Oviposition habitat selection by a mosquito in response to a predator: are predator-released kairomones air-borne cues?

Chemical cues play an important role in predator-prey interactions in aquatic environments (e.g., Petranka et al. 1987, Dodson et al. 1994, Wisenden 2000). Predator-released kairomones may induce morphological changes in prey (e.g., Laurila et al. 2004, Von-Bert and Stibor 2006), foraging changes by prey (e.g., McCarthy and Dickey 2002, Turner and Montgomery 2003), and behavioral responses of gravid prey females via oviposition habitat selection (e.g., Kiesecker and Skelly 2000, Orizaola and Brana 2003). There is a growing body of literature showing that a number of mosquito species detect some predators via chemical cues, causing them to avoid these predators when choosing an oviposition site Although a number of chemicals have been identified that mosquitoes respond to when ovipositing (reviewed in Bentley and Day 1989, Clements 1992, McCall 2002), little is known about how mosquito females detect predator-released kairomones in breeding sites, and no such kairomone has been chemically identified. Mosquitoes may detect chemicals from the air when the chemical possesses sufficient volatility or, in the case of low volatility chemicals, by a gustatory mechanism involving direct contact with the water (Clements 1992). Several studies show that mosquitoes chemically detect and avoid backswimmer species when ovipositing (e.g., Anopheles species: Munga et al. 2006, Culex/Culiseta species: Blaustein et al. 2004, Blaustein et al. 2005). Knowing whether the mosquito needs to touch the water or is able to detect the predator from the air will help narrow down the candidate chemicals that act as predator-released kairomones. We have previously shown that the mosquito Culiseta longiareolata Macquart chemically detects and avoids the predatory backswimmer Notonecta maculata Fabricius (Blaustein et al. 2004). Here, we conduct an artificial pool experiment designed to test whether C. longiareolata can detect the chemical cues from N. maculata, without touching the water. To assess whether C. longiareolata needs to make contact with the water to detect this predator, we used a behavioral assay in an outdoor experiment at the University of Haifa campus on Mount Carmel, Israel (32 0 44'N 35 0 01'E). Oviposition units consisted of a central, uncovered oviposition pool dug into the ground surrounded by an adjacent, above-ground, water-filled channel that was screened at the top (Figure 1). Thus, ovipositing mosquitoes could land on the water surface in the central oviposition pool but could not reach the water of the surrounding screened channel. All oviposition pools were identical in size (LxWxH: 60x40x15 cm), water volume and type …