The survey shows that objective estimates of the yield of maincrop potatoes can be obtained from small samples carefully selected and dug by hand. Samples taken from about 1000 fields gave estimates of the mean yield of all counties sampled with a standard error due to sampling of less than ± 0·2 ton/acre. The precision of the estimate could have been improved by a better distribution of samples among counties.

The results point to underestimation on the part of the official estimates, in each of the 3 years, especially in the case of high yields in particular counties, and in particular years. The discrepancy between the official and the survey yields is of the order of 1¾ tons/acre, after all necessary corrections have been applied to the survey yields.

The experience gained in the survey indicates that the method of sampling adopted provides an accurate and reliable method of estimating the yields of potatoes which could supplement, and, possibly, ultimately replace the present official estimates if more accurate estimates are required. A national scheme, properly designed, which would include all the potato-growing areas in due proportion should not be unduly expensive to operate. Estimates so obtained would not only be generally more accurate than those obtained by the present official method, but, perhaps more important, would indicate far more closely the fluctuation in yield from year to year.