Integration of multiple intraguild predator cues for oviposition decisions by a predatory mite.

In mutual intraguild predation (IGP), the role of individual guild members is strongly context dependent and, during ontogeny, can shift from an intraguild (IG) prey to a food competitor or to an IG predator. Consequently, recognition of an offspring's predator is more complex for IG than classic prey females. Thus, IG prey females should be able to modulate their oviposition decisions by integrating multiple IG predator cues and by experience. Using a guild of plant-inhabiting predatory mites sharing the spider mite Tetranychus urticae as prey and passing through ontogenetic role shifts in mutual IGP, we assessed the effects of single and combined direct cues of the IG predator Amblyseius andersoni (eggs and traces left by a female on the substrate) on prey patch selection and oviposition behaviour of naïve and IG predator-experienced IG prey females of Phytoseiulus persimilis. The IG prey females preferentially resided in patches without predator cues when the alternative patch contained traces of predator females or the cue combination. Preferential egg placement in patches without predator cues was only apparent in the choice situation with the cue combination. Experience increased the responsiveness of females exposed to the IG predator cue combination, indicated by immediate selection of the prey patch without predator cues and almost perfect oviposition avoidance in patches with the cue combination. We argue that the evolution of the ability of IG prey females to evaluate offspring's IGP risk accurately is driven by the irreversibility of oviposition and the functionally complex relationships between predator guild members.

Offspring predation risk is a major force driving the evolution of oviposition behaviour (e.g. Thompson & Pellmyr 1991; Murphy 2003; Rieger et al. 2004). To maximize fitness, ovipositing females should be able to evaluate their offspring's predation risk precisely and adjust oviposition site selection accordingly (e.g. Kats & Sih 1992). This is especially true for species without extended parental care because in such species selective egg placement is often the only means to enhance offspring survival. Oviposition site selection based on offspring's predation risk is well documented for classic predator–prey interactions in both terrestrial (Grostal & Dicke 1999; Mira & Bernays 2002; Moon & Stiling 2006) and aquatic systems (Murphy 2003; Blaustein et al. 2004; Rieger et al. 2004; Hirayama & Kasuya 2009), but has been rarely assessed in predator–predator interactions such as intraguild predation (IGP; Walzer et al. 2006; Walzer & Schausberger 2011).

Selective egg placement implies accurate perception and interpretation of cues signalling the presence of an offspring's predator because recognition errors may have fatal consequences for the offspring. Classic predator cues always indicate predation risk for prey but this is not the case for cues emitted from a predator guild member (Durant 2000). In mutual IGP systems, the guild members commonly pass through an ontogenetic role reversal from intraguild (IG) prey to a food competitor or IG predator or both (Polis & Holt 1992). For example, juvenile guild members are commonly IG prey of larger and more mature guild members, but are only food competitors of guild members in the same developmental stage. If a juvenile individual reaches the next developmental stage earlier than another similarly aged guild member, then the developmentally more advanced, larger individual becomes both a food competitor and an IG predator of the smaller, developmentally less advanced individual, which shifts to a potential IG prey. The function and interpretation of guild member-associated cues, however, is influenced not only by the life stages involved (Walzer et al. 2004), but also by the structural complexity of the habitat (e.g. Seelmann et al. 2007) and availability of shared prey (Walzer & Schausberger 1999a). Additionally, IG predator cues may indicate the presence of shared prey and guide another predator to the shared food source (e.g. Pangle & Holekamp 2010). Thus, the function of these cues is more complex and strongly context dependent than in classic predator–prey interactions rendering the interpretation of single guild member-associated cues ambiguous. Consequently, in mutual IGP systems, IG prey females should be selected for the ability to integrate multiple cues in IG predator recognition.

The integration of multiple cues in predator assessment by an ovipositing female may be advantageous for several reasons. Multiple cues may provide information about different aspects of the IG predator leading to cue-specific behavioural responses of IG prey (multiple message hypothesis; Johnstone 1996). Alternatively, multiple cues may provide the same information and allow increased accuracy of the IG prey response (backup signal hypothesis; Johnstone 1996; Rowe 1999; Hebets & Papaj 2005). Multiple cues may also interact with each other (intersignal interaction hypothesis; Hebets & Papaj 2005) together providing new information for prey and consequently modifying or intensifying a given antipredator behaviour (for classic prey–predator interactions: Smith & Belk 2001; Wisenden et al. 2003; Amo et al. 2004; Cooper 2009). These hypotheses were originally developed for mate and kin recognition contexts (Hebets & Papaj 2005) but they are similarly applicable to IG predator recognition and anti-IG predator responses. Only little is known about the type of cues mediating IG predator recognition by IG prey (Eiben & Persons 2007; Chauhan & Weber 2008; Kishida et al. 2009; Seagraves 2009; Webb et al. 2009; Choh et al. 2010). In principle, such cues may be direct cues emitted by the predators themselves, acting as kairomones, and indirect cues emanating from IG or extraguild prey such as alarm pheromones released upon attack or injury from killed individuals, or a combination of cues.

We studied the effects of single and combined direct predator cues on prey patch and oviposition site selection by IG prey females in a guild of plant-inhabiting predatory mites (Acari: Phytoseiidae) consisting of the IG prey Phytoseiulus persimilis and the IG predator Amblyseius andersoni. These two species and another predatory mite, Neoseiulus californicus, constitute a natural guild in Sicily and elsewhere (De Moraes et al. 2004; Walzer & Schausberger 2011) sharing two-spotted spider mites Tetranychus urticae (Acari: Tetranychidae) as prey (e.g. McMurtry & Croft 1997). The guild members can mutually prey on each other but IGP is asymmetric with respect to IGP strength with A. andersoni being the strongest and P. persimilis being the weakest IG predator (Walzer & Schausberger 2011). For each species, larvae are the preferred IG prey, whereas gravid females are the most pronounced IG predators. Gravid females usually do not prey on each other owing to size constraints. A recent study revealed that the eyeless P. persimilis is able to discriminate chemosensory cues from the low-risk and high-risk IG predators, N. californicus and A. andersoni, respectively (Walzer & Schausberger 2011). Predator recognition induced threat-sensitive prey patch selection, oviposition site selection and total egg production in experienced P. persimilis females (reared in the presence of IG predators during juvenile development; Walzer & Schausberger 2011). Previous IG predator presence may have been indicated by direct IG predator cues (eggs or traces left by the IG predator female on the substrate or both cues) or by indirect IG predator cues (alarm pheromones of the extraguild prey). However, indirect IG predator cues such as the smell from killed individuals may have been additional cues (Grostal & Dicke 1999) but fall short of explaining interspecific threat sensitivity in IG prey response because such cues are present with every predator including conspecifics. Our objectives were to determine whether single direct IG predator cues from the high-risk IG predator A. andersoni, on eggs and traces such as metabolic waste products or chemical footprints left by ovipositing IG predator females on the substrate, or both cues, are needed for IG predator recognition by IG prey females of P. persimilis and whether the response of P. persimilis is changed by experience with the IG predators.

Methods

Species Origin and Rearing

Specimens of P. persimilis and A. andersoni used to found laboratory-reared populations were collected from herbs and apple trees in the state of Trapani, Sicily, in 2007. In the laboratory, the two species were separately reared on arenas consisting of plastic tiles resting on water-saturated foam cubes in plastic boxes half-filled with water (Walzer & Schausberger 2011). The predators were fed in 2–3 day intervals with two-spotted spider mites, T. urticae, reared on whole bean plants, Phaseolus vulgaris, by adding bean leaves infested with spider mites (for P. persimilis) or by brushing spider mites from infested leaves (for A. andersoni) onto arenas.

Experimental Procedures

To determine which cues mediate prey patch and oviposition site selection of IG predator-naïve and IG predator-experienced females of P. persimilis we conducted binary choice experiments using a two (female status: naïve or experienced) by four (choice alternatives to a prey patch with only live spider mites) factorial design.

Generating Naïve and Experienced IG Prey Females

To obtain naïve and experienced IG prey females of P. persimilis for experiments, two even-aged groups of 13–15 eggs each were randomly taken from the rearing unit and placed on separate leaf arenas. Each leaf arena consisted of a detached bean leaf (5 × 5 cm) placed upside down on a water-saturated foam cube in a plastic box half-filled with water. Water-saturated cellulose strips (1 cm height) at the edge of the leaf confined the arena and prevented the mites from escaping. The first group of eggs (to be used as naïve IG prey females in experiments) was placed on a leaf with surplus spider mites; the second group of eggs (to be used as experienced IG prey females in experiments) was placed on a leaf with surplus spider mites and five females of the IG predator A. andersoni (Walzer & Schausberger 2011). During rearing the IG predator females mainly preyed on spider mites but also killed some IG prey and produced eggs. Mortality of IG prey was about 15–20% higher in rearing units with IG predators than without but the set-up provided for random but not selective IGP (A. Walzer & P. Schausberger, unpublished data). Therefore, during development IG prey that were to be used as experienced females in experiments were exposed to all direct (mobile predators, their eggs, chemical footprints and metabolic waste products) and indirect (killed conspecifics and killed spider mites) volatile and tactile chemosensory cues possibly indicating IG predator presence (Walzer et al. 2006; Walzer & Schausberger 2011). The developmental progress of the P. persimilis individuals to be used as IG prey in experiments was observed and spider mite prey were replenished daily. After 10 days the IG prey females had reached adulthood and were mated and ready to be subjected to choice experiments.

Choice Experiments

Each experimental choice unit consisted of two similarly sized leaflets (5–7 cm2), taken from trifoliate bean leaves, placed upside down on the same foam cube in a plastic box (15.0 × 10.0 cm and 4.0 cm high) half-filled with water and connected with a wax bridge (1.0 × 0.5 cm) (Walzer et al. 2006; Walzer & Schausberger 2011). This set-up simulates a natural situation in which a gravid female searches for a suitable prey patch and oviposition site on a branch within a plant. In each choice situation one leaflet only harboured live spider mites, T. urticae (the extraguild prey), whereas the second leaflet harboured live and dead spider mites, plus either (1) IG predator eggs and traces of an IG predator female, or (2) traces of an IG predator female, or (3) IG predator eggs, or (4) no direct IG predator cues. Each choice situation may be encountered in nature, including (3), because traces left by phytoseiid females are partially volatile and hence dissipate over time (e.g. Janssen et al. 1997).

For the pre-experimental preparation of the prey patches the bridge connecting the two leaflets was blocked with a strip of moist tissue paper. Each leaflet of each choice situation first received 30 juvenile and four to seven adult T. urticae females. For treatments (1) and (2), after 24 h, five A. andersoni females from the rearing units were added onto the second leaflet and allowed to prey on spider mites and to oviposit. After another 24 h, the T. urticae females on all leaflets of all choice situations and the IG predator females in treatments (1) and (2) were removed and the IG predator eggs either reduced to five per leaflet (1) or completely removed (2). Furthermore, five IG predator eggs were placed on the second leaflet in treatment (3) using a fine brush. In treatments (3) and (4) spider mite predation by the IG predator was mimicked by puncturing and killing 30 juveniles and 20 eggs of the spider mites using a needle (Janssen et al. 1997). The number of killed prey corresponds to the average daily predation rate of five A. andersoni females (A. Walzer & P. Schausberger, unpublished data). Before starting the choice experiment, we adjusted the density of live spider mites to identical levels on both leaflets (60 eggs and 20 juveniles per leaflet) within each choice unit. This spider mite density allowed the IG prey female to reach the maximum oviposition rate during the 24 h experimental period and leave enough prey for her offspring to reach adulthood on each leaflet (Vanas et al. 2006; A. Walzer & P. Schausberger, unpublished data). Therefore, differences in prey patch choice and oviposition behaviour can be delimited to the presence or absence of IG predator cues and are interpreted as anti-IG predator but not merely anticompetitor behaviour.

Before starting the choice experiment we transferred single P. persimilis females into closed acrylic cages and left them without food for 12 h. Each cage consisted of a cylindrical cell (15 mm diameter and 3 mm height) with a fine mesh screen at the bottom and closed on the upper side with a microscope slide (Schausberger 1997). Only females producing at least one egg during the food deprivation period were used for experiments. Subsequently, each female was singly released in the middle of the wax bridge and given a choice between a prey patch with only live spider mites and a prey patch with live and dead spider mites and direct IG predator cues. Each choice unit and each female was used only once. Each of the four choice situations with either a naïve or an experienced female was replicated 21–32 times. The position of the experimental female was observed immediately after release, and then after 1, 2, 3, 4, 5, 6 and 24 h. The repeated observations should allow us to detect time-dependent changes in IG prey behaviour and thus provide indications about the nature (contact versus volatile) of the IG predator cues. Number and between-patch distribution of eggs produced by the IG prey female were recorded after 6 and 24 h.

Statistical Analyses

All statistical analyses were carried out using SPSS 15.0.1 (SPSS Inc., Chicago, IL, U.S.A.). Generalized estimating equations (GEE, binomial distribution with identity link function, autocorrelation structure between observation points; pairwise post hoc comparisons by least significant difference, LSD, tests) were used to compare the influence of IG predator cue type (IG predator eggs or traces of IG predator females or both direct IG predator cues or only live and dead spider mites) and experience of IG prey females with IG predators (yes or no) on prey patch selection by P. persimilis females (presence in the IG predator patch or not) over time (used as within-subject covariate; eight observation points; Hardin & Hilbe 2003). Generalized linear models (GLM) were used to analyse the effect of experience (yes or no) and IG predator cue type (IG predator eggs or traces of IG predator females or both direct IG predator cues or only live and dead spider mites) on total egg production (eggs laid in both prey patches combined; gamma distribution, identity link function) and the number of eggs deposited in the IG predator patch out of all eggs produced (frequency of events, binomial distribution, logit link function, pairwise post hoc comparisons by LSD tests). Total egg production data were transformed before analysis (1/x) and had homogeneous variances (Levene test: P = 0.102). In both GEE and GLM, the choice situation live spider mites versus live and dead spider mites was used as a reference category in parameter estimation to exclude effects of indirect IG predator cues (dead spider mites).

Results

Prey patch selection was affected by IG predator cue type and the interaction between cue type and experience but not experience as the main effect. The effect of cue type and the interaction between cue type and experience varied over time (Table 1). Pairwise comparisons with the reference choice situation (live versus live and dead spider mites; Fig. 1g, h) in parameter estimation revealed that prey patch selection was not affected by IG predator eggs alone (GEE: Wald , P = 0.274; Fig. 1a, b), but was affected by the traces left by the IG predator females (Wald , P = 0.048; Fig. 1c, d) and almost significantly affected by the combination of IG predator eggs + IG predator female traces (Wald , P = 0.059; Fig. 1e, f). When confronted with IG predator female traces, more naïve females were found in prey patches with IG predator cues after 6 h, whereas the distribution of experienced IG prey females did not change in this time period. Afterwards, both naïve and experienced IG prey females switched to IG predator-free prey patches (Fig. 1c, d). The combination of IG predator eggs+IG predator female traces triggered a preference for the prey patch without IG predator cues in both experienced and naïve IG prey females. However, most experienced females immediately chose the prey patch without IG predator cues and stayed there for the remainder of the experiment (Fig. 1f), whereas naïve IG prey females moved from the prey patch with both IG predator cues to the prey patch without IG predator cues only in the course of the experiment resulting in a clear preference for the latter patch after 24 h (Fig. 1e).

Table 1

Generalized estimating equations for the influence of experience and cue type on prey patch selection by P. persimilis over time (presence/absence in patch with direct IG predator cues)

Source of variation	df	Wald χ2	P
Experience	1	00.033	0.856
Cue type	3	17.286	0.001
Experience*cue type	3	09.249	0.026
Experience*time	1	00.174	0.677
Cue type*time	3	39.596	<0.001
Experience*cue type*time	3	10.757	0.013

Generalized estimating equations: binomial distribution, identity link function, autocorrelation structure between observation points, time as covariate.

Figure 1

Prey patch selection after first choice, and 1, 2, 3, 4, 5, 6 and 24 h by (a, c, e, g) IG predator-naïve and (b, d, f, h) predator-experienced P. persimilis females given a choice between a prey patch with only live spider mites (white bars) and a prey patch with live and dead spider mites with and without direct cues of the IG predator A. andersoni (black bars). (a, b) IG predator eggs (N = 32 and 27, respectively, for naïve and experienced); (c, d) traces of IG predator females (N = 30 and 23, respectively); (e, f) IG predator eggs and traces of IG predator females (N = 25 and 26, respectively); (g, h) without direct predator cues (N = 21 and 24, respectively).

Oviposition site selection was not influenced by experience (GLM: Wald , P = 0.082), but was influenced by cue type (Wald , P < 0.001), and the interaction of cue type and experience (Wald , P = 0.035). Pairwise comparisons with the reference choice situation (live versus live and dead spider mites) in parameter estimation revealed that only the combination of both IG predator cues (Wald , P < 0.001) but not IG predator eggs alone (Wald , P = 0.172) or IG predator female traces alone (Wald , P = 0.101) induced oviposition avoidance in the prey patch with IG predator cues. Pairwise post hoc comparisons of the interaction terms revealed that the significant interaction between cue type and experience was mainly due to a significant difference (LSD: P = 0.001) between naïve and experienced females in the choice situation with both IG predator cues, whereas experience did not have a significant effect in the other choice situations (Fig. 2).

Figure 2

Proportion of eggs (mean + SD) deposited by P. persimilis females in the patch with live and dead spider mites and with or without direct IG predator cues of A. andersoni when given a choice between a prey patch with only live spider mites and a prey patch with live and dead spider mites with and without direct predator cues of A. andersoni.

In neither of the tested choice situations did IG prey females produce eggs within the first 6 h. Total egg production (within 24 h) was affected by cue type (GLM: Wald , P < 0.001), but not experience (Wald , P = 0.968) or the interaction of cue type and experience (Wald , P = 0.332). Irrespective of experience, pairwise comparisons with the reference choice situation (live versus live and dead spider mites) in parameter estimation revealed that egg production was reduced in the presence of any direct IG predator cue (traces left by IG predator females: Wald , P = 0.001; IG predator eggs alone: Wald , P < 0.001; IG predator eggs+IG predator female traces: Wald , P < 0.001; Fig. 3).

Figure 3

Total egg production (both patches combined; mean + SD) by IG predator-naïve and predator-experienced P. persimilis females given a choice between a prey patch with only live spider mites and a prey patch with live and dead spider mites with and without direct cues of the IG predator A. andersoni.

Discussion

Prey patch selection and oviposition behaviour of the IG prey females of P. persimilis were influenced by single and multiple direct IG predator cues and IG predator experience of IG prey females. Additionally, prey patch selection of P. persimilis changed over time. The females responded to each cue type but different behaviours were differently affected by cue type and experience. Irrespective of experience, oviposition by IG prey females was reduced in choice situations with IG predator eggs alone, traces of IG predator females alone and the combination of both cue types, as compared to the choice situation without any direct predator cues. Similar response to different cues allowed an increased accuracy in the IG prey female response, according to the backup signal hypothesis (Johnstone 1996). In contrast and in accordance with the multiple message and intersignal hypotheses (Johnstone 1996; Rowe 1999; Hebets & Papaj 2005), IG prey females integrated multiple direct IG predator cues in prey patch and oviposition site selection. Irrespective of experience, IG predator eggs alone had no influence on prey patch and oviposition site selection, whereas traces of IG predator females triggered a preference for prey patches without IG predator cues after 24 h. Only the combination of both IG predator eggs and traces of IG predator females affected both prey patch and oviposition site selection of IG prey females. Experience did not qualitatively change the antipredator responses but increased the responsiveness of females exposed to the IG predator cue combination. Naïve females moved from the prey patch with IG predator cues to the prey patch without IG predator cues only in the course of the experiment, whereas almost all experienced females chose the prey patch without IG predator cues immediately after release. Similarly, experienced IG prey females had a stronger preference to oviposit in prey patches without IG predator cues than naïve IG prey females.

Information from Single and Multiple IG Predator Cues

Eggs and traces left by the IG predator females on the substrate apparently conveyed different information about the IG predator. IG predator eggs alone do not pose an immediate predation risk and may indicate that the IG predator has already left the site assuming that the traces left by the IG predator A. andersoni are partially volatile (see Janssen et al. 1997 for other predatory mite species). IG predator eggs may also serve as alternative or supplementary prey for the IG prey P. persimilis (Walzer & Schausberger 1999b). Accordingly, perception of IG predator eggs alone induced oviposition reduction in IG prey females but did not affect prey patch and oviposition site selection. IG predator traces alone could also stem from an unmated female, which cannot produce eggs (Sabelis 1985). Consequently, unmated females have much lower prey needs and probably a lower motivation to kill IG prey and thus pose a lower predation risk to offspring than a gravid female. None the less, an unmated IG predator female is an immediate competitor, explaining the preference of IG prey females for prey patches without IG predator traces with and without IG predator eggs. Only the combination of both IG predator cues (on eggs and traces left by females) indicates an immediate risk and spatial and temporal proximity of a gravid IG predator. Consequently, integration of both IG predator cues led to the most elaborate antipredator response, that is, selective egg placement by IG prey females.

In general, multiple predator cues may interact in additive or synergistic fashion in relation to antipredator responses (Smith & Belk 2001; Wisenden et al. 2003; Ferrari et al. 2008). However, neither IG predator eggs nor traces left by IG predator females alone but only the cue combination influenced oviposition site selection of P. persimilis making additive or synergistic effects unlikely. More likely, perception of a single IG predator cue enhanced the attentiveness of IG prey females to additional IG predator cues. Freshly deposited IG predator eggs are always accompanied by IG predator female traces, whereas older eggs and female traces can also be present alone. Enhanced attentiveness due to perception of a single IG predator cue and perception of an additional IG predator cue altered oviposition site selection of P. persimilis. An analogous response of P. persimilis females was observed in the context of oviposition site selection and sibling cannibalism (Schausberger & Hoffmann 2008). The females preferentially added new eggs close to their own old eggs when given a choice between young and old eggs because sibling cannibalism was lower in sibling pairs with longer hatching intervals. As in the present experiments, only the combination of cues on eggs and traces left by ovipositing females, but not the single components, allowed females to discriminate between old and young eggs (Schausberger & Hoffmann 2008). Mechanistically similar antipredator responses to multiple cues have also been described for fish (Hartman & Abrahams 2000) and water beetles (Abjörnsson et al. 1997). In contrast to prey patch and oviposition site selection, IG prey females reduced egg production, irrespective of experience, in response to each IG predator cue alone and their combination. Such backup cues are particularly advantageous in noisy environments (e.g. Johnstone 1996). It seems that reduced egg production in high-risk environments is an immediate conservative innate response at a larger spatial scale (e.g. to IG predator presence on a plant or a branch of a plant) whereas subsequent integration of multiple predator cues allows more elaborate responses at a smaller spatial scale (e.g. between IG predator-free and IG predator-occupied leaves or prey patches within the same plant). Accordingly, we argue that the IG prey females perceived each choice situation as a risky environment and within these environments fine-tuned their residence and egg placement behaviours.

The combination of cues from IG predator eggs and traces left by IG predator females induced qualitatively similar (same direction) but quantitatively different (greater immediacy) behavioural patterns in naïve and experienced IG prey females. Experienced IG prey females were apparently more responsive to volatile cues emanating from IG predator patches than naïve IG prey females because most experienced IG prey females did not enter the prey patch with both IG predator cues even once during the experiment. This resulted in only four of 64 eggs deposited by experienced IG prey females in patches with IG predator cues as compared to 19 of 70 eggs by naïve females. The underlying learning mechanism could have been a sensitization-like process (repeated predator encounters during the learning phase increased the responsiveness of IG prey females to direct IG predator cues) or an associative mechanism such as operant conditioning (innate, predetermined avoidance of IG predator cues was strengthened by negative reinforcement through predator encounters or attacks during rearing).

Risk of Mutual IGP versus Classic Predation

We suggest that classic prey females that are able to perceive direct predator cues are less likely to be cue type sensitive in their response than the IG prey females observed in our study. Many predators deposit their eggs close to their prey to provide the hatching offspring with food, so that predator eggs usually pose a future threat for classic prey. Also predator traces and both cues combined always indicate a predation risk for a classic prey female and her offspring. Consequently, single direct predator cues unambiguously indicate the presence of a predator for classic prey females, whereas for IG prey females in mutual IGP systems, these cues, if perceived singly, may indicate a food source (IG prey), a food competitor or an IG predator of their offspring. For oviposition site selection, only the combination of cues, that is, both eggs and traces left on the substrate, was a reliable predictor of the immediate presence of a dangerous IG offspring predator, a gravid heterospecific female, reflecting the functional complexity of IG predator cues and the associated need to integrate multiple cues for optimizing oviposition decisions. In contrast, in classic predator–prey interactions single direct predator cues may convey sufficient information to modulate oviposition site selection based on offspring predation risk as, for example, has been documented for spider mites (Grostal & Dicke 1999), phantom midges (Berendonk 1999) and mosquitoes (Blaustein et al. 2004; Hurst et al. 2010). None the less, to verify our conclusions on single and multiple cue use, future studies should experimentally compare the response of the predatory mites to cues of classic or unidirectional IG predators such as ladybird beetles (Putman 1955) and mutual IG predators such as heterospecific predatory mites.