Oecologia (2014) 175:565–575 DOI 10.1007/s00442-014-2930-x

Population ecology - Original research

Interactive effects of salinity and a predator on mosquito oviposition and larval performance Alon Silberbush · Ido Tsurim · Yoel Margalith · Leon Blaustein 

Received: 18 April 2013 / Accepted: 6 March 2014 / Published online: 26 March 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  Oviposition habitat selection (OHS) is increasingly being recognized as playing a large role in explaining mosquito distributions and community assemblages. Most studies have assessed the role of single factors affecting OHS, while in nature, oviposition patterns are most likely explained by multiple, interacting biotic and abiotic factors. Determining how various factors interact to affect OHS is important for understanding metapopulation and metacommunity dynamics. We investigated the individual and interactive effects of three water salinities (0, 15 and 30 p.p.t. NaCl added) and the aquatic predator Anisops debilis Perplexa (Hemiptera: Notonectidae) on OHS and larval performance of the mosquitoes Ochlerotatus caspius Pallas and Culiseta longiareolata Macquart (Diptera: Culicidae) in outdoor-artificial-pool and laboratory experiments. C. longiareolata inhabited only freshwater pools, strongly avoided pools containing A. debilis, and larvae experienced lower survival in the presence of A. debilis. Communicated by Steven Kohler. Y. Margalith: deceased. Electronic supplementary material The online version of this article (doi:10.1007/s00442-014-2930-x) contains supplementary material, which is available to authorized users. A. Silberbush (*) · I. Tsurim · Y. Margalith  Department of Life Sciences, Center for Biological Control, Ben-Gurion University of the Negev, P.O. Box 653, 8410501 Beer‑Sheva, Israel e-mail: [email protected] A. Silberbush · L. Blaustein  Community Ecology Laboratory, Institute of Evolution and Department of Evolutionary and Environmental Biology, Faculty of Natural Sciences, University of Haifa, 3498838 Haifa, Israel

Salinity concentration interacted strongly with the predator in affecting OHS and larval survival of O. caspius; oviposition increased with increasing salinity in the absence of the predator and decreased with increasing salinity in the presence of the predator. O. caspius larval survival in predator-free pools was lowest in freshwater and highest at intermediate salinity. In predator pools, survival was highest at high salinity, where predation rate was shown to be lowest in the laboratory. Our results highlight that assessing the role of single factors in affecting mosquito distributions can be misleading. Instead, multiple factors may interact to affect oviposition patterns and larval performance. Keywords  Anisops debilis · Culiseta longiareolata · Ochlerotatus caspius · Context dependence · Predation risk

Introduction Understanding the individual and interactive effects of biotic and abiotic factors in explaining the distribution and abundance of organisms remains a central goal in both basic and applied ecology (Schemske et al. 2009; Bradshaw and Holzapfel 2010). Ecologists are increasingly recognizing that non-random patterns of distribution may be explained not only by performance in a given habitat, but also by colonization or oviposition habitat selection (OHS) (Blaustein 1999; Binckley and Resetarits 2005; Morris et al. 2008; Juliano 2009; Refsnider and Janzen 2010). In determining habitat quality for potential habitat selection, organisms must integrate both abiotic and biotic conditions. Habitat preference for specific abiotic conditions that may influence food resources or physiological tolerance is not constant, but may change depending on the distribution of competitors [e.g., gerbils (Rosenzweig and Abramsky

13

566

1997); salmon and trout (Young 2004)] and predators [e.g., Notonecta nymphs (Sih 1980); gerbils (Kotler and Blaustein 1995)]. Similarly, when gravid females select sites to deposit progeny, the selection of abiotic conditions may depend on competitors and predators. In temporary pools, success of deposited immature stages varies greatly among pools, depending on physicochemical and biotic conditions (Binckley and Resetarits 2005). Consequently, locating a breeding site is crucial to fitness (Blaustein 1999; Resetarits 2001). Two factors that may strongly affect among-pool distribution and performance of immature mosquitoes are predation and water salinity. Predation has been shown to be an important force in negatively affecting insect larval survival in pool habitats (Blaustein et al. 1995; Bond et al. 2005). It is not surprising that a growing number of studies are showing that for those mosquitoes at high risk of predation at the larval stage, gravid females can detect the presence of predators via chemical cues (kairomones) and avoid predator-inhabited sites when ovipositing (reviewed in Vonesh and Blaustein 2010). Salinity, likewise, is an important factor affecting invertebrate community structure (Silberbush et al. 2005; Carver et al. 2009) including the development and survival of many mosquito species inhabiting ephemeral aquatic habitats (Clark et al. 2004; Van Schie et al. 2009). Similar to predator avoidance, gravid mosquito females have been shown to be capable of detecting salinity concentrations when choosing an oviposition site (Foley and Bryan 1999; Osborn et al. 2006). The relationship among salinity, OHS and fitness is not always straightforward (Roberts 1996; Foley and Bryan 1999; Osborn et al. 2007), most likely because salinity may interact with other factors such as predators, competitors, and resources to affect mosquito OHS and consequently, fitness (Osborn et al. 2006, 2007; Carver et al. 2010). Many previous studies on mosquito OHS were conducted under laboratory conditions and usually focused on a single factor affecting OHS (McCall 2002). However, the influence of a single factor may be context dependent (e.g., Sadeh et al. 2009). Temporary pools around the Dead Sea Area, Israel, vary greatly with respect to salinity, ranging from freshwater to considerably above-normal seawater concentrations (Dimentman and Margalit 1981; I. Tsurim, unpublished data) and with respect to predators. The predatory backswimmer Anisops debilis Perplexa (Hemiptera: Notonectidae) is very abundant in the Dead Sea area (Larsen and Blaustein 2005) across a wide salinity range, and early summer populations were commonly found to be >50 individuals per a 15-L sweep net sample (A. Silberbush, personal observations). A previous outdoor mesocosm study in this area, in which experimental pools ranged in water salinity, showed that Ochlerotatus caspius Pallas (Diptera:

13

Oecologia (2014) 175:565–575

Culicidae) is the most prevalent mosquito species and that its larvae are euryhaline, being found in high densities over an experimental range of 0–30 p.p.t. NaCl, but tend to be least abundant in freshwater (Silberbush et al. 2005). Congruent with this finding, a laboratory study, in the absence of any potential competitors and predators, found survival of O. caspius larvae across this salinity range to be lowest in freshwater (Silberbush 2004). The potential impact of A. debilis on mosquitoes has not been investigated, but other backswimmer species have been shown to reduce mosquito oviposition (Eitam et al. 2002; Blaustein et al. 2004, 2005) and affect prey fitness via risk of predation as well as predation (Hampton et al. 2000). A. debilis may also have an indirect positive contribution in affecting mosquito larvae by causing a trophic cascade, generally increasing biomass of algae (Silberbush 2004), which is an important food resource for mosquito larvae. In addition to the direct effect of salinity on mosquito larval performance, it could also have an indirect effect via predator performance; although A. debilis was found in a wide range of salinities in natural pools (I. Tsurim, personal observation), it could primarily be adapted to a confined range of salinity levels and show reduced fitness or performance in other levels. In addition to predation and salinity, mosquito larval distribution is also known to be affected by inter-specific competition (Juliano 2009) and oviposition avoidance of competitors (Stav et al. 2005; Duquesne et al. 2011). A likely competitor of O. caspius larvae in freshwater pools of this region is the mosquito Culiseta longiareolata Macquart (Diptera: Culicidae). C. longiareolata has been shown to be a strong interspecific competitor of O. caspius in a laboratory setting (Tsurim et al. 2013) and a strong competitor and predator of anuran larvae in an outdoor mesocosm experiment (Blaustein and Margalit 1994). Silberbush et al. (2005) found that C. longiareolata larvae are restricted to freshwater pools though it was not determined whether this was due to OHS, larval performance or both. C. longiareolata is also very vulnerable to predatory backswimmer species (e.g., Blaustein et al. 1995; Eitam et al. 2002) and avoids ovipositing in pools containing these predators (Eitam et al. 2002; Blaustein et al. 2004). Thus, a salinity gradient and the presence or absence of A. debilis may indirectly affect O. caspius distribution among pools via its potential competitor, C. longiareolata. Here, we experimentally investigate the individual and interactive effects of water salinity and the predator A. debilis on OHS and larval performance of the mosquitoes O. caspius and C. longiareolata in an outdoor mesocosm experiment combined with supplementary lab experiments. Potential effects of predator and water salinity on female oviposition and larval survival for both mosquito species are shown in Fig. 1. We hypothesized that:

Oecologia (2014) 175:565–575

Fig. 1  Schematic diagram of hypothesized interactions between the predator Anisops debilis, the mosquitoes Culiseta longiareolata, Ochlerotatus caspius, and increasing water salinity. Arrows point to the affected species. Minus sign Hypothesized reduction of fitness, plus sign hypothesized increase of fitness, question mark unknown interaction prior to the study

1. O. caspius is distributed over a range of salinities in the absence of the predator, A. debilis. 2. C. longiareolata is found only in freshwater pools and avoids ovipositing in saline waters. 3. O. caspius and C. longiareolata are highly vulnerable to predation by A. debilis and thus avoid predator pools when ovipositing. 4. O. caspius, in integrating risk of competition, salinity and risk of predation, demonstrates a salinity × predator interaction with respect to an oviposition response. That is, we expected that fewer O. caspius eggs would be found in predator pools under saline conditions, but in freshwater, the presence of the competitor C. longiareolata in predator-free pools partially offsets the risk of predation in freshwater predator pools.

Materials and methods Field experiment Experimental design An outdoor artificial pool experiment was initiated on a sandy field adjacent to a date palm plantation near Ein Tamar, South Dead Sea Basin, Israel (30°56′18″N 35°22′47″E) where a previous outdoor pool experiment (Silberbush et al. 2005) and adult trapping indicated that O. caspius is locally common. Twenty-four green plastic pools (width × length × height: 50 × 40 × 20 cm), dug into the ground, were organized in a 4-column × 6-row grid. The four columns were 10 m apart and the inter-pool distance within each column was 1 m. On 12 February, we

567

added 35 L of tap water and 10 g of Koffolk mice pellets to each pool. The western half of each pool was covered by light-colored cardboard to reduce solar radiation and high water temperatures. We added 0 (freshwater), 15 (intermediate salinity) or 30 p.p.t. NaCl (high salinity) to each of eight pools, so that each of the four columns served as a statistical block, containing two randomly assigned pools of each salinity concentration. Pool communities were left to develop for 30 days with the salinity manipulation only (hereafter, “pre-predator” period) before adding the predators to half the pools. Typically, predators colonize temporary pools later than many prey species (Hein and Gillooly 2011). No predators colonized our pools during this period. We also maintained this pre-predator period for this 30-day time period to understand mosquito distributions in the absence of predators with reasonable statistical power. Then on 17 March, we initiated the “predator” period by introducing 25 early instar (second and third) nymphs of A. debilis to a randomly assigned pool for each salinity in each block (12 predator pools in total, four replicate pools for each predator × salinity treatment combination). We used juvenile-stage A. debilis individuals because during March, most of the individuals found in temporary pools were early instars. They are also logistically more convenient for experimental manipulation as juveniles cannot move among pools until adulthood. The backswimmers were collected from a nearby water reservoir and added to the pools on the same day. The predator density used was well within the range of densities observed in natural pools for this species during this season. Dead or missing Anisops individuals were replaced regularly during the course of the experiment. No Anisops individuals were found to colonize non-predator pools during the entire experiment. No other predators were found to colonize the pools during the prepredator period. During the predator period, 93 adult individuals of a small dytiscid species (