Risk of parasitism
A reduction of predation has been reported as an advantage of gregarious life styles in several animal taxa (Turchin & Kareiva, 1989; Uetz et al., 2001; Cresswell, 2002; Shiojiri & Takabayashi, 2003; Spieler, 2003). One proposed benefit of aggregations against predators is the avoidance effect (Hamilton, 1971; Turner & Pitcher, 1986). These authors suggest that a predator is less likely to find a particular aggregation rather than all the member individuals if they were scattered in space (Hamilton, 1971; Turner & Pitcher, 1986). Another proposed benefit of aggregations is the dilution effect or selfish herd, which suggests that the risk of predation for a single individual in an aggregation is lower than for an individual in solitude (Hamilton, 1971; Turner & Pitcher, 1986). Our results showed that A. grossa females experienced a reduction in the probability of egg mass parasitism when aggregated. Hence, egg-guarding females may benefit from both the avoidance effect and selfish herd dilution effect against parasitoids when aggregated. Female aggregations imply that females concentrate on one plant and do not put their eggs at risk by scattering them individually among nearby host plants. In addition, aggregated females may also benefit from the group if an attacking parasitoid can only parasitise a certain number of egg masses and then departs leaving other egg masses untouched. For instance, Plutella xylostella (L.) (Lepidoptera: Plutellidae) larvae benefit from a gregarious lifestyle provided that, regardless of how many larvae are present on the plant, its parasitoids (Cotesia plutellae Kurdjumov, Hymenoptera: Braconidae) parasitise only one or two individuals before leaving (Shiojiri & Takabayashi, 2003).
Parasitoids search for potential hosts using several signals, of which chemical cues seem to be the most important (Vinson, 1976; Fatouros et al., 2008; Wickremasinghe & Van Emden, 2008). These chemical signals have been reported to be those of the host's host plant, host plant emissions induced by feeding activity of the host, and host insect itself (Vinson, 1976; Fatouros et al., 2008). The present study shows that A. grossa egg-guarding females densely distributed over groups of host plants decreased the risk of parasitism on their egg masses. This contradicts what might be expected given that aggregations of females on clumped host plants may represent an increase in emission of chemical signals, and consequently, potentially increase parasitoid attraction. This apparent contradiction could be as a result of the added effect we found of egg-guarding female and host plant spatial aggregation on increasing the number of egg-guarding females on one plant. As mentioned above, female aggregation was the principal variable reducing the risk of parasitism. As a result, the effect of aggregation of egg-guarding females on reducing parasitism risk might be overpowering the effect of host plant and female spatial grouping on the attraction of parasitoids by concentrated chemical cues.
Determinants of the level of egg-parasitism
Subsocial maternal care strategies in treehoppers have been reported to be effective against predatory attacks but not against parasitoids (Eberhard, 1986; Perotto et al., 2002). In the present study with A. grossa, both maternal defensive behaviour and egg mass coverage decreased the level of parasitism. Alchisme grossa is a member of the tribe Hoplophorionini which has the highest degree of maternal care investment among the treehoppers (McKamey & Deitz, 1996). This highly developed maternal care may have evolved when treehoppers moved into higher altitudes where there was not enough ant diversity to support ant-membracid mutualism interactions for offspring defence (Wood, 1984). However, it should be noted that some hoplophorionine species occurring at low elevations, where ants are ubiquitous, also lack mutualistic relationships with ants (Wood, 1984). Thus, for A. grossa, subsocial maternal care is an adaptive response to increase offspring survival (Wood, 1984). The decreasing parasitism levels with increasing defensive behaviour observed in this study with A. grossa are congruent with selection pressure from predation and the evolutionary trends within the tribe. Moreover, as in all members in the tribe, A. grossa's maternal care also has associated morphological adaptations. The pronotum completely covers the insect dorsally, forming a shield-like structure. This is of great importance as it provides the female with a wide, solid dorsal surface under which she can hide and protect her egg mass (Wood, 1993). This is consistent with the trend of a decrease in parasitism with an increasing proportion of egg mass coverage. Studies in different Hemiptera support this point. Some Pentatomidae and Acanthosomatidae (Heteroptera) use their broad shield-like body to cover their egg masses (Eberhard, 1975; Mappes & Kaitala, 1994). In Elasmucha grisea (L.) (Acanthosomatidae), females adjust their egg masses to a size they can cover with their bodies, and eggs that a female fails to cover are readily parasitised (Mappes & Kaitala, 1994).
Aggregated A. grossa egg-guarding females experienced a dilution effect on the parasitism levels in their egg masses. This has also been reported in the wasp Crabro cribrellifer (Packard) (Hymenoptera: Sphecidae), which experiences a reduction in parasitism levels with increasing nest aggregation (Wcislo, 1984). Wcislo (1984) also suggests that these wasps choose to aggregate their nests because it is more likely that a neighbouring nest will be parasitised, in a classic selfish herd model. If, aggregated A. grossa females on a plant represent a numerical factor that divides among their egg masses the number of eggs a parasitoid can parasitise, the overall number of parasitised eggs on a plant would remain constant, regardless of how many females are guarding their eggs on it. In this study, however, the number of parasitised eggs on a single plant as well as the proportion of parasitised eggs decreased with an increased number of egg-guarding females on the plant. This suggests that A. grossa female aggregations were not only diluting parasitoid pressure among them, but discouraging the parasitoids from oviposition.
This discouraging effect on parasitoids might be explained by the positive effect female aggregation has on egg mass coverage combined with aggressive maternal care. Aggregated females increased their proportion of egg mass coverage by laying smaller egg masses, and benefitted from this behaviour through reduced parasitism. Additionally, A. grossa egg-guarding females sent substrate-borne alarm signals by drumming with their abdomen on the branch they were located on when threatened (L. Camacho & C. Keil, unpublished). This signalling has also been reported in the hoplophorionine treehopper U. crassicornis (Cocroft, 1996, 1999), and other treehoppers (Cocroft, 2003, 2005; Cocroft et al., 2008). Hence, if females attacked by parasitoids sent a distress signal, this may have put other aggregated egg-guarding females on alert and increase their irritability. This could explain why aggregated females intensified their maternal defensive behaviour. This phenomenon has been reported for Metepeira incrassate F.O. Pickard-Cambridge, a social spider that uses substrate vibrations on the web to warn other spiders about predatory attacks (Uetz et al., 2001). In that case, a reduction in the predator's success was found with an increase of the number of spiders on the web (Uetz et al., 2001). Accordingly, both female coverage of the egg mass and maternal care aggressiveness had a negative relation with parasitism levels in A. grossa. Therefore, aggregating A. grossa females enhanced their maternal care capabilities generating a better defence against parasitoids. Cooperative defence among individuals in insect aggregations has been reported in social and a few non-social insects (Evans, 1990). As a result, the dilution of risk of parasitism by aggregated A. grossa females by a selfish herd effect may also include a kin selection effect. These results might give a new dimension to the term ‘sub-sociality’ in Hoplophorionini treehoppers.
Further evidence of female aggregation as a defence mechanism against parasitism can be found in an A. grossa population in Tandayapa. Tandayapa is approximately 20 km east and 900 m higher in elevation than Mindo. Studies of this population of A. grossa found no egg mass parasitism. Only 38% of the egg-guarding A. grossa females in Tandayapa were aggregated (L. Camacho & C. Keil, unpublished) in contrast to 78% of the egg-guarding females from Mindo in this study. Moreover, females from Tandayapa were not as aggressive as those in Mindo when defending their egg masses. The individual egg masses were about twice as large as those from Mindo and consequently the proportion of the egg mass covered by the female was reduced (L. Camacho & C. Keil, unpublished). Therefore, it appears that in the absence of egg mass parasitism, A. grossa females in Tandayapa did not have such a strong propensity to aggregate and invest in maternal care strategies such as aggressive behaviour or minimising egg mass size.
Future studies should aim towards studying the genetic relationship among aggregated A. grossa females. When A. grossa immatures reach adulthood and gain flight capabilities, they begin to disperse in a similar fashion to U. crassicornis (Wood & Dowell, 1985), although sisters may stay on the host plant and rear their own offspring there. Hence, cooperation between egg-guarding sibling females would be strongly justified by genetic relatedness. Because of the tendency to aggregate for oviposition, it is possible that adult females have limited dispersal and hence may show some degree of relatedness. If that is the case, A. grossa's subsociality would include a kin selection component added to the selfish herd effect and simple cooperative behaviour proposed in this study. The existence of kin selection within the species' subsocial behaviour would put it, or even the whole tribe, a step closer towards the evolution of eusociality (Anderson, 1984).