STATISTICAL ANALYSIS OF MOLLUSCICIDE FIELD TRIALS.

The paper describes the methods used in the statistical analysis of field trials of acrolein, Bayluscide, ICI 24223 and NaPCP molluscicides. The study shows the importance of adequate sampling of the snail population both before and after the application of the molluscicide and the necessity of obtaining equally complete data from the controls. For kills above 99%, if the pretreatment collection contains fewer than 400-500 snails, the range between the confidence limits exceeds 5% and, if the pretreatment sample contains fewer than 1200-1500 snails, the range exceeds 2%. For all the molluscicides except acrolein, a cube-root transformation was found to be applicable. After treatment with acrolein, the snails were distributed at random. It most of the trials there were significant changes in the snail populations in the control canals, possibly due to sampling. No significant differences were observed between samplings taken 48 hours and one week after treatment; one of these could be omitted in future trials.