Introduction
Host-parasite systems provide effective models in community ecology as they allow for replicated units through individual hosts or populations and precise measurements of total species richness (Poulin, Reference Poulin2019). Within a sample of hosts, parasites may exhibit a range of interactions varying from antagonistic to facilitative (Benesh and Kalbe, Reference Benesh and Kalbe2016; Hoffmann et al., Reference Hoffmann, Horak, Bennett and Lutermann2016). Non-random co-occurrence patterns may not only form from species interactions, but also due to responses to environmental conditions. The environment of a parasite includes both the host environment (i.e. host characteristics) and external environment (i.e. host population abundance and habitat of the host; Mouillot et al., Reference Mouillot, Šimková, Morand and Poulin2005; Van den Wyngaert et al., Reference Van den Wyngaert, Vanholsbeeck, Spaak and Ibelings2014; Dallas et al., Reference Dallas, Laine and Ovaskainen2019). Given that hosts are the habitat for parasites, host traits can alter parasite co-occurrence patterns through habitat preferences along with other host-specific characteristics like immune response (Dallas et al., Reference Dallas, Laine and Ovaskainen2019). Additionally, as ectoparasites are in direct contact with both their host and the external environment, abiotic factors can also play a role in structuring ectoparasite communities (Poisot et al., Reference Poisot, Guéveneux-Julien, Fortin, Gravel and Legendre2017). The recent use of joint species distribution modelling (JSDM) allows for the evaluation of associations between parasite species while identifying the role of host traits and external environmental factors (Dallas et al., Reference Dallas, Laine and Ovaskainen2019). Therefore, this approach enables the identification of whether co-occurrence patterns between parasites are non-random and whether species associations are shaped by host or external environmental factors can be identified.
Parasite communities can be highly dynamic, with parasite co-occurrence patterns varying over time and space (Krasnov et al., Reference Krasnov, Stanko, Lareschi and Khokhlova2020). Uneven distributions of parasite species are especially common and can have important ecological consequences on host mortality and immune response, as well as parasite competition and reproduction (McVinish and Lester, Reference McVinish and Lester2020). Parasite aggregation is a common ecological pattern that has been observed in many species, including small mammals (Buchholz and Dick, Reference Buchholz and Dick2017; Krasnov et al., Reference Krasnov, Shenbrot, Korallo-Vinarskaya, Vinarski, van der Mescht, Warburton and Khokhlova2019; Herrero-Cófreces et al., Reference Herrero-Cófreces, Flechoso, Rodríguez-Pastor, Luque-Larena and Mougeot2021). An aggregated pattern is one whereby many hosts are infected with few or no parasites, but some hosts are infected with many species of parasites. Aggregation generally decreases heterospecific co-occurrences; however, parasite species may influence individual host fitness during parasitic encounters, leading to outcomes other than aggregation (Morrill et al., Reference Morrill, Dargent and Forbes2017). Parasite communities remain largely understudied compared to free-living communities (Hoffmann et al., Reference Hoffmann, Horak, Bennett and Lutermann2016; Dallas et al., Reference Dallas, Laine and Ovaskainen2019). Further investigation of parasite co-occurrence patterns is required to gain insight into community dynamics of parasite assemblages and whether ectoparasites of red squirrels conform to the patterns found in similar host-parasite systems (including aggregation).
North American red squirrels (Tamiasciurus hudsonicus) play critical roles in their ecosystems as ecosystem modifiers for mammalian predators (Posthumus et al., Reference Posthumus, Koprowski and Steidl2015), important nest and seed predators (Pelech et al., Reference Pelech, Smith and Boutin2010; Steele and Yi, Reference Steele and Yi2020), and carriers of zoonotic pathogens (Bangari et al., Reference Bangari, Mouser, Miller, Stevenson, Vemulapalli and Thacker2007; Stenger et al., Reference Stenger, Clark, Kváč, Khan, Giddings, Prediger and McEvoy2015). Ectoparasites of red squirrels have received attention, primarily around their impact on host reproduction and general infestation patterns. For instance, fleas and mites of red squirrels are known to influence reproductive success, though the direction of this relationship can vary (Gooderham and Schulte-Hostedde, Reference Gooderham and Schulte-Hostedde2011; Patterson et al., Reference Patterson, Neuhaus, Kutz and Ruckstuhl2013). Additionally, flea infestation patterns show seasonal variation, with males more commonly infested (Gorrell and Schulte-Hostedde, Reference Gorrell and Schulte-Hostedde2008; Patterson et al., Reference Patterson, Neuhaus, Kutz and Ruckstuhl2015). Given their ecological importance, red squirrels offer a valuable system for studying parasite community ecology, where parasite interactions could have implications for host fitness and broader ecological patterns.
Co-occurrence patterns of ectoparasites (two flea species and a chigger mite) on red squirrels were examined. This study evaluated (1) whether ectoparasites exhibit non-random co-occurrence patterns and (2) which, if any, host or external environmental factors influenced parasite assemblages. It was expected that any non-random co-occurrence patterns would be positive associations, as commonly seen in studies of ectoparasites on small mammal hosts (Krasnov et al., Reference Krasnov, Matthee, Lareschi, Korallo-Vinarskaya and Vinarski2010; Nava and Lareschi, Reference Nava and Lareschi2014; Colombo et al., Reference Colombo, Nava, Antoniazzi, Monje, Racca, Guglielmone and Beldomenico2015; but see Hoffmann et al., Reference Hoffmann, Horak, Bennett and Lutermann2016; Veitch et al., Reference Veitch, Bowman and Schulte-Hostedde2020). It was also expected that parasite species co-occurrence patterns would be shaped by host and external environmental factors. Flea recruitment rates are often linked to the host's biology, while mite recruitment is more dependent on external environmental conditions (Linardi and Krasnov, Reference Linardi and Krasnov2013). If there are species associations between flea species, host traits will be more influential. If there are species associations between flea and mite species, either host traits or external environmental traits may play a role.
Materials and methods
Study area and field sampling
This study was conducted in Algonquin Provincial Park, Ontario, Canada (45°54′ N, 78°26′ W) from May to August of 2013 and 2015. North American red squirrels were captured in a 23-ha grid in an area of mixed deciduous-coniferous forest using Tomahawk live traps (Tomahawk Live Trap Co., Hazelhurst, Wisconsin, USA; detailed methods in Gorrell and Schulte-Hostedde, Reference Gorrell and Schulte-Hostedde2008). Tomahawk live traps were mounted 20-m apart at a height of approximately 1.5-m from the ground on platforms that were attached perpendicularly to randomly assigned mature trees that were large enough to attach and support the traps. Traps were padded with polyester stuffing and baited 06.00–19.00 h with an apple slice and a ~10-g mixture of peanut butter and oats, then checked in less than or equal to 2-h intervals.
Captured red squirrels were transferred to a handling bag, sexed, and assessed for reproductive status (non-reproductive or reproductively active). Squirrels were considered reproductively active if males were scrotal or females were lactating. Individuals were weighed using a Pesola® scale (±0.1 g) and received two metal ear tags with unique alphanumeric codes (National Band and Tag Co., Newport, Kentucky, USA). All methods used were reviewed and approved by the Animal Care Committee (ACC) at Laurentian University, Sudbury, Ontario, Canada, protocol number 2013-05-01.
Collection of ectoparasite specimens
Ectoparasites were collected from squirrels during each capture with a metal flea comb (teeth spacing <300-μm, one-tenth the size of the smallest fleas; Burgham Ltd., Toronto, Ontario, Canada) by combing ten times down the mid-back from the neck to the base of the tail, and ten times down the ventral surface from the sternum to the genitals. Ectoparasites collected were preserved in Eppendorf vials with 70% ethanol. Red squirrels in Canada are known to carry a variety of ectoparasites, including fleas (Opisodasys pseudarctomys, Orchopeas caedens, Monopsyllus vison, Taropsylla coloradensis), lice (Hoplopleura sciuricola, Neohaematopinus sciurinus), ticks (Ixodes scapularis, Ixodes angustus), and mites (Trombiculidae family; Gorrell and Schulte-Hostedde, Reference Gorrell and Schulte-Hostedde2008; Bouchard et al., Reference Bouchard, Beauchamp, Nguon, Trudel, Milord, Lindsay, Bélanger and Ogden2011; Patterson et al., Reference Patterson, Neuhaus, Kutz and Ruckstuhl2013; Bobbie et al., Reference Bobbie, Schmidt, Foley and Schulte-Hostedde2017). While combing is a reliable method for examining fleas, lice, and ticks, it is possible that mite species can be missed through visual inspection and combing (Beaumont et al., Reference Beaumont, Beaumont and Waterman2019); therefore, we may have been unable to identify additional mite species. However, this is a common method of quantifying ectoparasite communities on small mammals (e.g., Buchholz and Dick, Reference Buchholz and Dick2017; Pero and Hare, Reference Pero and Hare2018; Beaumont et al., Reference Beaumont, Beaumont and Waterman2019).
Taxonomic identification of ectoparasite specimens
Fleas were sent to the Canadian Centre for DNA Barcoding (CCDB) at the University of Guelph, Ontario, Canada. A glass fibre protocol (Ivanova et al., Reference Ivanova, deWaard and Hebert2006) was used to extract DNA from the macerated flea tissues; the 658-bp target region of the COI gene was amplified by polymerase chain reaction (PCR). Each 12.5-μL PCR mixture included 6.25-μL of 10% trehalose, 1.25-μL 10× PCR buffer, 0.625-μL (50-mm) MgCl2, 0.125-μL (10-μ m) of each oligonucleotide primer, 0.0625-μL (10-mm) dNTPs, 0.06-μL Taq polymerase and 2-μL ddH2O + 2-μL template DNA (Hajibabaei et al., Reference Hajibabaei, Singer, Hebert and Hickey2007). PCRs were run at the following thermal cycle conditions: 1-min at 94°C, followed by five cycles of 30-s at 94°C, 40-s at 50°C, and 1-min at 72°C, followed by 35 cycles of 30-s at 94°C, 40-s at 55°C, and 1-min at 72°C, and finally 10-min at 72°C. DNA extracts were PCR amplified using the forward and reverse primer-pair C_LepFolF (5′-ATTCAACCAATCATAAAGATATTGG-3′) and C_LepFolR (5′-TAAACTTCTGGATGTCCAAAAAATC-3′) respectively (Stein et al., Reference Stein, White, Mazor, Miller and Pilgrim2013). PCR products were bidirectionally sequenced using Sanger sequencing with BigDye v3.1 using an ABI 3730 × l DNA Analyzer (Applied Biosystems, Foster City, CA).
The Refined Single Linkage (RESL) algorithm was used to cluster species (Ratnasingham and Hebert, Reference Ratnasingham and Hebert2013). Sequences and other pertinent specimen data (ex. date collected, its host, etc.) were uploaded to the Barcode of Life Data Systems (BOLD), and their OTUs were ascribed a Barcode Identification Number units (BINs); each BIN is populated by individual specimens having high sequence similarity and connectivity (Ratnasingham and Hebert, Reference Ratnasingham and Hebert2013). Nucleotide sequence homology searches were performed on the sequences obtained from the CCDB using NCBI BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi).
Following sequence analysis and clustering, specimens were mounted on slides similar to the procedures laid out in Richards (Reference Richards1964). Maceration was completed by sequencing technicians at the CCDB. The exoskeletons were then transferred singly into 70% then 95% ethanol solutions for 3–5 min each for dehydration. Immediately after, they were placed overnight in oil of cloves. Specimens were mounted with Permount medium (Fisher Scientific) and placed in a drying oven at 50°C for a week. Further batches of slides were alternatively left to dry on the counter overnight. These mounts were imaged and mailed to Dr T. Galloway (University of Manitoba, Canada) for morphological identification. Dr Galloway was unaware of the DNA-barcoding results at the time of morphological identification, therefore providing a separate identification process to further support the DNA barcoding results. Flea morphological identification matched barcode species identification, apart from a single instance (which was excluded from the statistical analysis). One flea specimen failed the sequencing process and was only identified to species morphologically by Dr Galloway. The mite specimens did not undergo barcode species identification but were stored in 70% ethanol and mailed to Dr H. Proctor (University of Alberta, Canada) for morphological identification (Bobbie et al., Reference Bobbie, Schmidt, Foley and Schulte-Hostedde2017).
Statistical analysis
Recaptures of individuals showed that fleas and mites take ~3–4 days to recolonize a host after parasite removal (data not shown). Therefore, recaptures of individuals that had ectoparasites removed were not included in the dataset unless at least a week had passed since ectoparasite removal. We only focused on occurrence in our statistical analysis, as parasite removal may influence intensity measures. Furthermore, only ectoparasites with at least 10 occurrences were included to avoid complications due to small sample sizes. This led to the exclusion of a single flea species (Opisodasys pseudarctomys). Ten flea specimens were missing information on the host's ID number or reproductive status and were also excluded from the statistical analysis. Analyses were conducted using statistical software package R version 4.0.2. Parasite prevalence, defined here as the proportion of host individuals infested with a parasite, was calculated by dividing the total number of infestation occurrences by the total number of host captures. Confidence intervals (95%) for parasite prevalence were calculated by Clopper-Pearson's exact method for binomial proportions (‘GenBinomApps’ package version 1.1). We ran linear mixed-effects models (‘nlme’ package version 3.1–149) with days since previous capture and Julian date as fixed effects and individual ID number as a random effect, but there was no effect of days since previous capture on occurrence of the flea species Ceratophyllus vison (β = 0.001, P = 0.674), Orchopeas caedens (β = −0.002, P = 0.315), or the mite Neotrombicula harperi (β < 0.001, P = 0.958; Pinheiro et al., Reference Pinheiro, Bates, DebRoy, Sarkar, Heisterkamp, Van Willigen and Maintainer2012). Individuals with only a single capture were excluded from the linear mixed-effects models.
We fitted the model using the ‘boral’ package version 1.8.1 (Hui, Reference Hui2016). This statistical method uses a model-based, parsimonious approach to ordination, with a generalized linear model and incorporated latent variables. Boral also accounts for host and external environmental factors while examining residual co-occurrence patterns (i.e. identification of species associations after controlling for investigated predictors). The model uses a species occurrence matrix as the response variable and host and external environmental covariates as explanatory predictors. We fitted a correlated response model, which examined how the parasite assemblage is explained by the host and external environment (Hui, Reference Hui2016). Host sex and date were incorporated as fixed effects to describe the host and external environment. Date was centred and scaled by the mean and standard deviation. Host ID was included as a random effect.
We ran the Bayesian MCMC sampler in boral allowing for two latent variables with 200 000 iterations, the first 100 000 discarded as burn-in, and the remaining thinned by a factor of 100. The selection of two latent variables compromises between model complexity and appropriate evaluation of species co-occurrence patterns after controlling for host and external environmental predictors (Letten et al., Reference Letten, Keith, Tozer and Hui2015; Warton et al., Reference Warton, Blanchet, O'Hara, Ovaskainen, Taskinen, Walker and Hui2015; Taranu et al., Reference Taranu, Pinel-Alloul and Legendre2021). Random row effects were included to account for spatial variation. Including random row effects is equivalent to including a random intercept in mixed models, as a normal distribution is drawn with mean zero and unknown variance (Hui, Reference Hui2016). The model was evaluated using stochastic search variable selection (SSVS) to determine predictors included in the final model (George and McCulloch, Reference EI and Mcculloch1993). Only predictors with an SSVS mean >0.5 were included in the final model (sex: mean = 0.61, s.d. = 0.36; date: mean = 0.82, s.d. = 0.29). Predictors with an SSVS mean <0.5 were removed in a backwards stepwise format. This included the following predictors: host body mass, host reproductive status, host population abundance, and year removed based on SSVS values. Convergence was assessed using Dunn–Smyth residual and normal quantile residual plots and the Geweke diagnostic (Supplemental Information; Geweke, Reference Geweke, Bernado, Berger, Dawid and Smith1992). We identified the relative importance of host and external environmental factors for each parasite species and constructed a horizontal line plot showing 95% highest posterior density (HPD) intervals for the column-specific regression coefficients. We determined parasite species co-occurrences due to host and external environmental responses using the function ‘get.enviro.cor’ and the remaining co-occurrence patterns after controlling for predictors using the function ‘get.residual.cor’. To measure how well the predictors described the species assemblage, we calculated a proportional difference in the trace of the residual covariate matrix between the correlated response model and a pure latent variable model (where species are regressed against unknown covariates to produce an unconstrained ordination for visualizing site and species patterns; Warton et al., Reference Warton, Blanchet, O'Hara, Ovaskainen, Taskinen, Walker and Hui2015; Hui, Reference Hui2016). Plots of co-occurrence patterns were produced using the function ‘corrplot’. A horizontal line plot of the predictors and a plot for the estimates of variance partitions (Supplemental Information) were constructed using the functions ‘gg_coefsplot’ and ‘gg_varpart’ respectively from ‘ggboral’ package version 0.1.7.
Results
The final dataset included 53 red squirrels (19 females, 34 males) with 207 captures, ranging from 1 to 19 captures for each individual. From DNA barcoding and morphological species identification, three prominent ectoparasite species were identified and analysed; two flea species (C. vison and O. caedens) and a chigger mite species (N. harperi) (Table 1). Eighty-four flea specimens were successfully sequenced and NCBI Blast identified two flea species with more than 95% identity with the known subtypes in GenBank (Table 2).
Ectoparasite prevalence and number of red squirrels (n = 53 individuals, 207 captures) infested
Individuals captured multiple times that were infested by an ectoparasite species in at least one capture are included in the count of infested individuals.
Highest identity coverage NCBI BLAST hit sequences from the COI gene for flea specimens collected from red squirrels (n = 84 sequences)
There were no significant associations between parasite species after controlling for variation in occurrence with host sex and date. There was a negative association between the flea O. caedens and mite N. harperi and a positive association between the flea C. vison and N. harperi explained by host or external environmental responses (Table 3). Parasite species had varied responses to host sex and date (Fig. 1). O. caedens flea occurrence was greater on male than female hosts. All parasite species occurrences varied with date, where C. vison flea and N. harperi mite occurrences increased between spring and summer, and O. caedens flea occurrence decreased. Notably, the estimated residual covariate matrix decreased from 31.46 to 17.48 from a pure latent variable model to the correlated response model, suggesting that the investigated predictors explained only 13.98% of the co-occurrence patterns between parasites.
Summary of correlations between ectoparasite species (fleas (Orchopeas caedens and Ceratophyllus vison) and mite (Neotrombicula harperi)) on red squirrels (n = 207 captures)
Bolded terms are those with 95% highest posterior density (HPD) intervals that do not include 0 (considered to have strong support).
Effect estimates of host and external environment covariates as predictors of ectoparasite species occurrence. Circles represent posterior mean coefficients and horizontal lines represent 95% highest posterior density (HPD) intervals. Vertical dotted lines indicate the zero value.
Discussion
A combination of positive, negative and random co-occurrence patterns was observed between the three prominent ectoparasite species on red squirrels. While we expected largely aggregative patterns, there were no significant associations between parasite species pairs when controlling for host sex and date. The significant associations observed, both positive and negative, were identified between the flea species and N. harperi mite and were explained by date. These trends suggest that associations between flea and mite species on red squirrels are shaped by the external environment rather than host traits or species interactions.
We did not identify any significant associations between parasite species pairs after controlling for host sex and date, indicating a lack of species interactions. Ectoparasite communities of small mammals tend to exhibit non-random co-occurrence patterns, particularly aggregation (Krasnov et al., Reference Krasnov, Matthee, Lareschi, Korallo-Vinarskaya and Vinarski2010; Nava and Lareschi, Reference Nava and Lareschi2014; Colombo et al., Reference Colombo, Nava, Antoniazzi, Monje, Racca, Guglielmone and Beldomenico2015), but this is not always the case (see Krasnov et al., Reference Krasnov, Stanko and Morand2006). Notably, Eurasian red squirrels (Sciurus vulgaris; Romeo et al., Reference Romeo, Pisanu, Ferrari, Basset, Tillon, Wauters, Martinoli, Saino and Chapuis2013) and invasive Pallas's squirrels (Callosciurus erythraeus; Mazzamuto et al., Reference Mazzamuto, Pisanu, Romeo, Ferrari, Preatoni, Wauters, Chapuis and Martinoli2016) have hosted poor parasite assemblages with little variation in composition. Our investigated parasite community was also very small, which may provide little opportunity for transmission across hosts (Romeo et al., Reference Romeo, Pisanu, Ferrari, Basset, Tillon, Wauters, Martinoli, Saino and Chapuis2013). This suggests that the investigated ectoparasites of red squirrels are largely unstructured by interspecific competition or facilitative processes and follow what we would expect from stochastic processes.
While parasite occurrence patterns do not seem to be altered by parasite species interactions, significant co-occurrence patterns were identified in the correlated response model, suggesting that host and external environmental factors play a stronger role. N. harperi mites had a negative association with O. caedens fleas and a positive association with C. vison fleas. These co-occurrence patterns align with temporal trends in occurrence on red squirrel hosts that were also identified from our model (i.e. associations with date). Temporal trends in the external and host environment may play more of a role in whether parasite species pair co-occur more or less often than expected by chance. Ectoparasites have direct contact with the external environment, compared to endoparasites within the host's body (Bush et al., Reference Bush, Fernández, Esch and Seed2001), and consequently, co-occurrence patterns may be much more structured by these environmental conditions that change over time. However, it is important to note that the investigated predictors explained only ~14% of co-occurrence patterns between species and date may play a limited role in influencing the association between the investigated species pairs. This may also suggest that even weak associations of ectoparasite occurrence with date could structure co-occurrence patterns between species.
Notably, we only saw trends between N. harperi mites and both investigated flea species, but not between the two flea species. N. harperi mites had the strongest response to sampling date, which may explain this result. Temporal trends in the external environment, particularly ambient temperature and relative humidity, are well known to strongly impact the imago and pre-imago stages of mites and fleas, with downstream effects on survival, development time, and patterns of blood digestion (Linardi and Krasnov, Reference Linardi and Krasnov2013). Given that mite recruitment, compared to fleas, is often more tied to the external environment, we may expect greater variation in mite occurrence over time associated with changes in temperature and humidity (Linardi and Krasnov, Reference Linardi and Krasnov2013). This is especially true in Trombiculid mites, such as N. harperi, which are only parasitic in their larval stage and are often associated with particular soil or habitat characteristics rather than hosts (Timm, Reference Timm and Tamarin1985). Therefore, the stronger ties that N. harperi shares with the external environment rather than the host environment may explain why date shaped associations between N. harperi and the two flea species, but not between the two flea species.
The negative trend of O. caedens flea occurrence over the sampling period was surprising, given that there was an increase in infection patterns over the same sampling period in the study area 10 years prior (Gorrell and Schulte-Hostedde, Reference Gorrell and Schulte-Hostedde2008). However, a similar negative trend was seen over the same seasonal sampling period in eastern chipmunks (Amin, Reference Amin1976) for a flea in the same genus, Orchopeas h. howardii. Moderate differences between years in seasonal trends of Orchopeas sexdantatus flea infestations on desert woodrats have also been observed, with occurrence patterns varying with humidity (Lang, Reference Lang2014), suggesting that we may expect differences in occurrence patterns of our investigated flea species between different years of study. Temporal temperature trends can shape flea occurrence patterns through impacts on flea oviposition, egg clutch size, immature development and survival (Bossard, Reference Bossard2022). Thus, we may expect changes in O. caedens flea occurrence throughout our study.
N. harperi mite and C. vison flea occurrence increased over the sampling period. This was expected, as previous studies have noted that both of these parasite groups tend to exhibit seasonal patterns where they increase during the warmer months. N. harperi mites tend to emerge in May or June and reach higher occurrences in the summer (Brennan and Wharton, Reference Brennan and Wharton1950). Our sampling period falls within periods where infestation rates should be increasing, as chigger mites (such as N. harperi) are only active during warm months in northern temperate areas (Dietsch, Reference Dietsch2005). Additionally, fleas in the Ceratophyllus genus often exhibit peaks during summer months (Samurov, Reference Samurov1990; Cyprich and Krumpal, Reference Cyprich and Krumpal2001; Haukisalmi and Hanski, Reference Haukisalmi and Hanski2007). The results of our study are consistent with previous work on N. harperi and C. vison on mammalian hosts.
Approximately 14% of the variation in co-occurrence patterns between the investigated parasite species was explained by our predictors, indicating that there may be additional factors that play a more substantial role in shaping parasite communities in this system. While we did investigate date, particular seasonal conditions such as temperature, rainfall, and humidity can influence flea occurrence on squirrels (Goldberg et al., Reference Goldberg, Conway and Biggins2020; Smith et al., Reference Smith, Smith, Working, Russell, Krout, Singh and Sih2021). Flea species can have similar or differing physiological tolerances to environmental conditions, which can shape their occurrence patterns (Smith et al., Reference Smith, Smith, Working, Russell, Krout, Singh and Sih2021). Precipitation levels from the previous year can have a strong effect on flea abundance on squirrels (Goldberg et al., Reference Goldberg, Conway and Biggins2020). For chigger mites of small mammals, occurrence may vary with habitat type (Choi et al., Reference Choi, Kim, Kim and Kwon2019; Matthee et al., Reference Matthee, Stekolnikov, van der Mescht, Froeschke and Morand2020). The host community within each habitat can also influence the occurrence of chigger mites, as additional common host species may facilitate greater mite abundance (Matthee et al., Reference Matthee, Stekolnikov, van der Mescht, Froeschke and Morand2020). The abundance of host species that spend more time on the ground near these soil-dwelling mites, such as voles, may provide food resources to support the mite population (Veitch, Reference Veitch2020). Therefore, further investigation of environmental conditions, such as temperature, precipitation, humidity, habitat type, and the host community may provide better explanatory predictors of parasite co-occurrence on red squirrels.
The occurrence of O. caedens fleas was greater on male red squirrels. Male host bias is a common trend in fleas parasitizing rodents (Perez-Orella and Schulte-Hostedde, Reference Perez-Orella and Schulte-Hostedde2005; Krasnov et al., Reference Krasnov, Bordes, Khokhlova and Morand2012) and has been observed in other study systems of red squirrels (Patterson et al., Reference Patterson, Neuhaus, Kutz and Ruckstuhl2015). However, a female-bias in flea parasitism was previously identified in this host population (Gorrell and Schulte-Hostedde, Reference Gorrell and Schulte-Hostedde2008). Inspection of the raw dataset from Gorrell and Schulte-Hostedde (Reference Gorrell and Schulte-Hostedde2008) demonstrated that there was higher flea parasitism on males recorded in June and August, suggesting that higher flea occurrence on male red squirrels is not uncommon in this population. Male squirrels often maximize number of matings with associated costs to their body condition and immune function (Scantlebury et al., Reference Scantlebury, Maher McWilliams, Marks, Dick, Edgar and Lutermann2010). Higher testosterone levels in males can lead to immunosuppressive effects (Folstad and Karter, Reference Folstad and Karter1992; Zuk and McKean, Reference Zuk and McKean1996; Foo et al., Reference Foo, Nakagawa, Rhodes and Simmons2017; but see Rolff, Reference Rolff2002). Males also often engage in behaviours that increase encounter rates such as increased mobility in larger or overlapping home ranges, which increases contact between hosts and the likelihood of potential transmission (Krasnov et al., Reference Krasnov, Bordes, Khokhlova and Morand2012). Male biases in O. caedens flea parasitism could be related to fitness benefits to the parasites, as fleas may feed more effectively and produce more offspring on male hosts (Khokhlova et al., Reference Khokhlova, Serobyan, Krasnov and Degen2009). Therefore, male red squirrels may be more susceptible to infestation of fleas. However, given that the investigated predictors explained ~14% of co-occurrence patterns between parasite species, further experimentation is needed to examine variation in flea occurrence with host sex.
The ectoparasite community of red squirrels investigated in this study did not appear to be structured by parasite-parasite species interactions but largely by changes in infestation patterns over our sampling period. External environmental conditions that fluctuate over time, such as temperature, rainfall, and humidity could play strong roles in structuring ectoparasite infestations and the impact of these factors on ectoparasite communities should be further explored. Ectoparasite communities of red squirrels may be shaped by seasonal or temporal shifts in the physical environment, but the host and external environmental factors that shape these associations are yet to be fully identified.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0031182024001513.
Data availability statement
Sequences, images, and original trace files are available for viewing in the project ‘Ectoparasites from Squirrels in Algonquin Park’ (ESAP), on the Barcode of Life Data Systems (http://www.boldsystems.org/).
Acknowledgements
The Algonquin Wildlife Research Station provided logistic support. Thanks to Dr Terry Galloway and Dr Heather Proctor for species identification of the flea and mite species, respectively. Thanks to Alex Borisenko, Susana Romo, and the rest of the Collections team at the Canadian Centre for DNA Barcoding (CCDB) in Guelph, Ontario, for genetic analysis of flea specimens. We also thank Alana Sajatovic for investigation, Colleen Bobbie, Trisha Hewitt, Elliott Schmidt, and Charlie Wilkes for collection of flea specimens, Pauline van Leeuwen for input on formal analysis, and Dr Dave Shutler, Breanna Pyott, and Sarah Falconer for constructive comments on an early version of this manuscript.
Author contributions
J. V. conceived and designed the study. J. V., A. S.-H., and D. K. curated the data. J. V. conducted the formal analysis and created the visualizations. A. S.-H. and J. B. secured funding. J. V. developed the methodology. J. V. prepared the original draft. A. S.-H. and J. B. reviewed and edited the manuscript.
Financial support
The Ontario Ministry of Natural Resources and Forestry (J. B.), Ontario Parks (J. B.), and the Natural Sciences and Engineering Research Council of Canada (A. S.-H., grant number 298367-2009) provided financial support.
Competing interests
The authors declare there are no conflicts of interest.
Ethical standards
All methods used were reviewed and approved by the Animal Care Committee (ACC) at Laurentian University, Sudbury, Ontario, Canada, protocol number 2013-05-01.
Open access
