MULTIVARIATE ANALYSIS OF YIELD AND RELATIVE CONTRIBUTION OF VARIABLES TO ITS VARIATION IN COTTON YIELD

The stepwise multiple linear regression analysis was used to study the
relationship between the variables using two models. While the first one seed cotton
yield (ken.lF.) is considered the dependent variable. In the second one lint yield
(ken.lF.) is considered the dependent variable. Three variables were significantly
contributing to variation in seed cotton yield, these variables were number of
bolls/plant, boll weight and seed index with relative contribution equal to (R2%)
91.69% for all variables and contribution 91.63% for acceptance variables.
Meanwhile two variables were significantly contributing to variation in lint yield, these
variables were number of bolls/plant and boll weight with relative contribution 91.17%
for all variables and 91.14% for acceptance variables. The path coefficient analysis
indicated that number of bolls/plant and boll weight were the most prominent direct
and indirect effects on seed cotton yield and lint yield. Total contribution of these
characters over all variation in seed cotton yield (mod.1) and lint yield (mod.2) were
91.61 % and 90.63%, respectively.
In general the results obtained herein indicated that number of bolls/plant
and boll weight were the major and most consistent source accounting for variation as
total contribution.