Abstract

                                Bengal gram was grown under three differont plent spacings (30 x 10 cm, 45: 10 cm and 30x20 cm) during 1973 Kharif season at Tamil Nadu Agricultural University experimental farm. At the time of harvest, 50 plants in the centre of each plot were selected at random and data on grain weight por plant and grain number per plant were collected. Considering replication as primary sampling units and plants within plots secondary sampling units. variance components viz. variance among plants within plots were estimated for the two characters seperately at 'each spacing using analysis to variance technique. It was observed that more than 86 per cent variability was attributed among plants within plots and rest among plots. Plants spaced with 45 x 10 cm and 30: 20 cm exhibited more variation then plants with 30x 10 cm suggesting selection of more sample plants in wider spaced treatments for estimation purposo. The relationship between mean and variance indicated that the two characters can be analysed with anginal values and their means compared in comperative experiments. To estimate grain weight per plant and grain number per plant. with a precision of 5 par cent standard error, a minimum of 116 and 65 plants respectively were considered necessary.

Key words :