Yield stability studies in some sweet potato (Ipomea batatas(L) Lam) cultivars in Kenya

Abstract

Sweet potato is an important food crop in Kenya not only because it is a good source of carbohydrates, vitamins and minerals but also because it fits very well in the existing cropping system in Eastern Province of Kenya which is mainly semi-arid except for some hill masses which receive more rainfall. There is however only limited information on the adaptability and performance of either the local or improved cultivars over a wide range of environments. A study on yield stability was therefore undertaken to identify the cultivars having specific or general adaptation. The study was conducted during short rains 1986-1987 and long rains 1987 in eastern Kenya. Ten varieties, six of which were exotic introductions, three local land races and one hybrid selection made at Katumani were grown at five environments, Katumani, Kitui, Kampi ya Mawe, Kabete and Kiboko. Of these environments, Katumani, Kitui, Kampi ya Mawe and Kiboko are in agroecological zones 4' and 5 receiving limited and irregular rainfall. Kabete is in agroecological zone 3 and receives adequate rainfall. The experiment was planted in a randomized complete block design with four replications. The stability for yield was studied usLng the methods developed by Finlay and Wilkinson (1963) and Eberhart and Russell (1966). Genotype x environment interaction for tuber weight, vine weight, mean number of tubers per plant, mean vine length, percent dry weight of vines and tubers were highly significant. The genotype x season interaction was also highly significant showing the importance of testing the genotypes over different seasons. The genotype x season x environment interaction was highly significant for tuber weight vine weight, mean number of tubers per plant, mean vine length and percent dry weight of tubers and vines. Using Finlay and Wilkinson's (1963) and Eberhart and Russell's (1966) stability models it was possible to categorize the different varieties into different groups using the stability parameters. Varieties KSP3, KSP17, KSP23, KSP21 and KSP102 had average stability for tuber yields with regression coefficients of approximately 1.0. Varieties KSP15, KSP20 and KSP97 had above average stability wit regression coefficients above 1.0 while varieties KSP11 and KSP135 had above The exotic cultivars were relatively higher yielding than the local land races and had better tuber yield adaptation over a wide range of environments compared to the local varieties. The exotic varieties KSP21, KSP17, KSP102 and KSP3 showed good general adaptation while some exotic varieties KSP15 and KSP20 showed specific adaptation to the better endowed environments since they had below average stability. Both the genotypic and phenotypic correlation coefficients between tuber weight and mean number of tubers were positively correlated. The correlation coefficient between vine weight and mean vine length to fresh tuber yield was either very small or negative indicating that these two traits only played a negligible role in fresh tuber yield production. The direct path values between mean number of tubers per plant and fresh tuber yield were high confirming the role of mean number of tubers per plant as a major determinant of final tuber yield in sweet potato cultivars.