The results of an interim examination of the experiments at Rothamsted and Woburn on the effects of nitrogen, phosphate and potash on a six-course rotation of crops are reported.

It has become apparent that the main aim of the experiments, to relate the effects of the fertilizers to weather conditions, is not being fully realized. This is in part a consequence of the design, which has inherently low accuracy and results in low responses to phosphate and potash. In addition, faulty randomization and, at Woburn, fertility irregularities and pest damage have further reduced the accuracy attained.

The faulty randomization also seriously affects the estimates of experimental error given by deviations from smoothed response curves. Alternative estimates of error are provided to overcome this difficulty.

The mean yields and mean standard responses over the period 1931–55 are presented.

There is very little year-by-year variation in the responses to phosphate, but the responses to nitrogen and, for some crops, the responses to potash show sufficient variation to justify further analysis. As a first step the regressions of the responses to nitrogen on winter rainfall and on rainfall during the growing season and, for the root crops only, the regressions on sowing or planting date are briefly considered.