Euphytica 29 (1980) 673-683 BREEDING CASSAVA FOR RESISTANCE TO CASSAVA MOSAIC DISEASE S. K. HAHN, E. R. TERRY and K. LEUSCHNER International Institute of Tropical Agriculture (IITA), P.M.B. 5320, Ibadan, Nigeria Received 10 January 1980 INDEXWORDS Man&or esculenta, cassava, Bemisia tabaci, whitefly, cassava mosaic disease, screening, mechanism of resistance. SUMMARY Cassava mosaic disease (CMD) is one of the most serious and widespread diseases throughout cassava growing areas in Africa, causing yield reductions of up to 90%. Early research on breeding of cassava (Manihot esculenta CRANTZ) for resistance to CMD in Africa is reviewed. Changes in population size and in activity of the white-fly vector of CMD (Bemisia rabaci GENN.) in relation to changes in environmental conditions such as amount and distribution of rainfall, light intensity and temperature are discussed in relation to screening for resistance to CMD. Over the past eight years, significant progress has been made at the International Institute of Tropical Agriculture (IITA). Resistance to CMD has been successfully incorporated into high yielding cultivars of acceptable quality. The CMD resistant material has been evaluated and many promising clones have been selected in various countries in tropical Africa and India. The resistance has been effective in those countries INTRODUCTION Cassava is grown widely as one of the most important staple food crops in the tropics, where it is adapted to diverse environmental conditions and farming systems. It is the seventh most important food crop throughout the world (HARLAN, 1976). Both tu- berous roots and leaves are used, the former as a major source of carbohydrates and the latter as a vegetable and source of vitamins, minerals and protein. Cassava mosaic disease (CMD) first reported in East Africa by WARBURG in 1894 (c.f. STOREY & NICHOLS, 1938) is the most widespread cassava disease in tropical Africa and India; it is not reported from Latin America where the genus Manihot is in- digenous. Cassava is believed to have been first introduced by the Portuguese in the delta of the Congo River and to the Guinea Coast during the last part of the sixteenth century. It was probably introduced to East African countries bordering the Indian Ocean during the eighteenth century (JONES, 1959). Annual losses due to CMD as high as 11% have been reported (PADWICK, 1956); and yield losses for individual susceptible cultivars have been mentioned to range from 20 to 90% (BECK&CHANT, 1958; BRIANT &JOHNS, 1940). STOREY & NICHOLS (1938) demonstrated that the CMD agent is carried in cuttings of diseased plants, and is transmissible through a graft and also by means of an insect vector, Bemisia spp. The causal agent has not been identified, but virus etiology is suspected (STOREY & NI- CHOLS, 1938; BOCK & GUTHRIE, 1976; ROSSEL& THOTTAPPILLY, 1978). In traditional 673 S. K. HAHN, E. R. TERRY AND K. LEUSCHNER African agricultural systems where inputs are low, the use of infected propagating material is common, so ensuring the perpetuation of the disease. The development of disease resistant cultivars is the only means of effective control. Although several research groups have attempted to breed cassava for resistance to CMD, only the research in Tanzania (East Africa) and Nigeria (West Africa) will be discussed. In East Africa, breeding for resistance to CMD was initiated by Storey in 1937 (NICHOLS, 1947). Low levels of resistance were found in the cultivated species, Manihot esculenta, from East Africa and other cassava growing countries. However, progenies of the third backcross of ceara rubber (Manihot glaziovii MUELLER VON ARGAU) to the cultivated cassava (M. esculenta) showed promise of good CMD resistance (NICHOLS, 1947). Moderately resistant clones with reasonable yields were selected from the interspecific crosses, after which the research was discontinued in 1958. NICHOLS (1947) stated that since immunity appeared to be unattainable, the most viable alternative would be to secure a high level of resistance. In the event that resistance breaks down, selection for tolerance would be the next line of action. In Nigeria, cassava breeding was initiated by Faulkner in 1932, discontinued in 1939 and resumed again in 1953 by Vernon. Investigations were continued until 1960 by Beck, who introduced seeds from Storey’s third backcrosses of the interspecific cross and identified the clone 58308 as resistant in 1958 (EKANDEM, 1970). Although the breeding work was discontinued in 1961, the resistant clone 58308 was maintained and has shown resistance for over 20 years under heavy CMD pressure. Tuberous root yield from this clone, however, was poor, both in quantity and quality (HAHN et al., 1973). The International Institute of Tropical Agriculture (IITA), located at Ibadan, Ni- geria, established the Root & Tuber Improvement Program in 1971 and assigned high priority to cassava. The breeding strategy was to incorporate disease resistance into susceptible but well adapted local cultivars. A previous paper (HAHN, 1978) reviewed our work on breeding for bacterial blight resistance. This paper reviews the recent work on breeding for resistance to cassava mosaic disease at IITA. VECTORANDENVIRONMENTALEFFECTONINCIDENCEANDSEVERITYOFCMD CMD incidence depends on the availability of inoculum, which in turn depends on the density and activity of the vector (Bemisia tabaci) population. Incidence is closely related to the number of whiteflies trapped (Fig. 1) (LEUSCHNER & TERRY, 1976). Populations of whiteflies change with environmental conditions such as rainfall distri- bution (Fig. l), light intensity and temperature. Whitefly population density increases as rainfall increases up to 280 mm per month. The activity of whiteflies was high at the temperature regime of 27-32°C but was low outside this range (Fig. 2). The activity of whiteflies was low when light intensity was low. The environmental conditions which apparently favour population build-up and activity of whiteflies are rainfall between 150-280 mm per month, temperatures within the range 27-32°C and solar radiation of 400 g-Cal/cm’ (LEUSCHNER, 1978). There are interactions between rainfall, temperature and solar radiation with chan- ges in density and activity of whitefly populations. There appears to be no significant difference between susceptible and resistant cultivars in number of adult whiteflies 674 Euphytica 29 (1980) CASSAVA MOSAIC DISEASE RESISTANCE Jlh’J. JUL. AUG. SW. OCT. MONTH OBSERVED Fig. 1. Monthly changes of and relation between incidence of cassava mosaic and number of whiteflies trapped and rainfall distribution (after LEUSCHNER & TERRY, 1976). observed on them in the field (Table 1). Temperature affects CMD symptom expression; a high temperature (35°C) sup- presses symptom development (CHANT, 1959; TERRY, 1979). CMD incidence was altered by lime application; 0.5-1.0 t/ha increased CMD incidence (EDWARDS & KANG, 1978). In the acid soil CMD incidence or severity is less. CMD incidence tends to increase as soil fertility decreases after clearing bush. Whether this is due to micro- nutrient imbalance and/or due to decrease in organic matter is not known. CMD incidence was lower during the dry season and in areas at higher elevation (above 500 m) and with lower (less than 900 mm per annum) and high rainfall (above 1500 mm per annum). CMD incidence decreases as the plant becomes older (from five to six months after planting). Table 1. Average number of adult whiteflies and pupae per fully expanded young leaf on susceptible and resistant cultivars (after LEUSCHNER, unpublished data). Period observed 60444 (susc.) Isunikakiyan (susc.) 58308 (res.) adult pupae adult pupae adult pupae Peak period 5.8* 7.0 3.3 10.5 4.8 8.5 Low period 1.1 3.4 0.8 1.7 1.0 3.0 *Data taken once a week and averaged over mean figures based on 7 observations for the peak period (30 May - 18 July 1976) and 22 observations for the low period (25 July - 26 Dec. 1976). Euphytica 29 (1980) 675 S.K.HAHN,E.R.TERRY AND K.LEUSCHNER 14-22 22-27 27-32 32-38 38-48 TEMPERATURE REGIME (“Cc) Fig. 2. Population size and activity of whiteflies in relation to temperature (after LEUSCHNER, 1978). SCREENING FOR RESISTANCE TO CMD Screening for resistance to CMD is based on phenotypic expression of symptom severity in the field relying upon natural infection by viruliferous whiteflies. Such field screening must be done in locations and seasons where inoculum from diseased cassava is present, where whitefly populations are high and where the average temperature is relatively low (below 30°C). Screening for resistance to CMD is more effective in areas with an annual rainfall of 1000-l 500 mm, with elevation less than 500 m, with average temperatures of about 20-25 “C, and with acid (about pH 4-6) and poorer soils. IITA is situated in an ideal area with most of the above conditions for effective screening for resistance to CMD. At IITA screening cassava breeding material for resistance has been practiced during the first (in early rainy) season (April-July) when the environmental conditions are most favourable for the vector in terms of its population build-up and activity, and for disease in terms of its infection and symptom development. The rating for reaction to CMD is based on five classes (Fig. 3) of severity, viz., class 1 -no symptoms observed; class 2 -mild chlorotic pattern over entire leaflets or mild distortion at the base of leaflets only with the remainder of the leaflets appearing green and healthy; class 3 - moderate mosaic pattern throughout the leaf, narrowing and distortion of the lower one-third of leaflets; class 4 - severe mosaic, distortion of two- 676 Euphytica 29 (1980) CASSAVA MOSAIC DISEASE RESISTANCE Fig. 3. Classes of cassava mosaic disease (CMD) severity: class 1 - no symptoms and class 5 most severe symptoms of CMD. thirds of the leaflets and general reduction of leaf size; and class 5 - severe mosaic distortion of the entire leaf. Leaf area of the fully developed leaves of two cultivars in relation to CMD classes is shown in Fig. 4. Leaf area was drastically reduced as the CMD score increased. BREEDINGMETHODS Since no adequate level of resistance to CMD was found within A4. esculenta, the breeding approach used for CMD resistance in the early stages by Storey and Nichols involved interspecific hybridization between cultivated cassava (M. esculenta) and other related Manihot species, particularly M. glaziovii, followed by backcrossing to cultivated cassava to recover the positive agronomic characters of cassava as well as resistance to CMD (NICHOLS, 1947; DOUGHTY, 1958). Euphytica 29 (1980) 677 S. K. HAHN, E. R. TERRY AND K. LEUSCHNER - TMS 30555. Y= 337.0-56.96x -x TMS 30395. Y= 441.2-66x "0 I ; 3 4 5 CMD SCORE Fig. 4. Leaf area of fully developed leaves of two cultivars in relation to cassava mosaic disease (CMD) incidence. Cultivar TMS 30395 is resistant to CMD and TMS 30555 is moderately resistant. DOUGHTY (1958) and JENNINGS (1976) attempted crosses among selected promising clones which were derived from the interspecific cross, with resistance to CMD and reasonable yields, in order to recombine genes for resistance to CMD and to in- corporate genes for desirable agronomic characters. At IITA, a recurrent selection system has been used to improve populations for CMD resistance and other agronomic characters, maintaining a large genetic variation. Resistance alone was improved in one cycle, taking l-2 years; but it took 4-5 years to combine it with a high yield potential. Introgression of exotic sources from other continents, especially Latin Ame- rica, into IITA breeding populations has been emphasized after achieving adequate resistance to CMD, envisaging that untapped genetic resources should be fully ex- plored for further improvement of cassava. GENETICSOFRESISTANCETOCMD Resistance to CMD was reported by NICHOLS (1947) to be genetically controlled and DOUGHTY (1958) suggested that it was polygenic, which was confirmed by HAHN & HOWLAND (1972) and HAHN et al. (1974) who found that it inherits largely in an additive way. From a seven parent diallel cross, it was concluded that CMD resistance 678 Euphytica 29 (1980) CASSAVA MOSAIC DISEASE RESISTANCE was recessive with a heritability of about 60% (HAHN et al., 1972). This recessiveness was confirmed by JENNINGS (1976) from a six parent diallel cross. It was reported by HAHN (IITA, 1972; IITA, 1973; IITA, 1977; HAHN, 1978, HAHN et al., 1980) that resistance to CMD is genetically correlated (r = 0.46-0.85) with resistance to bacterial blight disease (CBB). PROGRESS OF CMD RESISTANCE BREEDING AT IITA IITA’s cassava breeding program has utilized clone 58308 extensively as a source of resistance to CMD and bacterial blight. Clone 58308 had high breeding values for resistance to both CMD and CBB. Also, wild cassava from Nigeria which showed resistance to CMD, probably a natural hybrid between cultivated cassava and ceara rubber, which is native to latin America and has been grown in Nigeria for several centuries, was used in the crossing program in order to introduce an additional source of resistance to CMD. Several local Nigerian cultivars with some degree of CMD resistance were also used as sources of low levels of resistance; we presume these cultivars have evolved through natural selection following natural hybridization with- in M. esculenta and between M. esculenta and M. glaziovii. IITA-improved families exhibited superior resistance to CMD under IITA con- ditions, compared to those from Latin America, East Africa and India (Table 2). The Table 2. Comparison of cassava families (seedling) from various regions for resistance to CMD at IITA, Ibadan, in 1974. Region* CMD score Families % families Average tested class 1 and 2 score 1 2 3 4 5 IITA 5 91 170 40 2 314 32 2.80 E. Africa 0 2 12 19 36 69 3 4.29 CIAT 0 7 66 50 82 205 3 4.01 India 0 0 0 0 6 6 0 4.80 *Open pollinated seeds from different cultivars were introduced from various regions. IITA refers to the open pollinated seeds from IITA improved clones. exotic sources have been substantially improved for resistance to CMD over the years through alternating selection and recombination with IITA populations, while main- taining other desirable traits (IITA, 1973). The sources from Latin America (about 2000 accessions) and from India, showed a high susceptibility to CMD. Families from Zaire were as susceptible as the Nigerian local cultivar (Table 3). Crosses of exotic sources with IITA breeding populations showed a rapid improvement in terms of CMD resistance and yield. Yield of cassava was a function of CMD incidence (r = -0.98); as CMD incidence decreased yield increased remarkably (Table 3) sugges- ting that if the level of resistance to CMD is further increased, yield may accordingly increase. IITA CMD resistant breeding materials have been tested in many countries in West Africa, Central Africa, East Africa and India. The CMD resistance in all those places Euphytica 29 (1980) 679 S. K. HAHN, E. R. TERRY AND K. LEUSCHNER Table 3. Comparison of cassava cultivars from diverse sources for average CMD scores and yields at IITA, Ibadan, Nigeria in 1977. Source Clones tested** CMD score Yield (f/ha) IITA 2398 2.6 26.4 E. Africa x IITA* 60 2.7 25.6 Brazil x IITA* 218 2.9 24.6 Zaire 106 3.6 15.4 Local cultivar 1 3.4 12.3 India 83 4.3 0.40 Brazil 384 4.2 0.36 * E. Africa x IITA and Brazil x IITA are crosses of the clones derived from the seeds introduced from E. Africa and Brazil with IITA improved cultivars. **Single replication except of the local cultivar (Isunikakiyan) (n = 119). Correlation between CMD resistance scores and yield, r = -0.98. has been effective (IITA, 1973-78). This absence of regional differences in resistance and its polygenic nature suggest that the resistance may be durable, but whether or not it will prove to be race-non-specific depends on information on pathogenic variation. Over the past 8 years, signilicant progress has been made at IITA in producing improved cassava cultivars with resistance to two major diseases, CMD and CBB (HAHN et al., 1980). Improved cultivars are high yielding, have desirable root charac- teristics, are resistant to lodging and have low. HCN. The quality of cassava products has been tested and found acceptable. Improved cultivars out-yielded local cultivars by a factor of 2 to 18 times in Nigeria, Sierra Leone, Liberia, Gabon, Zaire, Zanzibar, Seychelles and India, especially under severe CMD and CBB epiphytotic conditions. The Nigerian unimproved local cultivar Isunikakiyan was compared with TMS 30395, an improved CMD resistant cultivar, for increase in CMD incidence as a result of vector inoculation. The rate of increase of CMD incidence over time fitted the equation Y = l/(a + b/X), where Y is percent plants with CMD symptoms and X is month after planting. It is assumed that CMD pressure through vector activity in- creases with time. The equations fitted to the CMD incidence data from the cultivars Isunikakiyan and TMS 30395 were Y = l/(0.0085 + 0.0062/X) and Y = l/(0.0348 + 0.1115/X) respectively (Fig. 5). The r2 values between the observed and estimated were 0.94 and 0.98 for Isunikakiyan and TMS 30395, respectively. The constant a may be a measure of the level of resistance to infection exhibited by cassava plants when exposed to CMD-carrying whiteflies. The estimated a values were 0.0085 for the susceptible local cultivar, Isunikakiyan, and 0.0348 for the resistant improved cultivar, TMS 30395, respectively, the latter being about 4 times higher than the former. RESISTANCE MECHANISM RUSSELL (1978) proposed six types of resistance to virus diseases: 1) immunity; 2) resistance to virus infection; 3) resistance to establishment and spread of virus in host plants; 4) resistance to virus multiplication; 5) tolerance; and 6) resistance to vectors. There were similar numbers of whiteflies observed on both susceptible and resistant cultivars. This indicates that it is not likely that there is resistance to the vector. There 680 Euphytica 29 (1980) CASSAVA MOSAIC DISEASE RESISTANCE Fig. 5. Performance of a resistant and a susceptible cassava cultivar to cassava mosaic disease (CMD). I I 2 3 4 5 6 MONTH AFTER PLANTING were remarkable differences among the cultivars in average CMD ratings ranging from class 1 to 5 and in percent plants without symptoms after six years of field exposure to high disease pressure by planting cuttings of about the same length (25-30 cm) each year (Table 4). However, there was no cultivar without CMD symptoms. Once in- dividual plants of both susceptible and resistant cultivars became infected, systemic invasion occurred and typical CMD symptoms were expressed. It is difficult to postulate the mechanism of resistance to CMD, particularly because its causal agent is not identified and experimental information on it is lacking. Ho- wever, the fact that resistant cultivars showed significantly lower CMD severity rating and small percentages of disease-infected plants for long periods of time under heavy disease pressure, even though systemic invasion of the disease occurred, suggests that Euphytica 29 (1980) 681 S. K. HAHN, E. R. TERRY AND K. LEUSCHNER Table 4. Percent plants of four cassava cultivars without CMD symptoms after six years of field exposure to high disease pressure (after E. R. TERRY, 1978). Cultivar Plants observed Symptomless plants number % TMS 30001 1789 1717 96.0 TMS 30395 2375 2273 95.7 TMS 30572 2826 669 23.7 lsunikakiyan 1050 0 0 the rate of disease spread is limited in plants of the resistant cultivars. As stated earlier, there was a remarkable difference between the resistant and susceptible cultivars in the rate of infection by viruliferous whiteflies. This indicates that an important aspect of resistance to CMD is resistance to vector infection and another aspect is resistance to spread of CMD within plant. Whether resistance to CMD is due to physical barriers or antiviral factors produced in response to infection needs to be investigated. CONCLUSION Sources of resistance to cassava mosaic disease (CMD) are available and have been successfully incorporated into high yielding cultivars of acceptable quality. The re- sistance to CMD has been effective in many countries in tropical Africa and India. This absence of regional differences in resistance and its polygenic nature suggest that resistance may be effective for a long time in several localities. The resistance was originally derived from a related Munihot species. Cassava is an allotetraploid and genetically a heterozygous plant. This would further support that resistance to CMD may be durable over localities and years. Resistance to CMD appears to be resistance to vector infection and to spread of the pathogen within the plant. ACKNOWLEDGMENTS The authors thank Dr D. J. Allen, Dr I. W. Buddenhagen, Mr H. R. Rossel and Dr G. Thottappilly for their useful suggestions and comments. The authors particularly express their sincere appreciation to Mrs A. K. Howland for her tireless efforts and dedication toward cassava breeding at IITA during 1972-l 977. The authors also thank Dr J. E. Parleviet for reading the manuscript and his useful suggestions and comments. REFERENCES BECK, B. D. A., 1971. The breeding goals in a cassava breeding program in West Africa. Paper presented at the Guinea Zone Conference, University of Ibadan, Nigeria. pp. 5. BECK, B. D. A. & S. R. CHANT, 1958. A preliminary investigation of the effect of mosaic virus on Manihot utilissima POHL in Nigeria. Trop. Agric. (Trinidad) 35: 59-64. BOCK, K. R. & E. J. GUTHRIE, 1976. Recent advances in research on cassava viruses in East Africa. Report of an Interdisciplinary Workshop held at Muguga, Kenya, IDRC-071e. BRIANT, A. K. & R. JOHNS, 1940. Cassava investigations in Zanzibar. E. Afr. Agric. J. 5: 404412. 682 Euphytica 29 (1980) CASSAVA MOSAIC DISEASE RESISTANCE CHANT, S. R., 1958. Studies on the transmission of cassava mosaic virus by Bemisia spp. (Aleyrodidae). Ann. Appl. Biol. 46: 210-215. DOUGHTY, L. R., 1958. Cassava breeding for resistance to mosaic and brown streak viruses. 1958 EAAFRG Annual Report. pp. 48-55. EDWARDS, D. G. & B. T. KANG, 1978. Tolerance of cassava to high soil acidity. Field Crops Res. 1: 337-346. EKANDEM, M. J., 1970. Cassava research in Nigeria before 1967. Federal Report of Agric. Research, Ibadan, Memo No. 103. pp. 16. HAHN, S. K., 1978. Breeding cassava for resistance to bacterial blight. PANS 24(4): 480485. HAHN, S. K. & A. K. HOWLAND, 1972. Breeding for resistance to cassava mosaic disease. Proc. Cassava Mosaic Workshop, Ibadan, Nigeria. pp. 37-39. HAHN, S. K., A. K. HOWLAND & C. A. OKOLI, 1974. Breeding for resistance to cassava bacterial blight at IITA. Proc. First Workshop on Cassava Bacterial Blight in Nigeria. Umudike, Nigeria. pp. 1 l-14. HAHN, S. K., A. K. HOWLAND & E. R. TERRY, 1973. Cassava breeding at IITA. Proc. Int. Trop. Root Crop Symp., Ibadan, Nigeria. pp. 410. HAHN, S. K., A. K. HOWLAND & E. R. TERRY, 1980. Correlated resistance of cassava to mosaic and bacterial blight diseases. Euphytica 29 : 305-3 11. HAHN, S. K., E. R. TERRY, K. LEUSCHNER, I. 0. AKOBUNXJ, C. OKALI & R. LAL, 1979. Cassava improvement in Africa. Field Crops Res. (in press). HARLAN, J. R., 1976. The plants and animals that nourish man. Scientific American 253(3): 89-97. IITA, 1973-78. Annual Reports, IITA, Ibadan, Nigeria. JENNINGS, D. L., 1976. Breeding for resistance to African cassava mosaic. African cassava mosaic, Report of an Interdisciplinary Workshop held at Muguga, Kenya. IDRC-07le, pp. 3944. Jo-, W. O., 1959. Maniac in Africa. Stanford Univ. Press, Stanford, pp. 60-87. LEUSCHNER, K., 1978. Whiteflies: Biology and transmission of African mosaic disease. Proc. of Cassava Protection Workshop, CIAT, Cali, Colombia. pp. 51-58. LEUSCHNER, K. & E. R. TERRY, 1976. IITA Annual Report. IITA, Ibadan, Nigeria. NICHOLS, R. F. W., 1947. Breeding cassava for virus resistance. E. Afr. Agric. J., 12: 184-194. PADWICK, G. W., 1956. Losses caused by plant diseases in the colonies. C.M.I. Phytopathological Papers, No. 1. pp. 60. ROSSEL, H. W. & G. THOVAPPILLY, 1978. IITA Annual Report. IITA, Ibadan, Nigeria. RUSSELL, G. E., 1978. Plant breeding for pest and disease resistance. Butterworts, London. pp. 209-229. STOREY, H. H. & R. F. W. NICHOLS, 1938. Studies on the mosaic diseases of cassava. Ann. Appl. Biol. 25: 790-806. TERRY, E. R., 1978. IITA Annual Report. IITA, Ibadan, Nigeria. Euphytica 29 (1980) 683