Invasive events appear to be highly idiosyncratic . Our understanding of what allows some introduced species to become invasive, while others do not remain poor, likewise there is little consensus on what makes some ecosystems more vulnerable to invasion than others . For example some studies indicate a link between species richness in the recipient community and the likelihood that it will be invaded Stachowicz et al. [55, 56], while others do not [57, 58]. Nevertheless, it is clear that the frequency of inoculation, or exposure to introduced species is important and this is part of the reason (perhaps the main reason) that coastal marine ecosystems suffer higher levels of biological invasions than almost any other ecosystem . Yet even within these systems it is by no means clear why only some inoculations with alien species lead to successful invasions. In the past the focus has been on the differing physiological abilities of introduced species and their native competitors (e.g. ) and little attention has been paid to the effects of behaviour, partially because the behavioural repertoires of these species, especially the benthic taxa, are extremely limited. Our results indicate that even very simple behaviours can have profound effects, contributing to effective habitat partitioning, by altering the physical environment as actually experienced by competing species.
The rocky intertidal zone is among the most physically harsh environments on earth  and physiological adaptation to environmental stress (e.g. wave action, temperature and desiccation) plays a major role in determining potential patterns of vertical distribution (zonation). Nevertheless, species interactions (e.g. predation, competition and parasitism) modify this crude template profoundly [47, 59, 60] and behaviour in turn can affect both stress tolerance and species interactions [16, 17, 61]. We show that two competing intertidal mussels have clearly different gaping behaviour that strongly influences their tolerance of environmental stress, consequently playing a determining role in their distribution and the dynamics of co-existence between an invasive and an indigenous species. Non-gaping behaviour allows the invasive species to colonise the higher mussel zone that is incompatible with the gaping behavior of the indigenous mussel because of the high risk of desiccation. The greater tolerance of Mytilus galloprovincialis to desiccation comes at the price of a higher production of Hsps. When aerially exposed, Perna perna has significantly higher water loss than the non-gaping species, probably due to both evaporation and incidental expulsion of water during valve closure (, KRN and GIZ pers. obs.). This provides an explanation for the fact that it exhibits higher mortality rates than M. galloprovincialis when the two are transplanted to the higher shore [44, 47]. These results are also in agreement with higher mortality rates of P. perna during air exposure at 17°C. In the field, the two mussel species are subjected to semidiurnal tides and they are usually exposed for less than six hours. The laboratory experiments lasted much longer than normal emersion, but produced mortality rates that presumably mirror sub-lethal fitness effects under field conditions where evaporative water loss is severe. Moreover, no intra-specific differences between mussels from different intertidal heights were observed in any of the parameters measured; this suggests that gaping behaviour alters water loss and desiccation but is related to species peculiarities rather than being a direct individual response to intertidal environmental conditions.
Previous studies have shown that, when kept in anoxic seawater, M. galloprovincialis and P. perna keep the valves open, but the latter is more sensitive to this condition and suffers higher mortality rates . In addition, both species can regulate oxygen uptake down to concentrations of approximately 2-2.5 and 3.4 ppm respectively [63, 64]. This is in accord with our results, indicating that during air exposure P. perna keeps the valves opened because of the need to rely on aerobic metabolism, while the invasive species has a higher tolerance of anoxic conditions and is able to maintain valve closure. Although it reduces desiccation rates, valve closure condemns mussels to a less efficient metabolism when catabolic, acidic end products have toxic effects and can accumulate to lethal levels [21, 34, 65, 66]. Respiration rates of organisms usually increase with temperature and several studies have shown that mussels' oxygen consumption increases exponentially with temperature (e. g. [31, 35]). Our results show that at 37°C mussels gape more than at lower temperatures suggesting that this behavior is linked to greater oxygen consumption, and supporting the important role of gaping in aerobic respiration . In addition the greater ventilation rate seems to explain the higher water loss (through evaporation and/or active expulsion) observed at 37°C.
High levels of Hsps expression in intertidal organisms have been related to several stressors, including heat and anoxic stress [30, 68]. The primary antibody used in this study is specific for both the inducible isoform, which is triggered by heat shock, and the constitutive isoform of Hsp70, which is affected by multiple physiological parameters. Previous studies found that there were higher levels of constitutive Hsp70 among high intertidal mussels in comparison to mussels from the low intertidal; in addition, constitutive Hsp70 levels in high intertidal bivalves were higher in summer than in winter [69, 70]. This study shows that the expression of Hsps is significantly higher in M. galloprovincialis than in P. perna. Unexpectedly, we did not detect any intraspecific difference when comparing individuals from different heights of the mussel zone. This suggests that different expression levels of Hsps between M. galloprovincialis and P. perna are species-specific rather than being related to vertical stress gradient in the intertidal (but ). Several studies on different species have shown that gaping does not play a role in body temperature [19, 35–37]; its most probable function is to allow aerobic respiration by maintaining an O2 gradient across the gills and the mantle wall [16, 35, 38]. This suggests that the higher stress experienced by M. galloprovincialis, as indicated by greater levels of Hsps expression, is related to anoxic stress rather than heat stress. Single stresses are rarely encountered in the natural environment. More typically, organisms encounter multiple stresses that interact with a variety of stress-responsive systems. In the case of sessile bivalves, a critical stressor that has been overlooked is simultaneous exposure to anoxia or hypoxia during periods of thermal stress. Since most periods of elevated ambient temperatures occur during emersion it will be important to address the question of whether anoxic and hypoxic conditions interact with heat-shock responses in intertidal bivalves.
Hawkins and Bayne  have estimated the cost of protein synthesis to constitute 20 - 25% of the energy budget of the mussel Mytilus edulis. This cost represents an additional energy burden because stress proteins do not directly contribute to growth or reproduction, and because under stress conditions Hsps may be synthesized preferentially, so that other proteins critical for the normal functioning of the organism are either synthesized at reduced rates or not synthesized at all. Furthermore, the function of stress proteins may require considerable ATP turnover; refolding of a protein may consume in excess of 100 ATP molecules [29, 72, 73]. Previous studies on intertidal mussels have shown that a trade-off exists between metabolically demanding processes such as attachment strength and gonad maturation . In P. perna and M. galloprovincialis; peaks in attachment strength coincide with periods of relatively low gamete production for both species, suggesting that they cannot afford to invest simultaneously in both processes . Energetic constraints can change spatially according to gradients of multiple physical factors and challenge co-existing species differently . P. perna attachment strength is higher than that of M. galloprovincialis, while the latter has a greater reproductive output [77, 78]. On the more wave exposed open coast, both species have to increase their attachment strength but, for the invader this comes at the cost of reduced reproductive output . The costs of Hsps expression can also affect fertility/fecundity, energy budgeting, and survival through the alteration of cell functioning and high energy consumption . The higher Hsps production of the invasive M. galloprovincialis could represent a competitive disadvantage, limiting its ability to survive on wave exposed shores where greater attachment strength is required.
The paradigm that assumes that lower and upper vertical distribution limits in the intertidal are solely set by biological and physical factors, respectively, has been challenged in the last few decades. Recent studies indicate that vertical zonation, and community dynamics in general, are largely driven by the interplay between environmental stress and species interactions [75–77]. Our two studied mussels show different tolerances to wave and sand stress, two of the main environmental factors affecting these intertidal communities. P. perna is more resistant to hydrodynamic stress than M. galloprovincialis. This explains the lower vertical limit of the invasive species and also explains the high mortality rates of M. galloprovincialis on wave exposed shores [17, 62]. On the other hand, the invasive species is less vulnerable to sand action . Sand tolerance does not have a relevant role in the vertical zonation of these species, but accounts for high mortality rates of P. perna during periods of high sand accumulation in mussel beds. Moreover, competition is at least in part influenced by physiological tolerance of physical factors. M. galloprovincialis survives well in the high zone, while transplanted P. perna dies . Our results clearly underline the crucial role of behaviour. It limits desiccation of the invasive species and promotes the invader's success in the high mussel zone. On the other hand, P. perna is restricted by its gaping behaviour to the lower shore where it is also able to exclude the other species . This agrees with the distribution of the invasive species on the west coast of South Africa, where P. perna is naturally absent and M. galloprovincialis extends its domain to the lower mussel zone.
Our results indicate that the outcomes of invasive/indigenous interactions are not necessarily solely affected by the environmental conditions prevailing in the invaded region, but rather by the conditions physiologically experienced by each species. Behaviour can dramatically moderate an organism's experience of environmental conditions and so dictate physiological reactions, and consequently the limits of tolerance limits to the stresses imposed by those conditions. As both indigenous species and invaders must respond to environmental variations, the difference in their responses can determine the success of the invader and how it interacts with native species [5, 6]. Here we show that even the simple behaviour of sedentary benthic animals can explain the mechanisms behind observed patterns of distribution and the ability of an introduced species to become invasive.