1 | Introduction
Parasites account for more than one-third of species on Earth and a great deal of biomass in ecosystems (Lafferty et al., 2006; Dobson et al ., 2008; Kuris et al., 2008); hence, host–parasite relationships are one of the most common biotic associations in nature (Hudson et al., 2006; Dobson et al ., 2008; Kuris et al ., 2008). Parasites damage host health via directly exploiting resources from hosts or indirectly causing physiological burdens (Poulin, 2011; Sheldon & Verhulst, 1996) and can be major drivers of host evolutionary changes (Paterson et al.,2010) and host population dynamics (Hudson et al ., 1998; Poulin, 2011). Parasite infections also occur as a consequence of poor host health (Lochmiller, 1996; Pederson & Greives, 2008; Beldomenicoet al ., 2008). Given that gaining or maintaining immunity is nutritionally costly for hosts (Lochmiller, 1996; Sheldon & Verhulst, 1996), host individuals without enough available food resources can be easily predisposed to higher parasite loads (Forbes et al ., 2016). These opportunistic infections in epidemiology may cause parasite outbreaks and finally crush wild populations (Lochmiller, 1996). Therefore, parasite infections and host health synergistically affect wild host and parasite dynamics.
Elucidating the synergy of parasite infections and host body conditions in the wild can advance our understanding of wild host population dynamics. However, most field studies have only examined cross-sectional correlations and have discussed one-sided causalities (Beldomenicoet al ., 2008). Field studies suggest a negative correlation between host body condition, and infection parameters generally indicate that parasites are causes of poor host condition (Harper et al ., 1999; Vicente et al., 2004; Sala-Bozano et al., 2012; Hasegawa et al., 2022a), although some showed poor condition situations such as food limitations may increase parasite prevalence and intensity because hosts compromise their immune functions under such situations (Forbes et al., 2016). Most importantly, when parasite infections are both the cause and consequence of a poor host condition, we can also expect positive feedback: an infected host with a poor body condition due to the parasite infection will be more susceptible to further infection (Beldomenico et al., 2008; Beldomenico et al., 2009 a, b; Beldomenico & Begon, 2010). Positive feedback may create heavily infected hosts, which could be “super spreaders” among the populations (Beldomenico & Begon, 2010). Further, positive feedback can decrease host survival and eventually undermine the host population (Beldomenico & Begon, 2010).
Both causalities and positive feedback are likely to occur but have rarely been demonstrated in natural populations. This is probably because tracking small and cryptic parasite infections is usually difficult without sacrificing host individuals, although longitudinal studies are one of the best ways to estimate the causalities in natural systems (Beldomenico et al., 2008; Telfer et al., 2010). Only a few studies have overcome these problems and specifically tested their causalities in wild conditions. A series of studies by Beldomenicoet al. (2008, 2009a, 2009b) successfully detected parasite infections on field voles Microtus agrestis in the field using a haematological method, and they monitored the infection status and host body condition combined with mark-recapture analysis of the host, clearly demonstrating positive feedback. A haemogram can be a useful indicator of infection; however, the authors did not observe parasites directly in the blood, and specific changes in infection intensity were not clarified. Blanchet et al. (2009) also demonstrated the causal relationships between parasite infections and host growth rates by estimating the growth of two host fishes from scales and otoliths, although these methods have potential estimation errors (e.g. Neilson, 1992), and the duration, frequency, and intensity of parasites before sampling were unknown. Moreover, these previous studies failed to evaluate host survival rate, even though positive feedback could likely cause host death in natural populations and, hence, affect wild host population dynamics (Beldomenico & Begon, 2010). Thus, previous findings have not sufficiently demonstrated the existence of condition–infection causality and positive feedback, necessitating more rigorous empirical evidence.
Here, we provide the first rigorous evidence of both causalities and positive feedback in wild populations by using a mark-recapture survey combined with structural equation modelling (SEM) in a wild stream fish–parasitic copepod system. SEM analysis is a powerful method for estimating the causalities in longitudinal datasets because of its simplicity and robustness (Fan et al., 2016). In fact, several studies have applied this approach to longitudinal studies and have revealed complex natural interactions (Almaraz, 2005; Byrnes et al., 2011). Our focused ectoparasitic copepod, Salmincolacf. markewitschi , is ideal for examining the causality between host body condition and parasite numbers because of their relatively large body size (2–5 mm; Kabata, 1969) and characteristic of attaching to the mouth cavities of host salmonid, white-spotted charr (Kabata, 1969), enabling us to track the change in infection intensity and host body condition longitudinally without sacrificing host fish. Further, previous studies have suggested that Salmincola spp. have negative impacts on host fitness components under rearing conditions, such as decline of fecundity (Gall et al., 1972), appetite (Nagasawa et al., 1994; Hiramatsu et al., 2001), and body condition (Nagasawa et al., 1998). Our previous studies also showed clear negative correlations between Salmincola cf.markewitschi loads and fish conditions in natural streams (Figure 1; Hasegawa & Koizumi under review 1). We conducted a mark-recapture survey of white-spotted charr Salvelinus leucomaenis andSalmincola cf. markewitschi infecting the host mouth cavity in the Shiodomari River in southern Hokkaido. We also evaluated the apparent survival rate of host fish during the mark-recapture period to assess whether positive feedback reduces host survival in the wild.