Discussion
In the host-parasite system we studied, chigger mites (E. alfreddugesi ) are highly abundant in the environment, and all age- and sex-classes (adults and juveniles; males and females) of S. undulatus are heavily parasitized (Pollock & John-Alder, 2020). However, we found no evidence that mites impose a growth cost, despite very high mite loads and consistent week-to-week rank ordering of mites on individual lizards. Indeed, within cohorts of yearlings, growth rate is unambiguously not correlated with mite load in either sex. Furthermore, across cohorts, both growth rate and mite loads were higher in 2019 than in 2016, which is opposite the prediction if mites imposed a growth cost. At any given mite load, females grow more quickly than males, which could suggest that mites impose a sex-biased growth cost. However, even in captivity in the complete absence of mites, SSD develops because females grow faster to become larger than males (Duncan et al., 2020). Thus, we find no evidence that growth rate in yearling juveniles of S. undulatus is affected by chigger mite ectoparasitism. It follows that the temporal correlation between the development of SSD and the seasonal attainment of sex-biased mite loads in July does not reflect a sex-biased growth cost of chiggers.
Our findings match those of several other studies that found no significant effects of mite parasitism in lizards. In S. virgatus , Abell (2000) reported no relationship between mite load and body condition, mating success, or survivorship to the following year, and Smith (1996) found no significant relationship between mite load and growth rate. In the anole species Norops polylepis , Schlaepfer (2006) found no sex differences in mite load and no growth costs associated with mite ectoparasitism. Mite parasitism was not associated with body condition in Gallotia atlantica lizards (García-Ramírez et al., 2005). Patterson and Blouin-Demers (2020) found no relationships between mite load and growth rates or survival in six different populations of Urosaurus ornatus . One caveat worth mentioning is that mite loads in the studies cited here were substantially lower than those on S. undulatus in the present study (S. undulatus : 0–435mites; U. ornatus : 0–120; N. polylepis : 0–37 mites, G. atlantica : 1–28 mites).
In contrast to our findings and the studies cited above, evidence of costs of mite parasitism has been reported in several species, including two close congeners of S. undulatus . An experimental study onS. virgatus found twice as many mites on males as on females and a strongly negative correlation between mite loads and growth rate (Cox & John-Alder, 2007). This result suggests that growth costs of mite ectoparasitism are greater in males than females (Cox & John-Alder, 2007), in contrast to what had previously been reported (Smith, 1996; Abell, 2000). In S. woodii , Orton et al. (2020) reported that both color quality and running endurance are negatively associated with mite load. In collared lizards (Crotaphytus collaris ), Curtis and Baird (2008) found evidence of a growth cost associated with heavy parasitism during the peak growing season. That study reported high mite loads, an average of 178.3 ± 65.2 mites on yearling male lizards, which are comparable to what we found on S. undulatus in the present study.
Overall, the inconsistent narrative regarding potential growth and fitness costs of mites suggests that effects of, or correlates of, mite parasitism can depend on which life history trait is being investigated and the timing of the studies, as ecological conditions (i.e., temperature, precipitation, food abundance, etc.) can fluctuate across the years and have different effects on life history traits. Furthermore, effects of mite parasitism can differ not only among species but also between populations of the same species as observed with the studies on S. virgatus (Smith, 1996; Abell, 2000; Cox and John-Alder, 2007). As such, it would be problematic to envision an overall evolved strategy to mite parasitism at a species level. Furthermore, future studies need to consider what life history trait is being examined when investigating potential effects of ectoparasitism.
While not a major consideration in our study, we posit that future studies on the costs of mite parasitism in lizards should consider whether the host population in the system would be expected to have evolved a strategy of resistance or tolerance towards the parasites in their study design. Parasites and their hosts coevolve in an evolutionary arms race in which parasites extract resources from their hosts, while hosts try to prevent or minimize the fitness costs of parasitism (May & Anderson, 1983; Restif & Koella, 2004; Carval & Ferriere, 2010). Within this context, hosts evolve strategies of resistance or tolerance to parasites, or some combination of the two. The ecological relationship between the populations of S. undulatus and E. alfreddugesi in this study (i.e., high environmental mite abundance and high host mite loads) is typical of a host-parasite system in which the host would be expected to have evolved tolerance to parasites (Råberg et al., 2009; Pollock & John-Alder, 2020). Hosts can evolve tolerance so effectively that the reduction in fitness caused by parasites may not be measurable (Råberg, 2014). Studies that do not find a growth cost associated with parasitism may be examining systems where the lizards have evolved a strategy of tolerance towards the mites. For example, it has been hypothesized that mite pockets evolved as a compensatory mechanism for reducing the harm caused by mite ectoparasitism because they concentrate mites in areas better equipped to heal and where they do not interfere with other functions, such as movement and vision (Arnold, 1986; Salvador et al., 1999; de Carvalho et al., 2006; Reed, 2014). Tolerance can be quantified using the slope of the relationship between fitness (or its proxy) and parasite burden (Burgan et al. 2019). Given that we found a slope of zero in the relationship between growth rate and mite load, our population of eastern fence lizards is highly tolerant of chiggers by the criteria we evaluated (Råberg et al., 2007; Råberg, 2014). Ornate tree lizards (Urosaurus ornatus ) have also been found to be highly tolerant of mite parasites (Paterson & Blouin-Demers, 2020). Host-parasite systems in which hosts appear to be highly tolerant of the parasite are not well-suited for investigating sex-biased costs of parasitism because any costs of parasitism are not easily measurable.
If chigger mites in our study population are in fact extracting a substantial cost from yearling S. undulatus , the cost may be at the expense of traits and functions other than growth. Rapid growth is expected to be under strong selection to ensure that yearlings grow to the minimum size of reproduction by the end of their first full activity season (Adolph & Porter, 1996; Haenel & John-Alder, 2002). Furthermore, female body size is positively correlated with clutch size, suggesting strong fecundity selection on body size and thus yearling growth (Angilletta et al., 2001; Haenel & John-Alder, 2002; Cox et al., 2003; Brandt & Navas, 2011; Jiménez‐Arcos et al., 2017). Thus, potential explanations for the absence of growth costs could include that costs of parasitism are 1) traded off against other functions, 2) too small to be of any consequence, or 3) compensated by increased dietary consumption. We now consider each of these possibilities.
If mites do impose costs on juveniles of S. undulatus , it seems likely that a potential growth cost might be traded off against other traits (Adolph & Porter, 1996). For example, reduced activity is one of the apparent costs of parasitism in common lizards (Lacerta vivipara ) and Western fence lizards (Sceloporus occidentalis ) (Clobert et al., 2000; Megia-Palma et al., 2020). In the present study, we did not quantify a specific measure of activity in S. undulatus . However, in a 2019 companion study conducted at our study site and on the same population of S. undulatus we found that home range area is greater in yearling males than in females (1101 ± 327 m2 vs 233 ± 62 m2) and is not correlated with growth rate (F2,1 = 0.14, p = 0.708), even while mite loads are greater in males than in females (Conrad, 2019; Yawdosyn, 2019). Home range area can be used as a proxy for energy expenditure on activity to assess the possibility of a trade-off against activity instead of growth (Christian & Waldschmidt, 1984; Hews, 1993; Main & Bull, 2000; John-Alder et al., 2009). Thus, our evidence suggests that mites do not impose an activity cost of mite ectoparasitism. However, while home range area can be used as a first approximation of activity, a study of the close congener S. occidentalis reported that lizards infected with malarial parasites maintained the same home range area as unparasitized lizards but showed reduced daily activity (Schall & Houle, 1992). Thus, time-intensive focal observations would be required for a definitive determination of whether mites impose an energetic cost on S. undulatus as seen through reduced activity.
Lacking evidence of costs of mite parasitism, it is possible that mites, even in high numbers, may not impose a measureable energetic cost, either directly or indirectly. We estimated the energetic cost of a single chigger mite to be approximately 0.004 J/day using estimates of chigger mite body size (body length = 0.4 mm, body mass = 0.9 μg; Johnson & Strong, 2000) and the metabolic scaling exponent of 0.75 (West et al., 2002). Taking the mean mite loads of yearling male lizards in June 2016 (58 mites), July 2016 (138 mites), June 2019 (93 mites), and July 2019 (119 mites), the average energetic costs of mite ectoparasitism for yearling male lizards would have been 0.232 J/day in June 2016, 0.552 J/day in July 2016, 0.372 J/day in June 2019, and 0.476 J/day in July 2019. For comparison, the energetic cost of growth is about 130 J/day for yearling males (Cox et al., 2005). Thus, in the aggregate, mites imposed an energetic cost that was a mere 0.002–0.004% of the energy cost of growth.
The energetic cost of direct energy extraction by mites is unlikely to be detectable as a growth cost because of its low value. These estimations, however, do not account for the potential indirect energetic costs of mounting an inflammatory immune response or the accompanying stress to the lizard in response to mite ectoparasitism, which we did not measure. However, potential energetic costs of immune or stress responses are far from certain. A meta-analysis by van der Most et al. (2011) found that selection for growth compromises immune function, but selection for immune function does not appear to affect growth. This suggests that the costs of growth are large relative to the costs of immune function. A study of parasites in side-blotched lizards,Uta stansburiana , found increased immunocompetence associated with parasitism, but no other measurable costs (Spence et al., 2017). Despite the lack of measurable costs in U. stansburiana , it can be costly for a host to up-regulate their immune system in response to parasites, forcing life-history trade-offs over time (Sheldon & Verhulst, 1996; Lochmiller & Deerenberg, 2000; French et al., 2009). Overall, while there may be some energetic cost to immunity, those costs are likely to small to be detectable as costs to growth. However, more research is certainly needed in this area.
Finally, growth and body condition may not have been associated with mite ectoparasitism in S. undulatus because environmental conditions favorable for ectoparasitism were at the same time favorable for abundant prey, including other arthropods and possibly even mites themselves. Thus, even while they were heavily parasitized by mites, lizards would be enabled to increase their energy consumption and compensate for costs of a high parasite load. The spring of 2019 had more precipitation and higher temperatures than the spring of 2016 (http://climate.rutgers.edu/stateclim_v1/nclimdiv/), and we saw an increase in both environmental mite abundance and lizard growth in 2019 as compared to 2016. Increased precipitation would have increased soil moisture and humidity, which are positively correlated with the abundance of arthropods, including mites (Zippel et al., 1996; Kardol et al., 2011; Prather et al., 2020). Temperature has also been shown to be positively correlated with arthropod abundance (Lessard et al., 2011). Mite activity is increased by a combination of moderately high temperatures and high humidity (Clopton and Gold, 1993). At the same time, higher temperatures can create opportunities for increased daily activity in S. undulatus , while also contributing to increased growth rates and faster maturation (Adolph and Porter, 1996). Greater arthropod abundance may have increased food availability for S. undulatus as well, facilitating increased growth. A greater understanding of the ecological interactions between S. undulatus , mites, and other arthropods could show if lizards predate upon the mites or increase their feeding rates on other arthropods due to increased mite ectoparasitism.
In summary, there are many possible explanations for why we did not find a growth cost associated with mite parasitism in either sex of S. undulatus in this study, even when males had higher mite loads than females. Female-larger SSD still developed in our study population. So, it is safe to conclude that sex-biased costs of mite parasitism are not contributing to the development of SSD in our study population ofS. undulatus . The lizards in our study population appear to be highly tolerant of mite ectoparasitism, but definitive experimental studies are required to rule out various trade-offs and effects of environmental food abundance. As such, while our study population also exhibits the common male-biased pattern of ectoparasitism, it also shows that we should be careful in making conclusions about the implications of sex differences in parasite intensity and prevalence.