Abstract

Variation, correlations and path coefficient analysis were calculated for root yield among 19 improved cassava clones for two harvesting periods over two years of evaluation in a randomized complete block design with three replications. Root width and leaf length were significant at 6 months harvesting, but not at 9 months harvesting. Estimates of phenotypic, genotypic and environmental variance for both harvesting periods provided evidence for genetic improvement. Heritability was low for root length, meaning that environment had a strong influence in this phenotype. The path coefficient analysis for data set at 6 months harvesting indicated that leaf width recorded the largest direct effect on tuber yield; this was masked by the indirect influence through cassava anthracnose disease, plant height, root length, percent starch, petiole length and top weight. The incidence of cassava anthracnose disease manifest in a significant negative correlation coefficient with root yield. The large positive direct effect of tuber/plant was masked by the indirect influence of petiole length and top weight. The dry matter yield had a positive direct effect on yield, but this was indirectly lowered by cassava anthracnose disease incidence and percent starch. At 9 months harvesting, leaf width had the largest direct effect on tuber yield, this was masked largely by indirect influence of percent dry matter. This study summarized the influence of leaf width, leaf length and tuber/plant as important selection indices for root yield at 6 months. The path analyses for both sampling months demonstrated that leaf width, leaf length, and tuber/plot are important selection indices for root yield.

Key words: Harvest periods, tuber yield, cause and effect relationships, heritability, path analysis and cassava.