A method of using the proportions of undamaged carrots or parsnips to estimate the relative population densities of carrot fly (<i>Psila rosae</i>) larvae, and its practical applications

Abstract

SUMMARY The plant‐to‐plant distribution of carrot fly damage was studied on carrots and parsnips in field and microplot experiments. The observed data were well‐fitted by the relationship In (–In q ) = In L + constant, where q is the proportion of undamaged roots and L is the number of larval mines in a sample and is proportional to the number of associated larvae, N . Hence, if N 1 and N 2 are the numbers of larvae at two levels of infestation on one occasion, the relative numbers present, N 2 : N 1 are estimated by In q 2 : In q 1 In this way the relative efficiencies of insecticide treatments can be determined within an experiment from only q 1 and q 2 and compared between occasions. A nomogram relating q 1 , q 2 and treatment efficiency is provided. The method can also probably be used to compare the relative susceptibilities of host‐plant cultivars to carrot fly attack. Guidance is given on the sample sizes needed to estimate q within 95% confidence limits and constraints on the application of the method are stated. An appropriate statistical procedure for analysing data sets on carrot fly infestations is by analysis of deviance and an example is given in an Appendix. The principles described can probably be adapted to other crop/pest problems.