Abstract:
A uniformity trial was conducted in bhindi intercropped with cowpea at the experimental field of College of Horticulture, Kerala Agricultural University, Vellanikkara during July 1993 to December 1993 to assess the nature and magnitude of soil heterogeneity, and to determine the optimum size of plot for increasing the efficiency of experiment with intercropping. Three different approaches have been attempted in the statistical analysis. At the time of harvest, the yield data from 320 plots each of size 0.60m x 0.45m were recorded separately after discarding the border rows.
It was observed that the index of soil heterogeneity was very high in all these approaches indicating that the contiguous plots are not correlated and the fertile spots are distributed randomly or in patches. It was also observed that an increase in the plot size in either direction decreased the coefficient of variation but the decrease was not proportional in all these approaches.
The empirical law suggested by smith gave a satisfactory fit to the data. All the other non- linear models tried also gave a satisfactory fit to the data in all these approaches.
The optimum plot size obtained by following different approaches indicated that 10 basic units plots(2.7m2) are optimum for conducting the intercropping experiment with bhindi and cowpea.
Comparison of the LER and the bivariate analysis method showed that for the same level of precision the LER method gives a smaller plot size implying that per unit cost of experimentation will be higher in bivariate analysis. For any choosen plot size the precision attained through LER analysis is larger than that through bivariate analysis. However, when the multivariate approach is adopted the optimum plot size was obtained as 10 basic units (2.7m2) as in the case of LER.
