School of Biodiversity, One Health & Veterinary Medicine

One of the main challenges facing population ecology is to explain the enormous spatial and temporal variation observed in patterns of abundance. Our understanding of these patterns has been illuminated by detailed studies of species with unstable dynamics (Stenseth et al. 1996; Hudson et al. 1998; Turchin 1999; Grenfell et al. 2001; Berryman 2002; Korpimaki et al. 2002; Klemola et al. 2003) . However, it is clear that such studies often fail to capture the variance in the dynamical patterns seen between populations. For example, the dynamics of many species with cyclic populations show variation in period, amplitude and shape across their range (Turchin 2003) , suggesting that this variation maybe the result of the tension between population processes and environmental conditions.

Populations are governed by a variety of biotic (eg predation & parasitism) and abiotic (eg climate) factors, and interactions between these factors are common in nature. Recent modelling and time series analyses have illustrated how biotic factors may interact with the perturbations caused by climate to affect dynamics and regional synchrony (Ranta et al. 1995; Grenfell and Finkenstadt 1998; Stenseth 1999; Cattadori et al. 2000; Bjornstad et al. 2002; Cattadori et al. 2005). Furthermore, the biotic factors may themselves interact (Newton 1998), and we know that such interactions can profoundly influence demography ( Karels et al. 2000; Zanette et al. 2003) and generate a wide range of dynamical outcomes (Packer et al. 2003). However, few field studies have considered how the interactions between biotic factors influence dynamics (Krebs et al. 1995) . In this proposal we seek to quantify the impact of interactions between biotic factors in their climatic context on population dynamics.

To tackle this issue, we will focus on a system dominated by two biotic processes that are present in a broad range of systems: intraspecific competition in the form of territorial behaviour, and parasitism. Territoriality is a widespread phenomenon, whereby individuals establish and defend a territory through increased aggressive behaviour. Similarly, parasites are widespread and t heir impact on reproduction and survival is now widely recognized (Gulland 1995; Albon et al. 2002; Hudson et al. 2002; Stien et al. 2002; Newey and Thirgood 2004) . Moreover, there is also increasing evidence that these two processes interact: parasites can influence aggressive behaviour, and increased testosterone dependent aggression is often associated with reduced parasite resistance (Folstad and Karter 1992; Hillgarth and Wingfield 1997) . In this proposal we seek to quantify the impact of interactions between territorial behaviour and parasitism on individual fitness, and then link individual behaviour to long-term demography through dynamic models to predict the impact on population dynamics and test these predictions using long-term time series.

1.1 From individual behaviour to population dynamics

In systems where multiple biotic factors are operating, individuals have to trade-off the risks and benefits of alternative strategies in order to maximise their fitness. The consequences of the decisions individuals take in these trade-offs may profoundly affect population size and dynamic stability (McNamara 2001). Moreover, given the trade-offs associated with territorial aggression, the fitness consequences of individual decisions are likely to depend on the decisions of others - for instance, the costs of fighting will depend on how aggressive opponents are likely to be. Game theory is the most appropriate tool for analysing such influences (Maynard Smith 1982).

In the basic game theoretic model of aggressive interactions for access to resources, the so-called Hawk-Dove game (Maynard Smith & Price 1973), evolutionarily stable (ES) levels of aggression depend crucially on the value of the resource being competed for and the costs associated with fighting. These values and costs are likely to depend on a wide range of factors including current levels of aggression within the population, parasites and environmental conditions. There is a long tradition of applying game theory to understanding levels of aggression in contests over resources (e.g. Maynard Smith & Price 1973, Grafen 1987, Eshel & Sansone 1995, Johnstone 2001) and progress has been made toward understanding the role of population density in determining evolutionarily stable levels of aggression in populations (Grafen 1987, Houston & McNamara 1991, Mesterton-Gibbons 1992, Eshel & Sansone 1995), and their influence on population dynamics (McNamara 2001). However, the behaviour of individuals in populations subject to interacting regulatory processes remains largely unexplored. Here we propose to use game theory to explore the strategic consequences of the interaction between aggressive territorial behaviour and increased susceptibility to infection by parasites. We will thus obtain specific parameters from this game theoretic approach and from new experiments to incorporate into a population model and compare the emergent dynamic properties against existing long-term time series.

1.2 The Study System

Red grouse Lagopus lagopus scoticus provide us with a unique study system with which to empirically tackle the role of biotic interactions and model the consequences for population dynamics. Grouse live on the heather-dominated moorlands of upland Britain and many populations are managed on private estates and harvested in autumn. Long-term harvest records are available for many of these estates, and their analysis shows that grouse populations exhibit unstable dynamics with geographical variation in cycle period ranging from 2 to 15 years (Haydon et al. 2002). This study system is relatively simple, and we have identified, manipulated and modelled the role of the dominant, destabilising biotic factors: territorial behaviour and parasites (Mougeot et al. 2003a,b; Matthiopoulos et al. 2003, 2005; Hudson et al. 1998; Redpath et al. In press a). Other processes, such as predation and food quality do not generate instability but do influence population growth rate (Moss and Watson 2001; Hudson et al. 2002). Traditionally, two, single-process hypotheses have been proposed to explain cyclic, grouse dynamics, one based on territorial behaviour and one on parasites. These have been treated as mutually exclusive.

1.3 Territorial behaviour and red grouse cycles

Red grouse are monogamous gamebirds that compete for territories in the autumn and defend them through to the following breeding season. They have conspicuous supra-orbital combs that are important signals in inter- and intra-sexual encounters (Moss et al. 1979; Mougeot et al. 2004; Mougeot & Redpath 2004). The size of these combs positively correlates with plasma testosterone concentration (Mougeot et al. 2005a) and relates to aggressiveness: males with bigger combs are more aggressive, more likely to obtain a large territory, hold larger territories and produce more young (Moss et al. 1994; MacColl et al. 2000; Redpath et al. In press b). The territorial behaviour hypothesis states that delayed density dependent changes in male aggressiveness leads to population cycles through its influence on territory size and recruitment rate of young males into the territorial population (Watson 1967 ; Moss et al. 1994; Moss et al. 1996 ; Matthiopoulos et al. 2003) . There is now strong evidence that these lagged-changes in aggressiveness are the result of changing kin structure within populations (Mountford et al. 1990;Matthiopoulos etal. 1998;Mougeot et al. 2005d) .

1.4 Parasites and red grouse cycles

The caecal threadworm Trichostrongylus tenuis is the main parasite of red grouse, and has important negative effects on the energetics, reproduction and survival of grouse (Hudson 1986b; Shaw & Moss 1990; Hudson et al. 1992; Hudson et al. 2002 ). This parasite has a direct life-cycle and no alternative hosts within the same habitat (Hudson 1986b). Prevalence is high, but intensity of infection varies considerably between years and locations as a function of both grouse density and rainfall (Hudson 1986a; Hudson et al. 1992; Moss et al. 1993; Cattadori et al. 2005). The parasite hypothesis states that cycles are caused by T. tenuis induced reductions in female grouse breeding production in tension with other features of the host-parasite system, specifically parasite induced mortality and the degree of parasite aggregation (Potts et al. 1984; Hudson 1986a ; Dobson & Hudson 1992; Hudson et al. 1992).

1.5 Interactions between territorial behaviour and parasites in red grouse

The territorial behaviour and parasite hypotheses were developed from field studies in different geographical regions, with differing climate, where the period and shape of the cycles differ greatly (Hudson 1992; Shaw et al. 2004). Thus it is possible that either the two processes are geographically distinct (Turchin 2003) , or both are operating across the species range to shape the dynamics. In a previous NERC grant ( NER/A/S/1999/00074) we undertook a wide range of population level experiments that confirmed the de-stabilising role of both processes in both regions and thus that the two processes were not geographically distinct (Mougeot et al. 2003a; Redpath et al. in press b ). Moreover, our individual-level experiments (Fig 1) have shown that parasites and aggressiveness interact in two ways (Fox & Hudson 2001; Mougeot et al. 2005b; Seivwright et al. 2005), and that some of these interactions are long-lasting. For instance, increased parasite intensities in male grouse with heightened aggressiveness in autumn only become apparent after one year (Seivwright et al. 2005) . Moreover, because grouse do not develop acquired immunity (Hudson & Dobson 1997) , these increases in parasite intensity may be maintained throughout their life. We are now at a point where we can examine how these interactions affect individual fitness. Furthermore, our knowledge of this system, developed from long-term studies, will enable us to link individual behaviour to long-term demography through dynamic models and predict emergent dynamics across gradients in climate.