Oviposition habitat selection in response to risk of predation in temporary pools: mode of detection and consistency across experimental venue

Natural selection should favor females that avoid ovipositing where risk of predation is high for their progeny. Despite the large consequences of such oviposition behavior for individual fitness, population dynamics, and community structure, relatively few studies have tested for this behavior. Moreover, these studies have rarely assessed the mode of detection of predators, compared responses in prey species that vary in vulnerability to predators, or tested for the behavior in natural habitats. In an outdoor artificial pool experiment, we tested the oviposition responses of two dipteran species, Culiseta longiareolata (mosquito) and Chironomus riparius (midge), to the hemipteran predator, Notonecta maculata. Both dipteran species have similar life history characteristics, but Culiseta longiareolata larvae are highly vulnerable to predation by Notonecta, while Chironomus riparius larvae are not. As their vulnerabilities would suggest, Culiseta longiareolata, but not Chironomus riparius, strongly avoided ovipositing in pools containing Notonecta. An experiment in natural rock pools assessing oviposition by Culiseta longiareolata in response to Notonecta maculata yielded an oviposition pattern highly consistent with that of the artificial pool experiment. We also demonstrated that the cue for oviposition avoidance by Culiseta longiareolata was a predator-released chemical: Notonecta water (without Notonecta replenishment) repelled oviposition for 8 days. Oviposition avoidance and mode of detection of the predator have important implications for how to assess the true impact of predators and for the use of commercially produced kairomones for mosquito control.     

Oviposition habitat selection by the mosquito Culiseta longiareolata in response to risk of predation and conspecific larval density

Abstract. 1. Most female Culiseta longiareolata (Diptera: Culicidae) avoid ovipositing in pools that contain the predatory backswimmer Notonecta maculata . Such oviposition habitat selection has been suggested to reflect a trade-off between the risk of predation on larvae and potential density-dependent fitness costs. This putative trade-off was examined. In particular, evidence was sought in support of direct female response to local heterogeneity in habitat quality.
2. Three habitat types were established using artificial outdoor pools: predator pools, and non-predator pools with either low or high densities of Culiseta larvae. During each experimental night, females were offered one of the three possible pair-wise treatment combinations.
3. The majority (≈88%) of females oviposited in low-density pools rather than in the predator- or high-density pools. Furthermore, a substantially higher proportion of females oviposited in predator pools when faced with the high-density alternative, however this was due largely to fewer females ovipositing in high- vs low-density pools.
4. Females of a second mosquito species ( Culex laticinctus ), the larvae of which are at a lower risk of predation, were predicted to exhibit weaker aversion to N. maculata ; this prediction was supported only weakly.
5. Oviposition habitat selection by female C. longiareolata does not appear to involve a behavioural response that is based on individual assessment of local heterogeneity in relative pool quality, at least not at the spatial scale examined here; alternative explanations are discussed.     

Chemical detection of the predator Notonecta irrorata by ovipositing Culex mosquitoes.

We tested the oviposition response of Culex mosquitoes to the predator Notonecta irrorata in an outdoor artificial pool experiment employing equal numbers of control and predator pools. There was a strong oviposition avoidance by Culex of Notonecta pools; 83% of egg rafts were found in control pools during the period in which Notonecta were present. After removing Notonecta, mosquitoes continued to avoid ovipositing in the former Notonecta pools for two additional days suggesting a predator-released kairomone as the cue used by the mosquitoes to detect the presence of this predator.     

Influence of the predatory backswimmer, Notonecta maculata, on invertebrate community structure

1. The impact of the predator, Notonecta maculata , on invertebrate community structure was investigated by a combination of a natural pool survey, an outdoor artificial pool experiment, and a laboratory prey preference experiment.
2. In a survey of forty natural pools in the Negev Desert, Israel, pelagic/neustonic species richness and overall species richness, but not benthic species richness, were negatively associated with N. maculata density.
3. In the artificial pool experiment, in which the presence or absence of Notonecta was manipulated, Notonecta caused a reduction in the richness of species colonizing the pools. Notonecta strongly reduced densities of the pelagic dipterans, Culiseta longiareolata (often 100% reduction of later-instar larvae and pupae) and ephydrids. Fewer Culiseta egg rafts were found in Notonecta pools, indicating that part of the observed reduction of Culiseta is not due to consumption of Culiseta larvae by Notonecta, but to habitat selection by ovipositing Culiseta females in response to risk of predation by Notonecta . Notonecta did not affect densities of (benthic) chironomids or the small cladoceran, Ceriodaphnia cf. quadrangula . Laboratory assessments of selectivity by Notonecta for the various prey matched patterns found in the field experiment.
4. This work provides strong evidence that N. maculata is an important organizer of community structure: it strongly reduces or even eliminates larger pelagic or neustonic species, but does not affect densities of small or benthic species.     

Modelling evolutionarily stable strategies in oviposition site selection, with varying risks of predation and intraspecific competition

Many ovipositing mosquitoes, as well as other species, can detect biotic factors that affect fitness. However, a female mosquito seeking a high quality oviposition site (e.g. one with low risk of predation and competition to her progeny) must often balance the competing risk of increasing probability of mortality to herself while she continues to search, against increased probability of finding a high quality site. Such oviposition site selection may affect adult population size. We examined a female mosquito’s expected strategy of oviposition site selection under conditions of varying predator prevalence and adult mortality risk, by combining a detailed structured population model with a Markov chain implementation of the adult behavioural process. We used parameter values from the specific mosquito-predator system, Culiseta longiareolata-Notonecta maculata, although the overall results can be generalised to many mosquito species. Our model finds the evolutionarily stable strategy of oviposition site selection for different parameter combinations. Our model predicts that oviposition strategy does not vary smoothly with varying environmental risk of adult mortality, but that certain oviposition strategies become unstable at some parameter values. Mosquitoes will distribute their reproductive effort between breeding sites of varying predation risk only when adult mortality is low or larval competition high. Our model predicts that females will continue searching for predator-free pools, rather than oviposit in the first site encountered, regardless of the risk of mortality to the adult. The ecological basis for a reproductive strategy with alternative behaviours is important for understanding the effect of biotic factors on the population dynamics of mosquitoes, and for the development of biological control strategies, such as the dissemination of predator-cue chemicals.     

Predation-dependent oviposition habitat selection by the mosquito Culiseta longiareolata: a test of competing hypotheses

We investigated the mechanism underlying oviposition habitat selection (OHS) in the mosquito Culiseta longiareolata. The putative outcome of a trade‐off between the risk of predation and detrimental density dependence, OHS in this species presents an opportunity to test two competing alternatives: (1) a polymorphic scenario, in which a fixed proportion of females constantly avoid ‘predator pools’, while the remainder oviposits at random; and (2) a monomorphic scenario, in which all females oviposit in predator pools with a certain probability. We present a conceptual framework that demonstrates how a simple experimental design – whereby predator incidence in artificial pools is alternated between 0.25 and 0.75 – can distinguish between, or refute, the two scenarios. Given the proportional use, by ovipositing females, of predator‐free pools observed under each treatment, and a bootstrap estimate of the ratio of daily oviposition rates, we find the monomorphic scenario twice as likely as the polymorphic.     

Effects of Anisops sardea (Hemiptera: Notonectidae) on oviposition habitat selection by mosquitoes and other dipterans and on community structure in artificial pools

We investigated the effects of the predatory backswimmer, Anisops sardea, on oviposition habitat selection of mosquitoes and other dipterans, and on community structure in experimental pools. We predicted that those dipteran species whose larvae were shown to be highly vulnerable to predation by Anisops would avoid Anisops pools when choosing an oviposition site. We established the following treatments in plastic tubs: (1) Control (without Anisops); (2) Free Anisops (ten Anisops within pool but not caged); (3) Caged Anisops (ten Anisops in cage). The pools were open to colonization by insects. We added resting stages of crustaceans and first instar larvae of the mosquitoes Culiseta longiareolata and Culex laticinctus. Among the dipteran species, Culiseta longiareolata, Culex laticinctus, Chironomus riparius (Chironomidae) and Forcipomyia sp. (Ceratopogonidae), only Culisetalarvae were highly vulnerable to predation. As predicted, based on larval vulnerability, Culiseta, but not the other species, avoided Anisops pools when ovipositing. Free Anisops reduced taxon richness. This reduction resulted largely from the elimination of the cladoceran Ceriodaphniasp. and Culiseta in most free Anisops pools. Thus, Anisops sardea structures the community, both by a behavioral response of prey to its presence and by consumption of prey.     

Interactive effects of salinity and a predator on mosquito oviposition and larval performance

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. 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.     

Can adults of the New Zealand mosquito Culex pervigilans (Bergorth) detect the presence of a key predator in larval habitats?

The influence of predators on mosquito populations may be direct through predation or indirect through sub‐lethal responses of adult mosquitoes in life history traits such as oviposition behavior. In New Zealand, the backswimmer, Anisops wakefieldi, is a common predator of mosquito larvae found in temporary and permanent water bodies. We predicted that the New Zealand native mosquito, Culex pervigilans, whose larvae are vulnerable to predation of Anisops, would likely avoid the containers with the presence of Anisops or its kairomone. We established temporary water containers without predators, free‐roaming predators, caged predators (which were unable to eat mosquitoes), or containers from which predators were removed immediately prior to the experiment (these containers would have remnant kairomones from the predators). Each treatment with Anisops had predator densities of one, three, or nine Anisops. Contrary to our predictions, when choosing oviposition habitats, Cx. pervigilans appeared to ignore the presence of free‐roaming Anisops, caged Anisops, and water with Anisops kairomone. We thus observed no significant differences between the numbers of egg rafts laid by Cx. pervigilans in the different predator treatments nor were the number of egg rafts significantly affected by the density of predators. Rather than the presence of predators, environmental factors including temperature, humidity, and pressure were significantly correlated with mosquito oviposition. These mosquitoes appeared to either ignore the presence of the predator, had no ability to detect the presence of predators, or perhaps the cues from Anisops predators were not sufficiently strong enough to alarm these mosquitoes. We argue that the mosquito has not evolved the ability to detect the presence of these predators while ovipositing.     

Oviposition habitat selection by the mosquito, Culiseta longiareolata: effects of conspecifics, food and green toad tadpoles

Abstract. 1. Ovipositing females should choose egg-laying sites which maximize the growth and survivorship of their offspring. Accordingly, they should avoid habitats with high densities of competitors and predators and should choose sites with adequate food. We tested experimentally the oviposition response of the mosquito, Culiseta longiareolata Macquart, to extra food and two species which are both potential competitors and predators of its offspring: conspecific larvae and green toad ( Bufo viridis ) tadpoles.
2. We conducted the experiment in outdoor artificial pools where wild C. longiareolata females were presented with choices of all combinations of the presence or absence of: (1) C.longiareolata larvae, (2) B.viridis tadpoles, and (3) extra food (a mixture of ground fish food and mouse chow).
3. Oviposition rates averaged approximately 4 times greater in response to extra food.
4. The presence of the tadpoles, particularly later in the experiment when they were larger, caused a significant reduction (approximately 50%) in oviposition rate.
5. We did not detect a statistically significant effect of conspecific larvae on oviposition habitat selection.
6. Our results indicate that Culiseta females choose habitats for oviposition based on food availability. They also tend to avoid habitats containing Bufo tadpoles, presumably to avoid predation and/or competition.     

Direct and indirect effects of a predatory backswimmer (Notonecta maculata) on community structure of desert temporary pools

Abstract.

• 1 We assessed experimentally the effects of the predatory backswimmer, Notonecta maculata, on naturally colonizing mosquito populations in artificial outdoor pools in the Negev Desert, Israel. A single Notonecta adult per pool (8–15 litres water) had a very large negative impact on populations of Culiseta longiareolata, the most common species found in natural local pools. Notonecta caused large reductions of Culiseta egg rafts and early-instar larvae (instars I and II) and virtually 100% reductions of late-instar Culiseta larvae (instars III and IV) and pupae.
• 2 Notonecta also caused a trophic cascade in the experimental pools; by preying on periphyton-feeding Culiseta larvae, Notonecta indirectly caused significantly higher densities of diatoms, the major component of the periphyton.
• 3 Surveys of nearby natural pools taken between March and May supported the experimental results: a strong negative association between Notonecta and Culiseta among pools occurred as Notonecta increased in numbers and became more widely distributed. Anopheles (occurring only in May) and Culex mosquito immatures were not negatively associated with Notonecta. Culiseta was not associated with surface vegetation whereas both Culex and Anopheles showed strong positive associations with surface vegetation both among and within pools. We attribute the negative association between the predator and Culiseta to local prey extinctions caused by Notonecta in individual pools. Culiseta, being an open water species, is apparently more prone to predation by Notonecta than the vegetation-dwelling Culex and Anopheles.

     

Influence of fire salamander larvae on among-pool distribution of mosquito egg rafts: oviposition habitat selection or egg raft predation?

Many amphibian populations are in decline worldwide. Surprisingly, few studies have examined how such declines may benefit mosquitoes. Amphibian larvae may compete with and prey upon mosquito larvae, and may alter oviposition habitat selection (OHS) of mosquito adults. However, often overlooked, observed among-pool egg distributions attributed to OHS may additionally or alternatively be explained by egg predation. Temporary pools of mountainous areas of the Mediterranean serve as larval habitat for both the mosquito, Culiseta longiareolata, and the salamander, Salamandra infraimmaculata. We found Culiseta larvae and egg rafts to be highly vulnerable to predation by pre-metamorphosing Salamandra larvae, but not to metamorphosing ones. In outdoor mesocosm experiments, oviposition avoidance by Culiseta females in response to caged Salamandra was not demonstrated regardless of salamander developmental stage. Egg raft abundance was significantly reduced in free-roaming, pre-metamorphosing Salamandra but not by metamorphosing ones. Thus, Salamandra larvae may have little deterrence on Culiseta oviposition. Instead, fewer egg rafts are attributed largely to egg predation. This study highlights the importance of egg raft predation in addition to OHS when interpreting the influence of predators on prey egg distributions. It also highlights that a cost of declining amphibian populations is their reduced impacts on mosquito populations.     

Experimental evidence for predation risk sensitive oviposition by a mosquito, Culiseta longiareolata

1. Females should choose to oviposit in habitats where risk of predation and competition are low. The ovipositional responses of a mosquito, Culiseta longiareolata, to a predator and to species sharing the same trophic level as this mosquito (controphic species) were assessed experimentally in outdoor artificial pools. 2. The predator, larval Anax imperator, which strongly reduced larval C. longiareolata survival, resulted in a 52% reduction of C. longiareolata egg rafts. The controphic species (primarily Daphnia magna), which had a small but statistically significant negative effect on the survival of C. longiareolata larvae, did not have a statistically significant influence on the number of egg rafts. 3. Laboratory trials indicated that only a small fraction of the reduced number of egg rafts seen in predator pools may be due to consumption of the egg rafts by A. imperator. 4. The experimental evidence indicates that the reduced number of C. longiareolata egg rafts found in the presence of A. imperator is due largely to oviposition habitat selection, i.e. C. longiareolata females choose pools with low risk of predation for their offspring.     

Rapid habituation by mosquito larvae to predator kairomones

Larvae of some species of mosquitoes have been shown to respond to water‐borne kairomones from predators by reducing bottom‐feeding and replacing it with surface filter‐feeding, which uses less movement and is thus less likely to attract a predator. However, if no predator attack takes place, then it would be more efficient to use a risk allocation strategy of habituating their response depending on the predator and the overall risk. The larvae of Culiseta longiareolata Macquart live in temporary rain‐filled pools, where they are exposed to a high level of predation. Within one hour, they responded to kairomones from dragonfly or damselfly nymphs, or to the fish Aphanius, by significantly reducing bottom‐feeding activity. Continued exposure to the predator kairomones resulted in habituation of their response to damselflies, a slower habituation to fish, but no habituation to dragonflies even after 30 h. In contrast, the larvae of Culex quinquefasciatus Say normally live in highly polluted and thus anaerobic water, where the predation risk will be much lower. They also showed a significant reduction in bottom‐feeding after 1 h of exposure to predator kairomones but had completely habituated this response within 6 h of continuous exposure. Some species of mosquito larvae can thus show a very rapid habituation to predator kairomones, while others only habituate slowly depending on the predator and overall predation risk.     

Predator-prey interactions and community structure: chironomids,mosquitoes and copepods in Heliconia imbricata (Musaceae)

Summary Evidence from both field observations and experimental work indicates that predation by larvae of a midge, Pentaneura n. sp. (Chironomidae), causes the low densities of mosquito larvae (Culicidae) found in the water filled bracts of Heliconia imbricata (Musaceae), microhabitats typically colonized by mosquitoes. This predation affects 2 species of mosquitoes, Wyeomyia pseudopecten, a resident species, and Trichoprosopon digitatum, a non-resident species. Predation keeps resident mosquito densities low while completely excluding the nonresident mosquito from the habitat. Both these effects of predation depend on the presence of an abundant alternative prey, an undescribed species of harpacticoid copepod found in the bracts. These copepod prey sustain chironomids when resident mosquito densities are low, permiting predator densities to remain high enough to exclude the non-resident mosquito. I discuss the evolutionary and ecological implications of predation structuring communities.     

Oviposition habitat selection by Anopheles gambiae in response to chemical cues by Notonecta maculata

A number of mosquito species avoid predator-inhabited oviposition sites by detecting predator-released kairomones. In the laboratory, we found that when offered de-ionized water and de-ionized water conditioned with Notonecta maculata, gravid Anopheles gambiae females preferentially oviposited into the former. We then conducted further experiments using two chemical components found in Notonecta-conditioned water, chemically pure n-tricosane and/or n-heneicosane, that was previously shown to repel oviposition by Culiseta longiareolata. These hydrocarbons failed to deter oviposition by An. gambiae females. Thus, different mosquito species may rely on distinct chemical cues to avoid predators. Identification and chemical characterization of such kairomones could facilitate innovative, environmentally sound mosquito control.     

Mosquito female response to the presence of larvivorous fish does not match threat to larvae

1. Several species with complex life‐history traits such as amphibians and insects with aquatic immature stages and terrestrial adults avoid ovipositing in pools containing larvivorous fish. This avoidance response was assumed to be a general one for most fish species. 2. The generality of ovipositing Culex to the presence of three, widespread larvivorous fish species was tested in a set of field experiments with artificial oviposition pools using caged fish. 3. Larval performance was further examined under actual predation by these three fish species. 4. Results show that ovipositing females responded strongly to the presence of caged mosquitofish, Gambusia affinis, while showing no significant response to the presence of caged green sunfish, Lepomis cyanellus, or the pirate perch, Aphredoderus sayanus. All three fish species consumed similar amounts of larvae. 5. This is the first example of species‐specific response differences to predators during mosquito oviposition habitat selection. These results point to the existence of predator‐released kairomones affecting mosquito behaviour. These kairomones may either be species‐specific or vary in concentration among fish, and probably have an important role in the understanding of mosquito spatial distribution.     

Attracted to the enemy: Aedes aegypti prefers oviposition sites with predator-killed conspecifics

Oviposition habitat choices of species with aquatic larvae are expected to be influenced by both offspring risk of mortality due to predation, and offspring growth potential. Aquatic predators may indirectly influence growth potential for prey by reducing prey density and, for filter-feeding prey, by increasing bacterial food for prey via added organic matter (feces, partially eaten victims), creating the potential for interactive effects on oviposition choices. We tested the hypothesis that the mosquito Aedes aegypti preferentially oviposits in habitats with predatory Toxorhynchites larvae because of indirect effects of predation on chemical cues indicating bacterial abundance. We predicted that A. aegypti would avoid oviposition in sites with Toxorhynchites, but prefer to oviposit where bacterial food for larvae is abundant, and that predation by Toxorhynchites would increase bacterial abundances. Gravid A. aegypti were offered paired oviposition sites representing choices among: predator presence; the act of predation; conspecific density; dead conspecific larvae; and bacterial activity. A. aegypti preferentially oviposited in sites with Toxorhynchites theobaldi predation, and with killed conspecific larvae, but failed to detect preferences for other treatments. The antibiotic tetracycline eliminated the strongest oviposition preference. Both predation by Toxorhynchites and killed larvae increased bacterial abundances, suggesting that oviposition attraction is cued by bacteria. Our results show the potential for indirect effects, like trophic cascades, to influence oviposition choices and community composition in aquatic systems. Our results suggest that predators like Toxorhynchites may be doubly beneficial as biocontrol agents because of the attraction of ovipositing mosquitoes to bacterial by-products of Toxorhynchites feeding.     

Predation by odonates depresses mosquito abundance in water-filled tree holes in Panama

In the lowland moist forest of Barro Colorado Island (BCI), Panama, larvae of four common species of odonates, a mosquito, and a tadpole are the major predators in water-filled tree holes. Mosquito larvae are their most common prey. Holes colonized naturally by predators and prey had lower densities of mosquitoes if odonates were present than if they were absent. Using artificial tree holes placed in the field, we tested the effects of odonates on their mosquito prey while controlling for the quantity and species of predator, hole volume, and nutrient input. In large and small holes with low nutrient input, odonates depressed the number of mosquitoes present and the number that survived to pupation. Increasing nutrient input (and consequently, mosquito abundance) to abnormally high levels dampened the effect of predation when odonates were relatively small. However, the predators grew faster with higher nutrients, and large larvae in all three genera reduced the number of mosquitoes surviving to pupation, even though the abundance of mosquito larvae remained high. Size-selective predation by the odonates is a likely explanation for this result; large mosquito larvae were less abundant in the predator treatment than in the controls. Because species assemblages were similar between natural and artificial tree holes, our results suggest that odonates are keystone species in tree holes on BCI, where they are the most common large predators. Received: 4 November 1996 / Accepted: 11 April 1997