Behavioural correlates of phenotypic plasticity in mouthpart chemoreceptor numbers in locusts

Rearing locusts in an impoverished chemosensory environment leads to fewer chemoreceptors developing on the mouthparts and antennae as adults but the behavioural relevance of these changes remains unknown. To address this question, locusts were reared for the final two larval stadia on either a single, nutritionally near-optimal synthetic food ('plain' pretreatment), or a diet comprising two nutritionally complementary foods containing two added flavours ('mix' pretreatment). Insects reared on the 'mix' diet had a mean 20% more chemosensilla on the maxillary palps than those fed on the 'plain' diet. Following an equilibration period, when all newly moulted adults could feed on two nutritionally complementary foods, insects were food deprived for 2 or 6 h, and then given a test meal of a single balanced food at one of two dilutions whilst their behaviour was recorded. 'Mix'-pretreated locusts had a shorter latency to feed and were more likely to reject the test food upon first contact if deprived for only 2 h; but if they did take a meal it lasted longer and contained fewer pauses. Using sensilla number as a covariate removed the statistical significance of pretreatment regime, indicating that sensilla number, or some close correlate of it, can largely account for the variation in behaviour. This suggests that sensilla numbers are behaviourally relevant; particularly where locusts are not greatly food deprived and faced with marginally acceptable foods.