Plant Chemistry and Insect Species Richness of British Umbellifers

Abstract

(1) Data on the number of insect species per host-plant, host-plant physical characteristics (geographic range, size and leaf form), the presence of biosynthetically distinct classes of secondary metabolites and intensity of biochemical investigation were compiled for the British umbellifers (Apiaceae) from the Phytophagous Insect Data Bank, previously published studies, and phytochemical literature, respectively. (2) Measures of the chemical diversity and uniqueness of each plant species were derived from the phytochemical data, and four hypotheses relating insect species richness to these chemical variables were explicitly developed and tested using single and multiple regression techniques. (3) The diverse defence and biochemical barrier hypotheses predict negative correlations between insect species richness and chemical diversity and uniqueness, respectively, because plants with diverse or unusual defences should show reduced insect colonization over evolutionary time. Alternatively, the common chemistry and enemy escape hypotheses predict positive correlations between insect richness and these two chemical variables, because plants with diverse chemistry have a higher probability of sharing chemicals common to other plants, facilitating host shifts, and because insect herbivores on biochemically unusual plants may escape from predators and parasites that use host odours to locate prey, again facilitating host shifts. (4) As in previous studies, size of plant geographic range, plant size and leaf form all influence insect species richness. Together, these variables account for 46% of the variance in insect species richness, leaving up to 54% of unexplained variance that could potentially be accounted for by host plant chemistry. (5) Insect species richness is not significantly correlated with chemical uniqueness; thus there is no support for the biochemical barrier or enemy escape hypotheses. (6) Insect species richness is positively correlated with chemical diversity, supporting the common chemistry but not the diverse defense hypothesis. However, support for the common chemistry hypothesis is tenuous, because of the confounding effects of intensity of biochemical investigation. When intensity of biochemical investigation is factored out, the effects of chemical diversity on insect species richness are still positive and significant but account for only 5% of the variation. (7) We conclude that there is no evidence from these data that biochemically diverse or unusual British umbellifers support less-species-rich assemblages of insect herbivores. While the data suggest that chemically diverse plants have more insect species, we consider this evidence to be weak, at best. (8) We discuss the need for further tests of these hypotheses and the types of phytochemical data required.