Estimates of gene effects through joint scaling test of three and six parameter and sequential fit model in five crosses for eleven characters were investigated. It was noticed that simple additive dominance model exhibited lack of good fit for all the traits. So, sequential fit model was searched after eliminating the non-significant parameters of six parameter model. Five parameter sequential fit model was observed for number of primaries per plant (cross 1), number of seeds per capsule (cross 4), oil content (cross 2 and 3), seed yield per plant (cross 1) and chlorophyll content (cross 3). Best fit four parameter sequential model was observed for number of primaries per plant (cross 2) and 1000 seed weight (cross 1). Higher order interactions epistasis or linkage were observed for days to 50 per cent flowering (cross 1, 2, 3, 4, and 5), days to maturity (cross 1, 2, 3, and 4), plant height (cross 3, 4, and 5), number of effective primaries per plant (cross 3, 4, and 5), number of effective capsules per plant (cross 2, 3, 4, and 5), number of seeds per capsule (cross 1, 2, 3, and 5), 1000 seed weight (cross 4 and 5), seed yield per plant (cross 4 and 5), oil content (cross 1, 4, and 5) and chlorophyll content (cross 1, 2, 4, and 5). Differential model schemes for same trait in different crosses were noticed in the present investigation. It was due to different parents involved with variable gene frequency with opposing and reinforcing genetic effects. The magnitude of [d] was relatively small to that of other genetic effects. This indicated that, additive genes are playing a minor role in the inheritance of these traits.