The design and analysis of an experiment to compare the yield of wheat grown continuously with that of wheat grown in two phases of a four-course rotation is described and the results discussed. A method of analysis which avoids making assumptions about the covariance structure of the errors derived from repeated observations on the same plots, which could be appropriate to other sorts of rotation experiments, is given.There were no differences between yields of the first and second wheat crops which followed oats after beans in a four-course rotation, but the average of these two crops was 0-75 t/ha greater than the yield of wheat grown continuously. The response to N was the same in the two systems of cropping. The only yield component affected by the system of cropping was grain mass.
TNTT? ODTTPTTONExperiments made deliberately to assess the effects of length of time under a particular crop Hanley & Ridgman (1978) reported the results of (including those mentioned above) have always ina long-term experiment comparing various stages volved as treatments a factor which has become of growing continuous wheat on heavy land at known as 'intake year', that is, within each repliCambridge using two varieties and two rates of cate one or more different crops usually called application of nitrogenous fertilizer. Their main 'break crops' are grown on one plot of each repliconclusion was that the yield of the second wheat cate in each of several successive years so that at crop was substantially less than that of the first the end of a suitable period the 1st, 2nd, . . .nth following oats after beans, but that there were no wheat crops are present in each replicate in the consistent yield differences between the second and same year. This involves a large amount of prelater crops. Increasing nitrogen application from liminary work for a small amount of results if n is 63 to 126 kg N/ha did not reduce the differences to be of any useful size. Carrying on such an experibetween first and later wheat crops.