Panicle surface area as a selection criterion for grain yield in pearl millet (Pennisetum glaucum)

In an experiment 24, experimental varieties of 3 composites, viz 'EC 87', ' EC 91' and 'HHVBC', were evaluated during rainy season of 1996 and 1997 in 11 environments at 3 locations for the effect of selection for particle sur face area on grain yield and its components in pearl millet [Pennisetum gla ucum (L.) R. Br. emend. Stuntz]. Singificant variation was observed among e xperimental varieties for all the traits. The actual gain in 'PCV 5' (with maximum panicle surface area) over the original population was 8.7% for gra in yield/m(2) and 7% for grains/m(2) across the composites. Breeder selecti on was etifective in improving grains/m(2) by an average of 4%, in 3 compos ites, but it was effective in improving grain yield/m(2) over. the original population by 3% in 'EC 91' only. The net increase in grain yield/m2 was a lmost similar to the prediction (+8.7% vs prediction of 10%). Also the sele ction for large panicle surface area, ic 'PCV 5', was more effective than b reeder selection for grain yield/m(2) by an average of 9% in the 3 composit es. The panicle surface area showed positive correlation with grain number/ m(2), grain size and grain yield/m(2) among experimental varieties ('PCV 1' to 'PCV 5') in 3 populations.