Published online by Cambridge University Press: 27 March 2009
(1) The desiderata of a sampling-method are outlined, and the particular case of sampling a large number of potato plots discussed.
(2) An analysis is made of the yields of 54 sub-plots of the Rothamsted Potato Experiment of 1928, both as estimated by a sampling-method and as determined by large-scale lifting.
(3) It is shown that most of the significant results of the experiment are obtained from the sample-yields, but that the higher standard error per plot obscures the effect of superphosphate.
(4) It is concluded that at Rothamsted 102, and at Woburn 56, plants would have to be lifted to give a sampling-error as small as 4 per cent. It would then be profitable only to sample experimental plots of 1/20th acre or more in area.
Finally it is a pleasure to record our indebtedness to Dr R. A. Fisher for much valuable advice and criticism: and to Mr H. J. G. Hines for assistance with the field work.