A Unique correlation has recently been obtained from some experimental data at Rothamsted. In the autumn of 1924 a set of 47 unmanured plots were laid out in Sawyer's Field, and were sown with wheat to test the uniformity of the fertility of the soil. The plots were each ·098 acre. The produce of grain and straw from each was recorded, and an examination of the figures showed that the straw yield recorded for plot 1 A was much lower than one would have expected under normal circumstances. A careful enquiry into the matter revealed that by an oversight a weighing of straw had been removed from the scales before the weight had been entered in the record. As it was impossible to verify the exact value of the missing entry, the yields from this plot were omitted from all statistical reductions of the data, leaving 46 plots upon which to base an estimate of the various statistics.

The soil was found to be very variable in fertility, the average yield of total grain per plot being 149·57 lb., with a standard error of 6·39 lb. and the average yield of total straw per plot was 194·35 lb., with a standard error of 7·55 lb.

The grain and straw yields of the 46 plots are shown in the accompanying diagram. The very close association between the total weight of grain and total weight of straw is evident. The amount of the relationship was measured by correlating the 46 pairs of values.