Kerr yield models for cotton Were applied to two cotton groups to determine the relative contribution of various yield components to lint yield/square meter and lint yield/plant (LY/P). The first group included the four Egyptian cultivars Ga 80, Giza 66, Dendera and Giza 75.
The second group ilud.ed the two Upland cultivars Mcnaire 220 and Stonevili. L3. Path coefficient results indicated that in both groups the number of boils per square meter and number of boll per plant B/P were the main contributors to lint yield variations per unit area and per plant, respectively. Second in importance were seed per boll and lint per seed. The joint effect of charcters contributing to yield of lint as indirect effects were very small. Similarly, the residual effects were trivial in magnitude, indicating that the main biological units contributing to lint yield were included in the models. Thus, in conclusion selection
to improve lint yield of cotton should be directed toward the prolificacy of B/ma or B/P and S/B or LIS in the breeding materia).