Data on Fresh Fruit Bunches weight (FFBW), number of bunches (BN) and average bunch weight (ABW) on 395 palms (23 Costa Rican and one Palode hybrids) for four years were subjected to different statistical models of Analysis of Variance (ANOVA). In first and second models, individual palm wise data and plot wise data, respectively were considered for studying the effects of hybrids, replication, years and hybrid X year interaction. In third model, plot wise data averaged over years were considered to study the effects of hybrids and replication. For above three characters scoring was done and promising hybrids were selected based on greater averages with lesser deviations within hybrid. The correlation between the scores of bunch weight and the total scores based on bunch weight, number of bunches and average bunch weight were found to be very high (r = 0.91), indicating the efficiency of scoring technique used in selecting promising hybrids.
Statistical models, oil palm, scoring method, correlation
