Experimental Error and the Field-Plot Technique with Potatoes (With One Diagram and One Text-figure.)

Abstract

The present investigation consists of the statistical analysis of a uniformity trial with potatoes conducted by Dr Kirk. In this study the standard error in per cent. of the mean decreased slightly with the increase in the widths up to plots 5 rows wide, but any further increase in the width of the plot resulted in the higher standard error. The fertility contour map of the field is given in Fig. 1 to show graphically the effect of soil heterogeneity on the yield. The increased size of the plot resulted in the decreased efficiency in the use of the land when the entire plot was harvested; in other words, given a piece of land of certain size, it is advantageous to have a greater replication of smaller plots than a smaller number of larger plots. Four-row plots proved to be the most efficient when the border rows are discarded. The superiority of long and narrow plots over shorter and wider ones is demonstrated.With the greatest pleasure I acknowledge my indebtedness to Dr Kirk, who supplied me with the data, to Dr Fisher for his valuable suggestions and to Dr J. Wishart for useful advice and criticism.