for each soil characteristic. Figure 3 show the probability integrals for soil carbon for three random pixels in Latín America. -4 probability of finding soil carbon in a given pixel -2 o 2 Ln organic carbon 4 Figure 3, Selected pi.-xels, the probability integrals for soil carbon. 4.1.2 User interface development. We have started development of the user interface. The point selectioo and mapping routines are being derived from those of MarkSim with the additioo of capacity for displaying probability maps. The interface to the mapping dll is determined and passing structures defined. We have determined an efficient way of holding the large data arrays for access by the Fortran dll. 4.1.3 R esults and Discussion. The object of this exercise is to produce a demonstration version of Homologue within this year for demonstration to donors for support for a full system. To this end we have restricted ourselves to using the existing 18km climate grids and sorne necessary function will not be developed untillater. One of these is the allowance for correlation amoog the soil characteristics. This can be overcome but needs sorne theoretical work. Extending the product to more precise climate grids will have to wait until the new indexing systems are fully developed. Implementing these will require substancial change to the way the data is accessed in future versions but this is not em-isaged to be a great problem. 4.2 Participatory Selection 4.2.1 Selection of elite clones by famers using in vitro propagated plants of Solanum quitoense (lulo) A large number of fruits of Andean origin have great potencial to become preroium products for local and export markets with a high economic retum for the farmers. S olanum quitoense, locally knmvn as lulo in Colombia and as naranjilla in other countries, is among these fmits. Th.is species is native from Colombia and Ecuador, and it is normally cultivated between 700 and 2000 meters above sea level. Sorne of the main attributes of this fruit include its high le\el of vitamin C, and the sub-shrubby perennial growth amenable for cultivation in hillsides and inter-cropping, aiding soil conservation practices. Recently in Colombia, naranjilla changed from being a fruit of local fresh consumption to become an important industrial fruit for juice and yogurt products, increasing its market value. A major constraint for the rapid adoption of naranjilla by the local farmers is the limited availability of elite germplasm free of pathogens from clona! propagation. The high level of heterozygosity of this species is reflected in tl1e high segregation of traits by its multiplication through botanical seeds. Rapid multiplication of high quality planting materials is of paramount importance. One of the main objectives of this project is to develop a protocol for in vitro propagation of naranjilla with application for conservation and rapid multiplication of elite clones free of pathogens. This protocol would facilitare the conservation and multiplication of high quality planting material aiding control of diseases which are one of the major constraints of this crop. In the previous two years we reported the development of an efficient protocol for the maintenance and propagation in vitro of dones selected from farmers field, for the plant regeneration from tissue cultures, and the preliminary evaluation of those plants in the field. This year we report the agronomic performance of those materials in the field, including quality traits of the fruit, as well the progress made to evaluare jointly wíth farmers the application of this technique at larger scale. 4.2.1.1 Matenals and Methods Plant material High quality and elite clones pro,-ided by the Andean Fruit Center (Centro Frutícola Andino - CEF A) were used. This collection includes naranjilla with or without thorns commonly grown by farmers. The plants were propagated in vitro or plants were regenerated follow:ing procedures as reported by Segovia et al. in SB2 Annual Report 2002. Plant and fruit evaluations. :\. small-scale field trial was conducted at 1700 m over sea level and a mean temperature of 22C to compare the growth, development and agronomic performance of regenerated plants respect to in vitro propagated clones. Fruits produced during a period of five months were harvested, counted, weighted, and classified by size according to standard scale. Premium size: > 5.5.cm diameter. Commercial size: from 3.5 cm to 5.5 cm in diameter. Not commercial: < 3.5 cm in diameter. The fruit maturation process (days from harvest to ripe to rotten) was evaluated using the Munsell color scale. Fruit quality trait analyses 'vere conducted at the Valle UniYersity (Cali). These analyses included Brix grade at 20C, relative humidity, acidity percemage, vitamin C, sugar content, pH. A sensorial panel analysis was conducted to evaluare the fruits for its appearance, color, aroma and flayor as fresh fruit or processed in juice, and compared with those available in the supernurket. 15 In vitro introduction of plants from greenhouse or field. Greenhouse or field grown plants were tested to select the best protocol for tbe in vitro introduction of elite fanner's clones from the field. In order to establish plants in the greenhouse, the recmrery of plants from axillary buds using shoot stakes was tested Stakes of 20 cm in length from ad,·entitious shoots witb 2-3 axillary buds were taken from selected plants in the field and disinfected first with 3 ml/ 1 of fungicide propamocarb HCL (Peroicur) for 5 min. The basal section of the stake containing the lower two buds was either: a) Cultured in water until root formation; b) Soaked in a solution of 10 mg/1 NAA for 3 days and then transferred to water until root formation; or e) Potted directly in sterilized mix of soil, sand and sugar cane plant residues (Franco et al., 2002). Treatments (a) and (b) were aerated with a fish tank pump system. In order to establish plants in t!Ítro, apical or axillary meristems from plants grown in the greenhouse or field were tested. Explants were surface sterilized and cultured in vitro on medium A instead of medium 17N (Segovia et al., 2002. SB2 Annual Repon). Riphampicine (antibiotic) 100 mg/1 was added to the medium to control bacteria] infection from .field explanrs, A Mean fruits per plaot GC GR TC TR Plant type Mean fruit weight per plaot GC GR TC TR Plant type Figure l . Fruit production io the field of in vitro propagated (C) and regenerated (R) plants derived from materials with thoms (T) or without thoms (G} since the presence of trichomes in the tissues prevented thoroughly disinfections. Evaluation of the technology with farmers. A total of 20 fanners from two regions of Colombia (Cauca y Huila) with commercial production of lulo were selected. Farmers were chosen based on the years of experience cultivating lulo, the number of lulo plants curremly grown and their interest to participate in this initiative. A strategy was planned jointly with the fanners to evaluate tbe potencial commercial use of in vitro generated plants as an alternative of planting material clean of pathogens. Tbis activity was conducted in collaboration with IPRA, the tropical fruit project of CIA T, and Corpoica- Popayán. 4.2. 1.2 Results and Discu.rsion Last year we reponed no differences in plant growth and development between regenerated and in vitro propagated plants. Likewise, no significant differences were noted between the regenerated and in vitro propagated plants respect to the total number and to tal weight of fruits produced (Figure 1A). However, the clones with thom s (I) showed higher disease and pest susceptibility that significantly reduced fruit production (Figure lA). The average productivity in S months of harYest indicated a mean number of fruits per plant of 142 for materials without thoms (G) and 28 fruits for T (Figure lA), and a total weight of fruits per plant of 5.58 kg for G and 1,57 kg for T (Figure 1 B). About 80% of the fruits Figure 2. (A) Plant without thoms. (B) Fruits showing commercial and non-commercial sizes. (C) Dark green flesh that showed high acceptance by panelist in the quality trait assay harvested from these clones showed a commercial size (from 3.5 cm to 5.5 cm in diameter). The rest of the fruits were classified as 15% small fruits ( < 3.5 cm in diameter) for the G plants, and 1 0% premium size (> 5.5 cm in diameter) for the T mateóals (Figure 2). Commercial production of lulo may yield per year about 135 fruits / plant and 9 kg fruits / plant when cultured at a densiry of 3,000-plants/ hectare (3 m 2 per plant, CCI). In this experiment plants were also planted at a distance of about 3 m2 per plant. Of the plants evaluated four were selected fo r their high yield potencial. These plants produced 9 kg fruits / plant during the Y2 year experimental harvest tin1e. These plants were selected; seeds and stakes of these plants were collected, and grown in the greenhouse and in vitro. In relation to fruit quality, no significant differences were noted between the in vitro propagated clones, the regenerated plants and the commercial fruits bought in the supermarket for most chemical traits evaluated. The exception was the content of reducing sugar, which was lower in the supermarket fruit possibly indicating a longer post-harvest time respect to the experimental fruits at the moment the analyses were conducted (Table 1). The two sensorial analyses conducted gave different results. The first panel with S panelists shows a higher acceptance for the experimental material~ for all traits, but the panel with 11 panelists indicated a higher acceptance for the supermarket fruit in relation to flavor and aroma, and higher acceptance for the experimental materials for color and physical appearance. This analysis seemed to be highly subjective, thus it may need to be conducted with a larger number of panelists to elucida te dearer preferences. No differences were noted in shelf life between the different materials. 17 Table 1. Chernical quality trait analysis of fruits from regenerated plants without thorns (GR), in vitro propagated plants without thoms (GC), regenerated plants with thoms (TR), in vitro propagated with thorns (fC), and bought in supermarket Trait GR GC TR TC Super- market Brix grade (at 20 oc) 11.5 11.6 11 .0 10.0 10.5 Acidity (%) 2.2 2.2 2.7 2.3 2.9 Hurnidity (%) 86.8 85.2 79.8 87 82.4 Reduced sugar (%) 4.6 5.5 5.0 4.9 0.7 Total sugar (%) 3.1 3.4 2.7 3.1 2.7 pH 3.2 3.2 3.2 3.3 3.2 Vitamin C (mg/1 OOg) 38.2 37.3 48.8 64.2 39.1 Attempts to est.ablish plants in the greenhouse from field grown materials indicated that the highest percentage of srakes with new shoots was obtained when the st.akes were planted directly into soil in the greenhouse. About 50% of the st.akes had healthy looking plantlets 1 month after planting (Figure 3). Explants derived from clones with thoms responded poorly, perhaps due to the weak stage of the donor materials since they were highly affected in the field. Once plants were est.ablished m the greenhouse st.akes from greenhouse-grown plants produced profuse roots when treated with aerated water, which ease the mass clonal propagation of plants in the greenhouse. A high percent.age of cont.amination was obtained when meristems were introduced in vitro directly from field grown plants. E>.:plant swvival (elongated shoots with roots) was increased from 40% to 60% when the antibiotic was added to the culture mediwn. o contamination was noted and at least 80% e:x-plant surrival was obtained when meristems deri•ed Figure 3. Percentage of stakes with new shoots derived from field-grown plants and treated in the greenhouse. (G) No thoms, glabrous. (f) Thoms. 60 -~ .. 50 - "' o 40 o .= "' .= -·~ 30 "' ~~ !': 20 "' -e c:c: 10 o: o Water NAA Asofrucol Treatments from greenhouse grown materials. Based on these results, successfu1 in tlilro introductions of elite materials selected m farmers fields can be achieved by est.ablishing first clonally propagated plants in the greenhouse, and use these plants as donors for the in vitro culture of meristems. In the case a direct introduction from the field into zn vi/ro conditions is needed, a more efficient surface sterilization protocol needs to be re•ised. A process was initiated to test with farmers the suitability of using in vitro propagated plaots as planting materials. The potencial advantage on the in tJitro source is the supply of pathogen free homogenous plants maintaining the selecred traits of the elite materials. Farmers were selected from rwo sites with commercial production of lulo (Pescador, Cauca, and Tierradentro Cauca-Huila). These sites include small, medium, and large-size production farms. The fanners haYe between 2 years ro 9 year of e:1 .. :perience cropping lulo, and 200-5,200 plant-size fanns. Fanners attended a workshop at CIAT with d1e objective to define joindy a strategy ro evaluare the potencial commercial use of the in rJitro plants. The advamages and lirnitations of using in vitro grown planrs were discussed. The farmers enluated the plants at d1e CIA T lulo e},:perimental plot. The criteria were established for the colla.boration. Fanners are highly enthusiastic with d1e project, and a teamwork approach will be in1plememed. For each site, two nurse!)· plots will be established. Farmers will select the best two plants currendy grown in their fields based on productiY:ity, fruit quality, and disease/ pesr resisrance. These plants will clonally propagared in the field, and explants will be established in the greenhouse at CIA T. Meristems from these phints will be introduced in tlilro. Pathogen-free in tlitro propagated plants will be grown in d1e nursery plots joindy with the field-clonally propagated plants, and seedlings derived from the seeds of d1e selected plants (standard propagation mode used by the farmers). The in tlitro, clonal-propagated and seed-propagated plants \v-ill be compared duoughout the production season. The nursery plots could also be used for the evaluation of new germphism, one of tbe main needs identify by the farmers, in addition to assisrance for an integrared crop and disease/ pest management approaches. Corpoica phiys a key role in these aspects, as CIA T is trying to strengthen their linkage to the project Farmers selected their best elite materials and currendy the plants are being clonally propagated in the field. Plants will be brought to CIA T ro initiate their production in tJitro. Field evaluations will be conducted first semester of 2004. 4.2.1.3 References CCI (Corporación Colombia Internacional): www.cci.org.co/ publicaciones / Tropico /TROPICOOS.htm CIA T, 2002. The deYelopment of methodology for in vitro multiplication, plant regeneration and genetic transformation of naranjilla Oulo). SB2 Annual Report. p: 304-306. George, Ed,vin F. 1993. Plant propagation by tissue culture. The technology. Part l. 2nd. Edition. Great Britain. p: 1361 Franco, G.; Bernal, J.; Giralda, M. 2002. El cultivo del lulo. Manual técnico. Asofrucol, Corpoica y F N F H . Manizales. 1 03 pp Segovia V. 2002. Optimización de la regeneracton de lulo Solamtm q11itoense orientada a la transformación genética de plantas. Uni,ersidad Internacional de Andalucía. Sede Santa Maria de la Rabida- Hueln. España. 74 p. 19 4.2.2 Genetic variabilityofthe Colombian coUection ofsoursop (Annona muricata L.) and related Annonaceae specie"s Colombia is rich in species of the family Annonaceae (Murillo 2001). Many species which produce edible fruits belong to this family,..roainly in the genera Annona and Rollinia, several of which are of actual or potencial economic importance. These include: soursop CAnnona muricata L.), cheri.moya (A. cherimola), anon (A. squamosa), anon amazonico (Rol/inia ed11fis), anona colorada (A. reticufata) and atemoya (A. squamosa x A. cherimo/a) among othe.rs. The rich genetic variability in the country contrasts with the limited availability of well characterized cultivars. Commercial nurseries offer two or three cultivars of soursop and few of other species. We deve1oped a methodology to propagate selected trees of soursop through in vitro micrografting. This methodology can be used for the production of clean planting material of soursop and other related species. This technique is only commercially useful if highly productive clones, of good agronomic performance, which produce fruits o f good quality are identified. The Colombian Corporation for Agricultura! Research - Corpoica Palmira station maintains a collection of germplasm of soursop and related species, with 32 entries of soursop and 12 entries of other Annonaceae. This activity is part of a project aimed to characterize, agro-morphologically and molecularly, using AFLPs, this collection and other available accessions (Royero et al. 2002, Mejia et al. 2002 and Saavedra et aL in preparation). During 2003 we continued with the molecular characterization of the collection and with the analysis of the information produced. Metbodology: Seventy eight accessions belonging to species of the genera Annona and Rollinia were characterized through the use of molecular markers of the AFLP type. The accessions were composed of 37 entries from A. muricata and 41 of related species, mainly from nacional gennplasm banks of soursop and related species, maintained at Corpoica - Palmira, but also from banks of other institutions, farms, markets and commercial plantations (see annual report of 2002 for the list of accessions). We applied the method of D ellaporta et al. (1983) for DNA e>..'traction, using PVP 1 g f 1 in the extraction buffer, and a chloroform / isoamylalcohol (24/ 1) cleaning step before D NA precipitation. The AFLP protocol followed was that described by Vos et al. (199 5), using AFLP Analysis System I Kit (Gibco BRL). In preliminary trials we evaluated 15 diffetent combinations of selective nucleotides fo.r AFLP amplification. Then, we selected three combinations that displayed the highest and "most readable" polymorphisms between soursop species and/ ot between Annonaceae species. They \vere E-ACT, M-CAA (combination F); E-AGC, M-CTC (combination M); E-AGC, M-CAA (combination N). Each accession was analyzed using two primer pairs. The combinari.on of primers F and M \Vas used for the soursop accessions and combinari.ons M and N for the Annonaccae accessions. Selecri.ve amplificari.ons were size-fracri.onated on 6% or 4% polyacrylarnide denaruring gels and Yisualized through sih·er staining. AFLP fingerprinting of each accession were converted into a similarity matrix, based on Nei and Li index (1979). The sirnilarity matrix was analyzed using NTSYS (Rohlf, 1994) computer program. Dendrograms were constructed by UPGMA method (Sneath and Sokal, 1973). Results and Discussion: 4.2.2.1 Genelic variabiliry qf the dflferenl acce.rsions o/ Annonaceae An example of the AFLP banding pattem produced by the different accessions is shown in Figure 1. The comparison of the similarity matri.~es obtained with the two different primer pairs, showed a high correlari.on (r =0.93-0.95 ) in the soursop accessions, as well as in the other Annonaceae accessions, indicari.ng that both primer combinari.ons detecred a similar leve! o f genetic variability, and that the data produced were reliable. The dendrograrn produced frorn the AFLP data of the Annonaceae accessions is shown in Figure 2. Tbere are two rnain clusters, with similariry level of 0.1 O, the first composed by a single accession, Cadmia (Cananga odorata), an Annonaceae of Asiari.c origin, included in this analysis as an out-group, and the second group includes the rest of accessions, which belong to the neotropical genera Annona and Rollinia. At a level of similarity of 0.55 (gray line in rhe Figure 2), nine subclusters can be detected (marked as groups G1 to G9 in the Figure 2). Since the taxonornic classificari.on of severa! accessions of these subclusters is known, it can be deduced that these subclusters represent at least nine different species, as follows: Gl . A . glabra; G2. unidenri.fied Annona spp.; G3. Rollinia spp; G4. A . montana; GS. A. muricala; G6. A . p11rpurea; G7. A. reticulata; G8. A. cherimo!a; and G9. A . .rquamosa. The A . .rquamosa x A. cherimola hybrids known as atemoyas, all fell within groups 8 and 9, which confirms thcir hybrid nature. A low similarity, or in other words, a high variability was found in the accessions of the subclusters G3, G4, G7, G8 and G9. Dueto the low similarity presented by sorne accessions (0.75 or less) of the same subcluster, it is not clear if sorne o f the subclusters such as G3, G4 and G7, are composed of accessions of one or more spec1es . .-\Jthough sorne of the subclusters are represemed only by few accessions, the variability found among them, can be exploited in breeding actiYiri.es. 21 Figura 1 Example of the AFLPs banding pattems of Annonaceae accessions obtained with the primer combination M (only the superior half of the gel is presented). Group 1 Gl0up3 Group4 G1oup5 G10up 7 G10up8 Oroup 9 A. glabra Roliinia spp A mont.2na A muricda A ret¡c:li~ A .. ci!~rimoia _4. squamosil. 1 z 3 4 5 ' '7 8 10 11 lZ 13 14 15 %!1 30 31 lZ 19 %0 Z1 ZZ Z3 '1:;' - -- 1<"-i - q~ . io:ziai l Ml .,_,,.., ... -..t ~~~: ~ .... lilllil ijiillí -- ~·~ ~" ¡¡.¡¡ .. = = -- ......,.. .- - ,... ,_ .. _ ,.,., -" ...._ ~·--.t.-... .......... ~.cr.r iíil!t! - ..... ,. 1 """ ... .. «•e 1 -· · . - ll!lli.l*l"' ••• , p ...... ~ - ~'-.·\·: l"l' ~ ........ - ...... _.., . ~,.. - u 'fff$r-· --~~"-'· - -- .. ~ _,1 .... Figure 2 Genetic similarity tree of 40 Annonaceae accessions based on AFLP fingerprinting with the primer combinations .M and N. On the right side of the dendrograin,~the different subclusters found at a b·el of spnilarity of 0.55 (grey line, subclusters marked with G 1 - G9), and the possible species that ea eh of them represem . ...-r-A!)J'{"\' ~ 1 -~ 1 .----------------- --·· ~~;~:;"·=-' G 1 A. glabra '----------·------------~~.;;,~-~·-• 1 G2 A11110110 spp. ~ ¡L_-----------·-----1 !4=:AS!•·1 -.,s•· ' ~=- • G3 Rolli11ia spp . • \..;r UHbO-~ '--- \!:.P(,\!J -U 1\SP \ -4 t.- 1 -------1._. - --tce=-----...... ·~. _- ---~~~:¡;;;~·· 1 G4 4 1110 1/({/1/{/ - ~ A:\10!\! \'!~1-1 - . ----: ... ~··-~ ~ -A~!O~ !15.'...! '----:U !UN \ ',44 L_---------------------L~-~~:~.= ~· ::~ 1 G5 A . III III'ÍC(f(G l'h::-..:no,..,_¡ 1[ r ----r---·- ~- -~~;.:·~=;~~;; 1. G6 A. purpuren L¡--¡ _ ___::=~======·~~~ :~: G7 A. reticulaw l [ ---~---:~JE~:i 1 GS A. cherimola ~~; ;~;:~·· ---e=. ~. t\SP \ 4J.1J 1 '--------ASP\~ 1·3 --------,,sr \'1()..1 G9 A. squamosa ,----C" \S!' \ ji•l -:! L.,.:A.SP ft4-l Anc.ml Uianc:t- l ,... 1 __,~:::¡::==¡::==;;:=:;::::=::r=:;=:::;===;:=:::1 = r==r==r=:;=:::;:::::::=;=:;==;:=;=::;< .a\lmu 0.00 0 . .:!5 0.50 0.75 100 'ei-Li Sumlamy Analysis of the genetic variability of the soursop accessions The 37 accessions of soursop presented similarities between 0.81 and 0.95. This similarity is higher than that observed in the subclusters of other Annonaceae 3, 4, 7 and 8 (corresponding to Rollinia spp., A. montana, A. reticulata, A. cherimola andA. squamosa respectively). No 100% similarity was found among the different soursop accessions analyzed, although they included two selected clones and seedlings derived from them (AMUR-H 3-1 and AMUR-H 2-1, derived from AMUR-H3 and AMUR-H2 respectively). This indica tes that no duplicares are found among the different accessions. :\t a leYel of similarity of 0.82 two subclusters can be obsen:ed . The subcluster 1 includes two accessions, the 6-3 of Palmira-Valle and the 2039-2 of Sonso-Valle. The subcluster 2 includes a total of 35 accessions of Venezuela (2014-2), Costa Rica (AMUR-H 3-1), the Colombian departments of Magdalena (AMGR-M5~2), Ancioquia (AMUR-A 7-2), Cundinamarca (201 5~ 1 ), Caldas (1918-5), Quindio (1946-1 0), and Valle del Cauca (the rest of accessions). There was no ob,-ious relation between the origi.n of the accessions and the le\el of similarity presenrcd by them. One explanatioo of this results can be that the accessions are not originally from the sites where they were collected, rather seeds have been transponed by man O\er the centuries. 23 The accessions of soursop showed very low similarity (0.22) with the next most similar subcluster G4 (A. montana). With other subclusters the similarity was 0.19 (G1 and G3) or less (0.16; G6, G7, G8 and G9). This may explain why interspecific crosses of soursop with A . squamosa, A. glabra, A. montana, Rollinia mucosa, or A. retiCIIlata have not yielded viable and fertile hybrids (Nakasone and Paull, 1998; Samuel et aL 1991; Mohd-Kahlid, 2002). This suggest that the efforts in breeding of this species should be concentrated in the germplasm available within the species. Other species not included in this study, may be genetically closer to soursop, and these could possibly be used in soursop breeding. The available variability of soursop at the interspecies level in the germplasm bank can be exploited to produce sexual hybrids from whose progeny, indiv--iduals 'vith superior agronornic performance or fruit quality, can be se1ected. The results of the present study stress the need to collects species with low representation in the germplasm bank, as well as of other species not yet represented in the collection. 4.2.2.2 Proposed taxonorníc classification ofthe different accessions analyzed in this study The taxonomy of most of the accessions included in this study is clear and doesn't lea\e room for misclassifications for most of the subgroups. However, according to our results, a few accessions were incorrectly classified in the bank, (highlighted in table 1). In Table 1, we presenta proposed classification of the accessions analyzed in this study according to AFLP sirnilarity data. For sorne of the accessions, this classification has to be confirmed with other classical taxonomic techniques (underlined in Table 1). Figure 3. Genecic similarity between A. mun'cata accesions, based on AFLP fmgerprinting with the primer combinations F and M. The accession ASP-GV34 was used asan outgroup in the analysis. Accessions 1920-1 (Ro/lima spp.), 1959-2 (A. montana), and 1958-3 (A. glabra) were previously characterized as soursops. However, this tree does not group them with soursops. l j ---1 ! \ l o.:o _1 040 060 o.so Similaridad de Nei·LI 19l$·S lO:lG-1 19S1-3 IIP<-l rr~ ~ ~ l·l liD-1 1913-D :1011·3 ~ Lf ....--ID·I ID·I ~ liD-9 19tH lliQS4 lOOikl :lfiD.I ~UtVI ~UR!Ii IJII>$,1 ~ID-1 ~tli F variation (mean square) Isolates 58 4758.9619** 0.0001 Coefficient of variation % 40.89736 •. ANOVA was performed on a single experiment; 59 rreatments were inoculated m a soursop cultivar, arranged in a randomized complete block design experiment with four plants by treatments and 4 replicares; 3 controls (inoculated with water). b Mean square obtained by area under disease progress cutYe (AUDPC) ** and NS represent significant effects at the P5: 0.01 and no significance at P ,- -~ ~·rt~ >··i7·t:· DF~~- . . . DAilmail Figure 4. Percentage of access to means of communication identified by users interested in forming part of the Thernatic Network on Tropical Fruits. Figure 4 shows the channels, or means of communication, through which information reaches users. 4.4 Fruit flies in two departments of Colombia. CIA T's decision to add tropical fruits to its commodity portfolio has stirred interest in analyzing the possible arthropod pest problems that might be associated with regional fruit production. The Andean region is characterized by numerous agroecosystems and an impressive diversity of crops, and this is especially true of tropical fruits. Each fruit species will have a particular pest complex associated with it. At this time, since no specific fruit species has been designated as the priority species, it was decided not to concentrate on a particular species. There are, however certain groups of pests that consist of a species 56 complex that can attack and damage numerous fruit species. The fruit fly complex IS certainly an example of a pest that can damage numerous fruit species. In Colombia, fruir flies are a serious problem and are found in nearly all of the fruir growing regions of the country. They are especially imponant and can cause considerably economic loss in the fruit e}..-port indusrry. In Latin ~-\merica, about 20 fruit fly species have been reported causing losses calculared at abour 25 million dollars per year. The objectives of this initial study are to: 1. Establish a reference collection of fruit fly (Anastrepha spp) from the fruir growing regions of Valle del Cauca, Tolima and Quindío Departmenrs of Colombia. 2. Sample and identify fruit hosts and tbe associared fruir fly species in the region. 3. D eYelop laboratory rearing methods to eyentually study rhe biology and behavior of these species. 4. Iniciare a literature search to establish a darabank of present knowledge on fruit flies in rhe regions. 4.4.1 Literature Search Fruit flies belong to the Order; Díptera and tbe Family: Tephritidae. Worldwide, approximately 4000 species have been described and 400 species are reported from the Americas (Núñez; 2000). In Colombia, the m ost important species belong to the genera Anastrepha, Toxotrypana, and Ceratitis. Of the three genera, Anastrepha is considered to be tbe most important econornically, owing ro che considerable damage ir causes on different fruit species throughour the continent (Caraballo, 2001). The origin of this genus (Anastrepha) is the neotropics and it consists of more than 200 described species, of which, four are considered most importanr economically, Anastrepha striata Schiner, in guayaba; A . fraterculus (Wiedmann) in peach, mango, plum and others; A. obliqua (Macquart) in mango and plum, and A. serpentina (Wiedmann) in níspero (persimmon), caimito (star-apple) and other sapotaceous fruits. In addition two species, Ana.r/repha pickeli and A . manihoti, attack cassava - ~- Figure 3 Figure 4 Figures 3 and 4. Larval development, pupation chambers for fruit flies (.Anastrepha spp); chambers contain humid sterile soil fot pupation. fruits (and under certain conditions, cassava srems) bur they are not considered as economically important. 57 Many Anastrepha species may be host specific; others will attack host plants within the same family. Examples of the latter include A. grandis attacking Cucurbitaceae, A. oblique attacking Anacardiaceae, A . setpentina on Sapotaceae, A. striata on Mirtaceae and A pallidipennis on Passifloraceae. Generalist species such as A. suspensa, the Caribbean fruit fly, A .fraterculus, the South American fruit fly andA. ludens, the Mexican fruit fly attack more than 60 diverse species. These species may also have numerous \\ril.d hosts that have not yet been identified. There are 15 Anastrepha species recorded in Colombia, usually found between 15 to 29°C and Table 1. Sites sampled in the Department of Tolima, Valle del Cauca and Quindío {Colombia} for fruit fly {Anastree.ha SEE} infested fruit. Department M unicipali ty Locality H ost Date Tolima Ibagué M. indica 28 - Ylll - 02 Tolima !bagué P. domutica 28 - \'III - 02 Valle del Cauca Cerrito Sta. Elena P.gttqyava 12 - IX - 02 Valle del Cauca Bolívar Plaza Vieja c. pap':)'O 24- IX - 02 Valle del Ca u ca Bolívar San Fdo. P. gucryava 24 - IX - 02 Valle del Cauca Bolívar San Fdo. A. dJiriml!)'a 24- IX- 02 Valle del Cauca Bolívar San Fdo. M. indica 24- IX- 02 Valle del Cauca Palmira CIAT P.guqyava 02 - X- 02 Valle del Cauca La Cumbre C. maxima 02- X- 02 Valle del Cauca Candelaria Cavasa C. papqya 17- X- 02 Valle del Cauca Candelaria Cavasa P. quadrangularis 17- X- 02 Valle del Cauca Candelaria Cavasa C. pubescens 17- X- 02 Valle del Cauca Candelaria Cavasa P. gucryava 17- X - 02 Valle del (auca Candelaria Cavasa A. muricata 17 - X - 02 Valle del Cauca Candelaria Cavasa A. chiriml!)'a 17- X - 02 Valle del ( auca Candelaria Cavasa M. indica 17 - X- 02 Quin dí o Montenegro Va raya P. domistica 26- IX - 02 Quin dí o Montenegro Varaya M. cordata 26- IX - 02 Q uin dí o Q uimba ya El Laurel P.gucryava 26 - IX- 02 Quin dí o Quimbaya Querman M. esculenta 26- IX - 02 Quin dí o Circasia La Cabaña P.gucryava 26- IX - 02 Quin dí o Circasia Barcelona Baja P.guqyava 26- IX - 02 Quin dí o Armenia La Primavera M. esculenta 28- IX- 02 from sea level to 2000 m.a.s.l. (Portilla, 1994). Adult Tephritidos are about the size of a housefly and characterized by various colors, but predominandy yellow and translucent wings with longitudinal or transverse sports and bands. Adults live 1 to 3 months and females sexually mature in 3 to 4 days, copulating frequendy (Portilla, 1994). Their biological development is influenced by humidity, temperatw:e, lighr, nati,-e 'egetanon, pupanon and m·ipositional subso:ate and food anilability. Eggs of Anastrepha spp are usually a pale white, transparent and oviposited indi,-idually. :\ fully de,·eloped egg is opaque and the first instar larne are eYident before hatching. Tbe larvae are wedge-shaped with a rounded posterior. They are usually cream colored to yellow, but color can be influenced by food. There are three larnl instars before pupation. Pupae are 1.4 to 1.8mrn long and light straw colored to dark brO\\n. After copulating, females O\Íposit v;rithin the host fruit and the emerging hm·ae pass their instars feeding on the fruit pulp. The third instar larYae leave the úuit and pupates in the soil. .Adults emerge within severa! days. The larval phase can nry from 13 to 28 days and pupae duration is 14 to 23 days. The prem·iposition period is approximately 13 days; females can deposit 1 O to 11 O eggs per batch in fruit. Third instar fruit fly larvae upon emerging from the fruit will pupate in the soil; pupa were remm·ed, washed in distilled water and placed in glass gars, also containing sterile soil, where adults emerged (Figures 5 and 6). Adults were maintained on water plus bee honey solution for 2 to 3 days until the complete coloration for each species was attained. Those specimens separated for identification were placed in 60% alcohol; others were mounted on entomological pins and stored in the CIA T Arthropod Reference Collection. Basically, three morphological characters are used in the identification of fruir flies; these are the thoracic design, the wing design and the female OYÍpositor. Based on these parameters, the wings and ovipositors of females were mounted to facilitare identification. This was done by removing the wing of each specimen and m-ipositor and placing tbem on a glass slide with Hoyers media. Identification was done at the ICA (Instituto Colombiano Agropecuario) Laboratorio de Sanidad Vegetal in Palmira (Valle del Cauca). Figure 5 Figure 6 Figures S and 6. Fruit flies pupae collected from development chambers are washed in distilled water and placed in glass gars for adult emergence. 4.4.2 Results The fruits collected in Quindío and Valle del Cauca were mango, guava, papaya, cassan, chirimoya, plum, zapallo (calabash), sour-sop (guanábana), zapote (sapodilla), passion-flower (granadilla) and para)uela (Table 2). 229 specimens were collected from these fruits and this resulted in six separate Ana.rtrepha species (Table 2). The species A. striata was collected 59 from guava in several localities in Q uindío and Valle del Cauca. There are several other species reported from guava from tbese regions induding A. fraterculus, A. oblicua and A. ornate. The fact that only A. striata was collected from guava may have sometbing to do with the timming of tbe collections, September to October 2002. Tiús supports the need to sample fruits throughout they year in arder to determine if seasonality exists for the different Anastrepha species and the time of fruit infestation. T able 2. Anastrepha (fruit fly) species collected from severa! hosts m Department of Tolima, Valle del Cauca and Quindío, Colombia (Sept. to Oct. 2002). Code H ost Department Municipality ldentification 01 M. indica Tolima !bagué Anastrepha oblicr1a Macquat 7<;f 6 o 02 03 04 os 06 07 08 09 10 11 12 13 14 P. dome.rtica Tolima lb agué P. gucryava Valle del Cauca Cerrito P. gucryava Valle del Cauca Bolívar P. gutryatJa Valle del Cauca Palmita C. maxima Valle del Cauca La Cumbre P. gutryava Valle del Ca u ca Candelaria P. doméstica Quindío Montenegro M. cordata P.guqyava M. escufenta P.guqyava P. gutryava M. esculenta Quin dí o Quin dí o Quin dí o Quin dí o Quin dí o Quin dí o Montenegro Quimba ya Quimba ya Circasia Circasia Armenia Anastrepha .fraterculus Wiedmann 11 <;f 10 o Anastrepha striata Schiner 14 <;f 21 o Anastrepha striata Schiner S <;f 7 o Anastrepha striata Schiner 136<;f 1SSo AnastrephagrandisTtochez 2<;f S o Anastrepha striata Schiner 3 ~ S o Anastrepha striata Schiner Oblicua 2<;f so Anastrepha nunezae Steysca120~ lOo Anastrepha striata Schiner 1 <;f Anastrepha pickeli Litna 3 <;f 4o Anastrepha striata Schiner 4 <;f 7 o Anastrepha striata Schiner S~ 7 o Anastrepha pickeli Litna 3 <;f 4o The identifying morphological characteristics of four of the collected species are shown in figures 8 to 11. A. Anastrepha striata B. Ovipositor C. Wing pattern Figure 8. This species, known as the guava fruit fly, primarily attacks fruit of the Mirtaceae family but may also infest mango and sour orange (Citrus aurantium). A. Anastrepha grandis B. Ovipositor C. Wing pattern Figure 9. At present, A. grandis is not considered of major economic importance in Colombia as it mostly attacks cucurbitaceous (i.e. watermelon). It is considered as a quarantine pest in Argentina and Uruguay and may eventually have greater importance in Colombia. 61 A. Anastrepha pickeli B. Ovipositor C. Wing Figure 10. This species has only been found attacking cassava fruits (and stems). When infesting cassava stems it can severe rotting due to the invasion of soft rot bacteria Erwinia caratavora. Tbe latter causes a reduction in the quality of planting material (stem cuttings). A. Anastreplza nunezae B. Ovipositor C. Wing ~-. -~ Figure 11. A. nunezae was found infesting zapote (Quararibaea cordata) especially between 900 to 1700 m.a.s.l. 4.5 Towards New Research Policies for Fruit Crop Research The Colombian Ministry of Agriculture and Rural DeYelopment (MADR) requested assistance from CIA T in designing research policies to support tropical fruit based enterprises. Working closely with the l\1ADR and CORPOIC\ we have developed guidelines for more effeccive use of research resources on Tropical Fruits. Although the policies are directly related to the Colombian situation, many of the recommendations are highly relevant to many other developing countries. Fruit producers and all the actors who form part o f the produccion chain 1 depend to a greater or Iesser extent on good technology in order to be competici,·e, parcicularly when tl1eir efforts are dedicated to export markets. Good technology is necessary to improYe productivity and tl1e quality of the final product and to maintain commercial enterprises in the face of new threats. Furthermore, effecci\e technology depends on continued research support. Thus, for any agro-business to be competitive it must be supported by research and technology development: fruits and \egetables are no excepcion to this rule. Although research is sometimes seen as just another expense, there is no doubt that investing in agricultura! research is highly profitable, providing economic benefits for the actors in the produccion chain, and social benefits in terms of lower prices and improved quality for the consumers. A meta analysis of the interna! tates of return indicated that these are high in all countries studied 2. In spite of these high rates of return, research in,·estment has been much lower in the developing countries than the developed countries in the past few decades. The research intensity3 in tl1e agricultura! sector is approximately 0.62 in the developing countries (and only 0.14% in Colombia), whilst it is about 2.6% in the developed countries4. It is probable that the low competitivity of Colombian producers is at least partially due to extremely low research intensity in agriculture over many years, and without doubt very little research has been carried out in the horticultural sector A marked difference between the developed and the developing countries is the lack of private sector research investment in the developing countries. Recently it has been suggested that private non-profit organizations could be an interesting altemaci,·e to close the gap berween prívate and public sector research in tl1e developing countries. In the case of Colombia this is already occurring with a considerable proportion of the total research effort being carried out by the CENis (Nacional Research Centres) financed by para fiscal funds obtained by levies on specific agricultura! products. The CENis now form an integral part of the nacional agricultura! research system. The CENis only support those proouction 1 The production chain defined as the "Group of economic agents who participare directly in the production, transformation and mo,ement ro the markets of rhe same agricultura! product". Duruflé, Fabre r Young. Translated to Spanish by IIC..-\) www.prompex.gob.pe/ prompex/ lnf_Secrorial/.A.gro/ PDFJ\L-\C:\ / PilarCoral.pdf 2 Julian M. Alsron, Connie Chan-Kang, l\ Iichele C. ?\farra, Philip G. Pardey, and TJ Wyarr. 2000. A ~Iera­ Analysis of Rates of Return to .-\gricultural R&D: Ex Pede Herculem?. Research Report 113 IFPRI, Washington, D.C..'\lston cr al 1999 3 Research intensity is deflned as research expenditure as a pcrcentage of the toral gross product value of rhe sector inrolved. 4 Philip Pardey and Nineke Bientema (2001 ) Slow Magic. www.ifpri.org/ pubs/ fp s/ fps36.htrn 63 4.5.1.2 New or Incipient Production Chains. The essence of the research and development in new or incipient production chains is to develop new options for the horticultura! sector. Currently there is tremendous interest in developing new options for export. Research and Development Characteristics. The initiative for developing new crops often comes from independent producers, however, the pioneers are often not able to appropriate a large proportion the benefits of their investment and the knowledge they have obtained. In other words they are producing a public good in the form of improved and accessible knowledge, which can be utilized freely by the rest of the population. For this reason private entrepreneurs are reluctant to develop totally new options on their own. In agriculture and horticulture the most successful crops are normally introduced species grown far from their centre of origin. Consequently there is a high probability that the most promising new options will be species from distant lands, and local development agencies and researchers are probably not aware of their existence. In addition, agencies that are looking for new options are accustomed to looking at what are considered promising local species, or those grown by neighbors. The research and development efforts geared to providing new options has to work sirnultaneously on two fronts: the research must determine what new products could be produced in a particular regions, and how those new products can be marketed. The oprima! strategy for obtaining information on potencial crops or species is by visiting areas similar to the target areas in terms of ecological conditions, but which are geographically distant. Once exciting prospects have been identified, exchange of germplasm and technical know will be necessary. On the other hand marketing may be a case of opening up totally new markets or of competing in previously established markets. Marketing of export crops will also have to be closely linked with quarantine problems of both importing germplasm, and eA'Porting fruit products with attention being paid to assessing risks. Implementation of the model • • • • It is suggested that CCI and other agencies continue to study the potencial for producting and marketing new products, and that theses agencies expand the horizons to include altematives from Africa and Asia. The marketing studies should be financed by ASOHOFRUCO, ONGs, official agencies such as MADR and Proexport, and internacional agencies. Universities and other research agencies such as CORPOICA and CENICAFE should assume the responsibility of looking for new crop options particularly well adapted to specific target regions of the country. The prospecting should be based on looking for altematives that grow well in areas that are ecologically similar to the target areas The MADR, Colciencias, Asohofrucol and Proexport should finance these initiatives as long term development efforts. • 4.5.1.3 The Centre of Phytosanitary Excellence should conrinue in its role of risk assessment broadening its scope to include a '\VÍder range of options, particularly those from other continents. Technolo!!J! E>..port The possibility of exporring fruit technolog:, as opposed to fruits themsleYles, is a nm'el and probably polemical idea for the developing countries. Nevertheless, there are two incontrovertible facts: the majority of important export crops are cultivated more widely outside their centre of origin, whilst at the same the basic germplasm required for their de,-elopmem resides principally in the cemers of origin. At present Colombia imports technology in the forro of improved seeds, a wide range of purchased inputs and most of the machinery for production and processing. ~-\t presem there is little incentive for rhe country to collect and de,-elop its germplasm if the likely beneficiaries are farmers ins other parts of the world. The question them becomes one of how to ensure that the exchange of knowledge aod germplasm can be made economically attractive for those in the centre of origin who collect, and irnprove germplasm, and for those who use the technology for production. It appears that there may be a possibility to exploít local germplasm and biodíversity by impro,-ing it and exporting the results. The policies required to achieve this are still being de,-eloped. 71