VARIABILITY ANALYSIS IN FIELD EXPERIMENTATION A CASE STUDY ON MAIZE
Loading...
Date
1995
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
AAU, Anand
Abstract
Variability as judged by the coefficient of variation in per cent (CV) is an indicator of reliability of the experimental estimates of the field experiments. CV depends on the experimental, soil and climatic factors. No work so far has been initiated on these aspects, particularly, for maize crop grown in the Middle Gujarat in agro-climatic conditions. The experiments on maize crop, conducted at the Maize Research Station, Godhra show very large variation (CV) in results, calling attention of Agricultural Statisticians to search on method(s) which may help in lowering CV of the field experiments. Similarly, there is no yardstick for determining upper bound of CV which can help in accepting or rejecting the experimental results. Therefore, the present study, being location as well as crop specific, was initiated (i) to study the experimental factors responsible for high variability in the field experiments conducted at Godhra Farm on maize crop, (ii) to study the distribution of CV observed in the field experiments, (iii) to develop yardstick for the reliability of experimental results, and (iv) to generate information on plot techniques including optimum plot size and shape for minimizing the experimental error. For achieving the goal, secondary data of 231 field experiments and primary data generated through the uniformity trial, on maize were analysed statistically. The results revealed the following. * The CV of maize experiments conducted at Godhra Farm was influenced by the experimental plot size, number of treatments and number of replications. * The CV of maize experiments followed positively skewed and leptocurtic distribution and gave best fit to the erlang distribution. * The yardstick of CV for the field experiments was obtained as 31 per cent, which is on higher side. * The optimum plot size through the univariate and multivariate approaches was found to be of 10 units (6.0 sq.m) with a shape of 5 rows each of 2 m length. * The soil heterogeneity indices revealed that the variation in soil fertility was in patches. * Block size (i.e., number of treatments) influenced the precision of the experimental estimates. * Among various experimental designs, latin square design was comparatively more efficient than other designs for field experiments on maize crop at the Godhra Farm. The present study indicated that the site (location) for the field experimentation on maize crop is not proper due to high soil fertility variation.
Description
Keywords
AGRICULTURAL STATISTICS, A CASE STUDY