Effects of within- and among-patch experiences on the patch-leaving decision rules in an insect parasitoid.

摘要

The present study aimed to address how an insect parasitoid makes patch-departure decisions from various types of host patches and how previous patch experiences in the environment modify this decision-making process. Experiments were done with the parasitic wasp Aphidius rhopalosiphi attacking the grain aphid Sitobion avenae. In the experiments, wasps were observed in a laboratory environment containing several patches of various host densities, and behavioural records were analysed using a Cox's proportional hazards model. Consideration of the effect of the within-patch experience gave a classic pattern of patch-leaving decision rules in parasitoids: A. rhopalosiphi used local information on host quality (i.e. numbers of ovipositions or rejections) and availability (i.e. patch density) to determine departure decision. However, consideration of previous patch experiences provided evidence that these departure rules are fundamentally dynamic, responding to the physiological state of the animal and the information it has about its environment. Results showed that A. rhopalosiphi decreased its tendency to leave the visited patch after an oviposition. However, when a female has already laid several other eggs in the environment, such an incremental mechanism gradually switched to a decremental one. Hence, A. rhopalosiphi responded to egg-load depletion by leaving the visited patches sooner and by depositing a smaller number of eggs in those patches, which probably led to a decreased level of superparasitism. Results also indicated that previous experiences enabled wasps to estimate spatial host distribution and then to adjust their behavioural decisions accordingly. Thus, A. rhopalosiphi was shown to adjust its patch residence time according to the quality and the number of the patches previously visited. These proximate mechanistic rules adopted by A. rhopalosiphi females are discussed in the context of general predictions from optimality models.