A Phenothiazine Experiment Statistically Treated

Abstract

In a previous paper, Peters & J. Leiper (1940) pleaded that, in anthelmintic experiments, the statistical problem of variation should be considered, not only so as to assess the significance of treatment differences after the results have come to hand, but also with a view to the prior construction of an experimental design capable of giving reliable answers to the questions posed. Reference was made to the “Enormous expenditure of time, labour and money” on anthelmintic experiments, the results of which usually remain undigested and of unknown significance.While the paper was passing through the press a symposium on phenothiazine as an anthelmintic, led and summarized by Taylor & Sanderson (1940), was published in the Veterinary Record. The symposium deals with “Work carried out in various parts of the country and co-ordinated through the Agricultural Research Council” it covers some 150 horses, 300 to 400 sheep, 70 to 80 goats, and a few cattle, pigs, and dogs. The drug was found to be remarkably efficient against horse strongyles and the stomach-worms of ruminants, and in varying degree less efficient against other worms.Extensive tables of egg-counts, worm-counts, and weights of animals are published in the symposium. It was usual to count eggs or weigh animals once before treatment and several times afterwards. The repeated after-treatment determinations are made in order to detect a general trend in egg-counts or weights: in other words there is no attempt to measure and make allowance for the very considerable upand- down variation from day to day.