ABSTRACT
Design for comparisons is discussed in B. W. Brown; design for mixtures is considered by K. M. Abdelbasit and R. L. Plackett; design for the division of fixed resources between control and experimental animals is discussed in P. F. Thall and R. Simon, with reference also to the weight to be placed on historical controls. Sequential methods clearly require relatively rapid responses, and they could be well suited to experiments on large animals, say, for which a non-sequential experiment would simply take up too much laboratory space. An attractive, practical feature of the standard up-and-down experiment is that it makes use of a limited number of doses, which may therefore be mainly made up in anticipation of the experiment. The main conclusion from M. Davis's comparison of various up-and-down and Robbins-Monro designs was the importance of incorporating an initial delay, corresponding to excluding initial trials when appropriate, and the good performance of delayed designs under fairly general circumstances.
