خشونت درون گروهی و بین گروهی بین سطوح آندروژن در گریگوریوهای تاریک را دگرگون نمی کند، با این حال خشونت مردانه توسط یک مسدودکننده آندروژن کاهش می یابد

Discussions about social behavior are generally limited to fitness effects of interactions occurring between conspecifics. However, many fitness relevant interactions take place between individuals belonging to different species. Our detailed knowledge about the role of hormones in intraspecific interactions provides a starting point to investigate how far interspecific interactions are governed by the same physiological mechanisms. Here, we carried out standardized resident–intruder (sRI) tests in the laboratory to investigate the relationship between androgens and both intra- and interspecific aggression in a year-round territorial coral reef fish, the dusky gregory, Stegastes nigricans. This damselfish species fiercely defend cultivated algal crops, used as a food source, against a broad array of species, mainly food competitors, and thus represent an ideal model system for comparisons of intra-and interspecific territorial aggression. In a first experiment, resident S. nigricans showed elevated territorial aggression against intra- and interspecific intruders, yet neither elicited a significant increase in androgen levels. However, in a second experiment where we treated residents with flutamide, an androgen receptor blocker, males but not females showed decreased aggression, both towards intra- and interspecific intruders. Thus androgens appear to affect aggression in a broader territorial context where species identity of the intruder appears to play no role. This supports the idea that the same hormonal mechanism may be relevant in intra- and interspecific interactions. We further propose that in such a case, where physiological mechanisms of behavioral responses are found to be context dependent, interspecific territorial aggression should be considered a social behavior.

A behavior is considered social if it has fitness effects on both its actor and receiver (West et al., 2007). Scientists discussing social behavior typically refer to intraspecific interactions (Blumstein et al., 2010), and as an example, a recent textbook on the topic did not mention a single case of interspecific interactions (Székely et al., 2010). This omission may look surprising since organisms are sometimes embedded in a heterogeneous network where interspecific interactions are frequent and exert profound effects on fitness (Bshary, 2001 and Peiman and Robinson, 2007). In addition, intra- and interspecific behaviors can be classified largely along the same line depending on their effect on the direct reproductive fitness of the different partners. At the exception of altruism which is not expected in interspecific interactions, behaviors directed towards con- or heterospecific individuals can be mutually beneficial, selfish or spiteful. From a modeling perspective, the main difference is that interaction partners belong to different gene pools in interspecific interactions, which causes mainly quantitative adjustments (Bergström et al., 2003 and Doebeli and Knowlton, 1998). Thus, the question arises whether a distinction between intra-and interspecific interactions is useful or whether it hinders a better integration of concepts. A classical argument for a distinction between intra- and interspecific behaviors originates from studies of aggression. Intraspecific aggression shown during escalated fighting is generally restricted to the reproductive season, sex specific and facilitated by androgen hormones such as testosterone (T) (Borg, 1994, Liley and Stacey, 1983 and Wingfield et al., 2006). In contrast, aggression directed against heterospecifics, like killing of prey, may be shown by both sexes and unrelated to T (Bernard, 1976, Gammie et al., 2003 and Giammanco et al., 2005). This argument takes a general behavior (aggression) as the starting point for the distinction. At the proximal level though, it was postulated early on that it would be more appropriate to explore the causation of intra- and interspecific behaviors in a specific context, rather than as a function of against whom it is directed (Huntingford, 1976). In order to do so we studied the relationship between hormonal correlates with intra- and interspecific aggression in the context of territoriality. In this context, the immediate consequences of an aggressive act by the territory holder on intra- or interspecific intruders are comparable, and according to Huntingford (1976) in such a situation the aggressive response would be best brought about by a shared physiological factor. We selected androgens as a candidate physiological factor, because castration and replacement studies have demonstrated a central role of these steroid hormones in the causation of aggressive behavior (Adkins-Regan, 2005). For intraspecific territorial aggression a two-way causal relationship has been described, where social challenges stimulate the release of androgens, which in turn facilitate territorial aggression towards intruders (Hau, 2007, Oliveira, 2004, Ros et al., 2002, Ros et al., 2004 and Wingfield et al., 1990). This relationship has been mostly investigated during the reproductive period, and has generally been shown to vary with the degree of social instability associated with parenting style, and mating system (Goymann, 2009, Hirschenhauser and Oliveira, 2006, Wingfield et al., 1990 and Wingfield et al., 2006). Indeed, it has been shown that socially induced increases in androgen levels, in particular in testosterone (T), can be induced experimentally by simulating a territorial intrusion using decoys (Desjardins et al., 2006, Gleason et al., 2009, Goymann, 2009, Hay and Pankhurst, 2005, Hirschenhauser et al., 2004 and Ramenofsky, 1984). In coral reefs, niche overlaps between species may lead to both intra- and interspecific competition for shelter or food (Holbrook and Schmitt, 2002, Muñoz and Motta, 2000 and Myrberg and Thresher, 1974). This is particularly evident in some species of damselfish that are called ‘gardeners’ as they grow and harvest algae as a food source (Ceccarelli, 2007 and Karino and Nakazono, 1993), and defend these in a permanent territory against a broad array of potential competitors throughout the year (Di Paola et al., 2012, Ebersole, 1977, Hata and Kato, 2002, Hata and Kato, 2003, Hata and Kato, 2006 and Hata et al., 2010). Our study species, the dusky gregory, Stegastes nigricans, forms colonies where each individual solitary defends its algal crop ( Karino and Kuwamura, 1997). Here we address two main issues. We first ask whether the relationship between androgens and intraspecific territorial aggression, as observed in reproductively active territorial fish ( Desjardins et al., 2006, Hirschenhauser et al., 2004 and Ros et al., 2004), hold outside of the reproductive period. We then question whether this relationship found for intraspecific territoriality would be similar when territorial aggression is directed against heterospecific intruders. In our first experiment, we measured the effects of both intra- and interspecific territorial aggression on plasma levels of androgens, focusing on T and 11-ketotestosterone (KT). These two androgens are behaviorally the most relevant in teleosts (Borg, 1994 and Kime, 1993). In a second experiment, we investigated any causal effect of androgens on territorial aggression by treating the residents with slow-release implants of flutamide, an androgen receptor blocker (Sebire et al., 2008), or castor oil as a control. We compared territorial aggression towards conspecific and heterospecific intruders to test whether androgens might act as a common causal factor for both types of responses. As a heterospecific intruder we selected another year round territorial damselfish, the herbivore Plectroglyphidodon lacrymatus. Under natural conditions, P. lacrymatus are often found on the edges of S. nigricans colonies and competition might occur over vacant territories which also represent a food resource ( Hata and Kato, 2006). Occasional aggressive interactions have indeed been observed between free living individuals of the two species (Vullioud, personal observation). Because the endocrine system varies with sex, we tested for possible sex-specific effects of androgens on territorial behaviors by including both males and females in our study.

The multivariate test (Table 1) showed a significant effect of the intruder (sRI) on behavioral response (effect sRI vs sRI-c: Pillai's Trace [1,42] = 0.79, p < 0.001). No significant effect of the intruder species, or of the gender of the resident was found (main and interaction effects: Pillai's Trace [1,42] < 0.095, p > 0.13). Post-hoc univariate tests showed a significant increase due to the intruder in agonistic displays and in aggression (Fig. 1; agonistic display, F[1,45] = 105.4, p < 0.001; aggression, F (1,45) = 85.2 p < 0.001) and a decrease in hiding duration (Fig. 1; F[1,45] = 29.85, p < 0.001). Levels of aggression, agonistic display and hiding duration during sRI tests were significantly related to each other (Spearman correlation, |rho| > 0.47, n = 46, p < 0.001).