- Top of page
In many group-living animals, within-group associations are determined by familiarity, i.e. familiar individuals, independent of genetic relatedness, preferentially associate with each other. The ultimate causes of this behaviour are poorly understood and rigorous documentation of its adaptive significance is scarce. Limited attention theory states that focusing on a given task has interrelated cognitive, behavioural and physiological costs with respect to the attention paid to other tasks. In multiple signal environments attention has thus to be shared among signals. Assuming that familiar neighbours require less attention than unfamiliar ones, associating with familiar individuals should increase the efficiency in other tasks and ultimately increase fitness. We tested this prediction in adult females of the group-living, plant-inhabiting predatory mite Phytoseiulus persimilis. We evaluated the influence of social familiarity on within-group association behaviour, activity, predation and reproduction. In mixed groups (familiar and unfamiliar), familiar predator females preferentially associated with each other. In pure groups (either familiar or unfamiliar), familiar predator females produced more eggs than unfamiliar females at similar predation rates. Higher egg production was correlated with lower activity levels, indicating decreased restlessness. In light of limited attention theory, we argue that the ability to discriminate between familiar and unfamiliar individuals and preferential association with familiar individuals confers a selective advantage because familiar social environments are cognitively and physiologically less taxing than unfamiliar social environments.
Group-living is a widespread phenomenon in both invertebrate and vertebrate animals (Krause and Ruxton 2002). Ultimately, animals live in groups for various reasons such as enhanced food exploitation, vigilance, protection from or defence against predators, breeding, mate choice, homeostasis or energy saving in group movements (Alexander 1974, Krause and Ruxton 2002, Earley and Dugatkin 2010). Proximately, group formation and cohesion may be based on two principal mechanisms: mutual attraction and/or response to external stimuli such as abiotic conditions or food sources (Alexander 1974, Krause and Ruxton 2002). In most cases, the proximate causes of group formation and cohesion are difficult to pinpoint because grouping is commonly guided by both conspecific and external stimuli (Krause and Ruxton 2002). While the mechanisms of group formation and cohesion, and the benefits and costs of group-living have been intensely investigated (Alexander 1974, Krause and Ruxton 2002, Earley and Dugatkin 2010), a more rarely addressed aspect of group-living is within-group association behaviour (Croft et al. 2005, Jordan et al. 2010) and its bearings for individual group members.
Within-group associations are commonly non-random and may be determined by group member characteristics such as life stage, sex, size, dominance rank or social famili arity (Conradt and Roper 2000, Krause and Ruxton 2002, Guttridge et al. 2009, Earley and Dugatkin 2010, Kaspersson et al. 2010). Here we focused on social familiarity, which requires the ability to discriminate familiar and unfamiliar conspecific individuals based on prior association (Waldman 1988, Grafen 1990, Mateo 2004). Preferential association of familiar individuals has been observed in many animal taxa such as mammals (Palphramand and White 2007), birds (Komdeur et al. 2004), fish (Höjesjö et al. 1998, Griffiths 2003), amphibians (Blaustein and O’Hara 1986) and arthropods (Gamboa 2004, Schausberger 2004, 2007, Strodl and Schausberger 2012a, b).
Familiarity is a common mechanism used to discriminate kin and non-kin (Blaustein and O’Hara 1986, Waldman 1988, Mateo 2004, Schausberger 2007). The adaptive significance of familiarity, if used as proxy of genetic relatedness, has been revealed for various group- living animals (Komdeur et al. 2004, Gamboa 2004, Schausberger 2004, 2007). If there is an obvious gain in inclusive fitness, preferential treatment of familiar individuals is commonly considered kin-selected behaviour (Hamilton 1964) and kin-selected behaviours targeted towards non-kin would be considered recognition errors. However, several studies showed that differential treatment of familiar and unfamiliar individuals may occur inde pendently of the degree of genetic relatedness and may be beneficial without indirect fitness gains, implying the existence of alternative or additional forces selecting for the ability to discriminate familiar and unfamiliar individuals. For example, familiarity of group members, independent of genetic relatedness, may increase foraging efficiency (Griffiths et al. 2004, Strodl and Schausberger 2012b) or may reduce aggressive and competitive behaviours (Ward and Hart 2003, Palphramand and White 2007, Höjesjö et al. 1998), emotional stress (Takeda et al. 2003) or predation risk (Chivers et al. 1995, Ward and Hart 2003, Strodl and Schausberger 2012a). While all these studies show the behavioural consequences of familiarity and commonly assume familiarity to be adaptive, only few studies linked the behavioural observations with the interrelated cognitive processes (Griffiths et al. 2004, Strodl and Schausberger 2012a), or provided a conclusive ultimate explanation for why the ability to recognise familiar individuals evolved in and should be beneficial for group- living animals. Moreover, only one study (Strodl and Schausberger 2012a) experimentally documented the adaptive significance, i.e. the effects of social familiarity on the prime fitness traits, survival and/or reproduction. Strodl and Schausberger (2012a) observed enhanced survival of socially familiar mites under the risk of predation as compared to unfamiliar mites.
A highly appealing, but rarely experimentally tested, ultimate explanation of the benefits of associating with familiar individuals and the interrelated cognitive processes is the implication of limited attention (Dukas 2002). Limited attention theory postulates that focusing on a given task has cognitive and associated physiological costs with respect to the attention paid to other tasks and may thus affect every major life activity such as foraging, anti-predation, reproduction, mating and social interactions. In natural environments, animals are constantly confronted with multiple signals they should or should not respond to. Cognitive processes, including shared attention to simultaneously present stimuli (or tasks such as foraging and neighbor inspection – both need at the cognitive level perception and processing of cues of prey/food and conspecific individuals), and the associated behaviours are inevitably linked (Bernays 2001, Dukas 2002, 2004). Attention per se can only be indirectly deduced from behavioural and/or neurobiological and physiological measurements (Dukas and Kamil 2000, Bernays 2001, Dukas 2002, 2004). In an ideal case one would integrate these diverse measurements (Dukas 2002, 2004) but this is extremely difficult to accomplish for most study animals. Thus, all previous studies on limited attention, including the pioneering study by Dukas and Kamil (2000) or more recently Griffiths et al. (2004), Purser and Radford (2011) and our own studies (Strodl and Schausberger 2012a, b), relied on behavioural observations alone to conclude on the interrelated cognitive processes. Previous investigations strongly suggested that the ability to behaviourally respond to signals is limited by various cognitive constraints, which are defined as anything that prevents, delays or increases the costs of focusing on a given task (Bernays 2001, Dukas 2002, 2004). Such constraints may be a low learning rate, insufficient perception, or an imperfect long-term and/or working memory. The classical example comes from blue jays simultaneously performing two different tasks (foraging and predator vigilance), limiting their ability to efficiently respond to a peripherally presented predator model (Dukas and Kamil 2000).
The implication of limited attention theory for group-living animals is that, if familiar group members require less attention than unfamiliar ones, assorting with familiar individuals should lead to increased efficiency in other tasks (Dukas 2002, Griffiths et al. 2004, Strodl and Schausberger 2012a, b, Zach et al. 2012). Thus, in light of limited attention theory, social familiarity, i.e. familiarity among conspecific individuals (Alexander 1974), may confer a selective advantage if it allows to switch attention from costly group member assessments or aggressive interactions to other major life activities such as foraging, predator avoidance, parental care or mating (Dukas 2002) and thereby enhances fitness. To date, the only experimental support for limited attention theory in the context of group-living comes from salmonid fish (Griffiths et al. 2004) and predatory mites (Strodl and Schausberger 2012a, b). Griffiths et al. (2004) demonstrated that individuals with a familiar conspecific neighbour responded more quickly to simulated predator attacks and had higher feeding rates than indivi duals with an unfamiliar conspecific neighbor. Likewise, juvenile predatory mites, Phytoseiulus persimilis, held in familiar groups responded more quickly to intraguild predator attacks (Strodl and Schausberger 2012a) and foraged more optimally, i.e. needed less prey at similar developmental speed and body size at maturity, than unfamiliar mites (Strodl and Schausberger 2012b). Nevertheless, except for the study by Strodl and Schausberger (2012a), which shows that socially familiar juvenile mites react more quickly to approaching predators and consequently survive longer under predation risk, experimental studies documenting the adaptive significance of social familiarity are lacking.
Here, we examined the adaptive significance of social familiarity in adult females of the group-living, plant- inhabiting predatory mite P. persimilis. We hypothesized that social familiarity reduces the cognitive (due to limited attention) and associated behavioural and physiological (due to decreased stress) costs of group-living. Within-groups, familiar mites should be more likely to associate with each other than unfamiliar mites and social familiarity should enhance the mites’ efficiency in foraging and/or reproduction. In the first experiment, we assessed the influence of social familiarity on within-group association behaviour of adult gravid P. persimilis females held in mixed groups of familiar and unfamiliar individuals by determining the familiarity status of each individual’s neighbours and the inter-individual distances of familiar and unfamiliar individuals. To assure that the establishment of social familiarity and its possible consequences occur independently of the degree of genetic relatedness (Schausberger 2007), we used females with varying degrees of genetic relatedness. In the second experiment, we assessed the effects of familiarity on general activity, predation, oviposition, inter-individual distances and offspring sex ratio of P. persimilis females living in groups consisting of either familiar or unfamiliar individuals. To determine if group size during the sensitive familiarization period affects the familiarization process and its possible consequences later in life, we used females familiarized in small and large groups.
- Top of page
Social familiarity had a decisive impact on within-group association behaviour and reproduction of the predatory mite Phytoseiulus persimilis. Females living in mixed groups of familiar and unfamiliar females preferentially associated with familiar females. This behaviour was more apparent in groups with homogeneous genetic background than in groups with heterogeneous genetic background. Females living in groups of only familiar females had higher oviposition rates, similar predation rates, lower activity levels and shorter inter-individual distances than females living in groups of only unfamiliar females. Limited attention theory postulates that in multiple signal environments attention has to be shared among signals, with familiar signals drawing off less attention than unfamiliar signals (Dukas 2002). Thus, our findings are consistent with the idea that familiar pre datory mite females preferentially assort with each other because a familiar social neighbourhood draws off less attention and allows direction of more attention and energy to fitness enhancing activities such as oviposition.
Our study documents and links the adaptive significance of social familiarity to relaxation of limited attention (Dukas 2002), reducing the physiological costs of infor mation gathering and processing in a group-living animal. In general, physiological tradeoffs may occur for various reasons and behavioural observations are only indirect indicators of attention. Nevertheless, in our experiments social familiarity, a cognitive trait, was the only manipulated factor, everything else was kept the same. Moreover, the assumption of the tight linkage between attention-related cognitive processes, fitness and behavioural traits is backed up by the recent finding that social familiarity enhances vigilance, measured in reaction time, and survival of P. persimilis under the risk of predation (Strodl and Schausberger 2012a). Reaction time is a most commonly used behavioural indicator of attention (Rowe 1999, Griffiths et al. 2004). Various studies suggested beneficial effects of social familiarity in group-living species con cerning foraging, anti-predator behaviour and agonistic inter actions (Chivers et al. 1995, Höjesjö et al. 1998, Utne-Palm and Hart 2000, Strodl and Schausberger 2012a, b). However, no previous study measured the effects of social familiarity on the reproductive success of group-living animals. For example, minnows in familiar shoals had higher food intake rates than minnows in unfamiliar shoals (Chivers et al. 1995). However, higher feeding rates do not necessarily convert into higher reproduction (Illius et al. 2002). In experiment 2, females living in familiar groups had higher oviposition rates than females living in unfamiliar groups. Classical optimal foraging theories predict that selection favours decision mechanisms, which optimize energy gain by maximizing food intake rate and/or minimizing handling and/or searching times (Giraldeau 2008). Consequently, in terms of limited attention, familiarity could have increased foraging success of familiar individuals (Griffiths et al. 2004). However, the 24 h predation rates of P. persimilis did not differ and possible differences in handling and searching times were negligible due to the high prey density. There was also no indication of partial vs complete consumption of prey eggs between-groups, which could have resulted in differences in energy extracted per prey item. Moreover, such differences could have been balanced by shorter handling times, reducing energy spent per item, and could have been compensated for by increased numbers of consumed eggs. We therefore argue that the observed lower general activity level of familiar females is the proximate explanation for their higher reproduction. Lower general activity is an indicator of decreased restlessness, which, at the physiological level, reflects less stress (Lincoln et al. 1998 for a definition of stress, Takeda et al. 2003), and, at the cognitive level, a reduced need to explore the immediate social surrounding, leading to lower energy expenditure for neighbour inspection and assessment (Lima 1998, Dukas 2002) and leaving more energy for egg production. In general, differences in energy expenditure may also be due to other constraints than behavioural performance such as oxygen availability, temperature, diurnal activity, life stage and phase, etc. However, these other constraints can be dismissed as potential explanations in our study because, except for social familiarity during the experiment, everything else, such as the conditions during pre-experimental rearing or life stage and phase, was the same for individuals of familiar and unfamiliar groups. An alternative behavioural inter pretation is that lower activity indicates a lower dispersal propensity of familiar individuals, leading to higher energy expenditure and reduced egg production in unfamiliar individuals. However, this explanation can be dismissed because mites held in familiar groups have a similar or greater propensity to disperse than mites held in unfamiliar groups (Zach et al. 2012). Thus, the activity levels measured in experiment 2 are only indicative of within group behaviour but not dispersal or between group movements.
Depending on the ecological context and the life-stages involved, P. persimilis is able to use at least three different perceptual mechanisms, enabling them to discriminate familiar from unfamiliar conspecifics or kin from non- kin: recognition via prior association, phenotype matching and self-referent phenotype matching (Schausberger 2004, 2005, 2007, Strodl and Schausberger 2012a, b). In the current experiments, P. persimilis seems to have used individual recognition of genetically determined tactile or short distance volatile, recipient-borne chemosensory cues learned through prior association. The response of females living in groups with homogeneous and hetero geneous genetic background did not fundamentally differ, implying that close genetic relatedness per se was not a precondition to familiarize and later discriminate familiar and unfamiliar individuals. Some fishes may even discriminate between familiar and unfamiliar heterospecific individuals (Ward et al. 2003, Valero et al. 2009). However, statistical analyses revealed sophisticated subtle differences in inter-individual distances (indicated by the three-way interaction). Familiar females with homogeneous genetic background were closer together than unfamiliar females at the beginning of the experiment (until 1.5 h) but not during the subsequent 4 h. This pattern might indicate that familiar females were more strongly attracted to each other in the initial phase of the experiment when they were heavily stressed and disordered due to having been transferred into a novel environment. After acclimatization, the predators resumed foraging and searching for optimal oviposition sites, activities expected to increase the inter-individual distances over time because food availability decreases. Nonetheless, the shortest distances and highest likelihood of familiar females being 1st neighbours occurred at the last observation point (after 24 h) when prey density was low but predator egg density high. Phytoseiulus persimilis females are well able to adjust patch-leaving to own and progeny prey needs (Vanas et al. 2006), to manipulate hatching asynchrony to reduce the risk of sibling cannibalism (Schausberger and Hoffmann 2008), and adjust egg placement to ensure that own offspring imprint on kin (Schausberger 2005). Therefore, it is possible that closer association after 24 h was additionally determined by the presence of eggs, i.e. familiar females depositing their eggs closer together, thereby amplifying the effect of familiarity on inter-individual distances. We did not observe this phenomenon in groups of females having a heterogeneous genetic background. In the heterogeneous group, familiarity did not have any effect on inter-individual distances whereas in the homogeneous group it had a significant effect at three observation points. These comparisons indicate that learning and responding to individual cues is perceptually more challenging in a genetically heterogeneous than homogeneous group and may consequently compromise precision in individual recognition. Furthermore, in experiment 2, we observed a more pronounced effect of familiarity on inter-individual distances than in experiment 1. Recognition was probably more precise in experiment 2 because individuals were perceptually less challenged due to lower within-group label variability.
We did not show that females discriminate between different individuals having exactly the same familiarity status, which is experimentally extremely difficult if not impossible to achieve (Thom and Hurst 2004), but we argue that individual recognition, which is known from various animals (Halpin 1980, Bonadonna and Nevitt 2004, Schausberger 2007, Sheehan and Tibbetts 2008), is nonetheless the most likely and plausible recognition mechanism. In experiment 2, group size during familiarization did not affect discrimination ability: females reared in groups of 17 were equally able to discriminate between familiar and unfamiliar individuals as were females reared in groups of 6. This outcome suggests that each female learned and memorized up to 16 individual labels. An alternative explanation to individual recognition is marking of group members leading to a shared phenotypic cue of familiar individuals. During familiarization each individual could have deposited a unique chemical marker on each encountered individual, allowing to later re-recognize this marked individual, or the markers of all group members intermingled and created a unique group-specific mixture that was learned and later re-recognized by each group member. Con- and/or hetero-specific marking is known from various arthropods (Breed et al. 1992, Ivy et al. 2005, Yew et al. 2009) but is unlikely in our experiments for the following reasons: 1) individuals touch each other with the 1st pair of legs for recognition but no glands are known on these legs that could possibly produce marking pheromones (Jagers op Akkerhuis et al. 1985); 2) this mechanism would require repeated marking because the mites moulted three times during the familiarization phase; 3) if every conspecific individual encountered is marked, the effect of familiarity should have vanished over time in the course of the experiment but the opposite was the case – the effect of familiarity was rather strengthened than weakened over time; 4) in experiment 2, every individual would have received up to 16 different markers likely resulting in complex mixtures and possibly masking each other, rendering re-recognition of unique markers extremely difficult.
In summary, our study suggests that social familiarity may reduce the costs of cognitively and physiologically challenging group-living. The observed adaptive significance of social familiarity linked to relaxed limited attention suggests that limited attention may be an important ubiquitous driver of the evolution of the ability to discriminate familiar and unfamiliar individuals in group-living animals. We expect further implications of social familiarity to major life processes of both adult and immature life-stages, not only in reproduction but also in anti-predator behaviours against heterospecific predators (Griffiths et al. 2004, Strodl and Schausberger 2012a), agonistic conspecific interactions (Utne-Palm and Hart 2000, Schausberger 2004, 2007), or juvenile growth and survival (Gerlach et al. 2007, Strodl and Schausberger 2012b). The idea that the effects of social familiarity should cascade to the population level and translate into higher population productivity due to optimized patch exploitation (Vanas et al. 2006, Zach et al. 2012) opens a promising avenue of future research.