Mosquito females quantify risk of predation to their progeny when selecting an oviposition site

TLDR: It is suggested, based on statistical and behavioural factors, that pairwise comparisons, and not simultaneous multiple-choice experiments, are the experimental design of choice to adequately test whether prey can quantify predation risk when selecting habitats.

Abstract: Summary 1. Numerous studies demonstrate that given the dichotomous choice of predator-free habitats vs. habitats containing predators, prey choose predator-free habitats when foraging for food or ovipositing. However, predation risk is rarely dichotomous in nature, and very few studies have assessed whether prey can quantify predation risk when selecting habitats. 2. It was shown previously that gravid females of the mosquito Culiseta longiareolata, when simultaneously offered pools with multiple choices of densities of the predator Notonecta maculata, oviposited more in the zero-predator density pools but oviposited less frequently and similarly across all other densities. This flat oviposition response across various Notonecta densities was in contrast to a decrease in mosquito immature survival with increasing Notonecta density. Here, we reconsider this question with the same species but with a different experimental design; rather than experimentally assessing multiple predator densities simultaneously, we offered only pairwise choices on any given night. Specifically, we offered ovipositing Culiseta females all pairwise combinations from no, low and high predation risk (0, 1 and 4 Notonecta per pool). 3. Overall oviposition was lower when mosquitoes could only choose pools containing Notonecta (1 or 4). In all pairwise comparisons, more females chose pools of lesser predation risk. Thus, gravid females of this species, and probably many other species, can quantify predation risk, and not only assess presence or absence of predation risk, when choosing oviposition sites. 4. This is the first demonstration that an ovipositing female of any species can quantify risk of predation. We suggest, based on statistical and behavioural factors, that pairwise comparisons, and not simultaneous multiple-choice experiments, are the experimental design of choice to adequately test this ability.