Rodent ectoparasite diversity in response to anthropogenic disturbance

Description

Notes

Methods

Study sites. Our study sites were in the municipalities of Nuevo San Juan, Tancitaro, and Uruapan in the state of Michoacan, Mexico, in the Trans-Mexican Volcanic Belt.  Our aim was to find the relationship between ectoparasite diversity and anthropogenic disturbance. We defined the disturbance status of our sampling sites based on their forest structure and cover. Five paired disturbed and undisturbed sampling sites were chosen to be, on average, 300 m apart between sites within a pair. Undisturbed sites had high forest cover (see results below), while disturbed sites had highly disturbed vegetation due to human activities. At each site, we measured plant (trees and shrubs) height, density, and growth form along three transects. Each transect was 50 m long and 2 m wide (Perkins et al., 2019). Each plant was identified at the species level with the help of community professionals from the Forestry Direction of Nuevo San Juan.

To sample rodents, at each site, we placed 50 Sherman traps in a 5 × 10 grid arrangement (spaced at 10 m). We baited the traps with a mixture of peanut butter, vanilla essence, and oats, which have given good results in similar environments (Vázquez, Medellín, and Cameron, 2000). Paired disturbed and undisturbed sites were sampled simultaneously on each event. Sites were sampled for two nights in both dry (May 2022) and wet (August 2022) seasons. As rodents can be sensitive to light, we avoided sampling on full moon nights. Each captured rodent was identified, photographed, and measured (total length, tail length, ear length, hind foot length measured in millimetres, and weight measured in grams). We inspected the fur of each rodent with entomological tweezers systematically on the dorsal area, the ears, and limbs, for approximately 2 min. All observed ectoparasites were removed. The use of anaesthetics during the inspections was not necessarily due to the docile nature of the individuals captured. However, signs of distress were assessed prior to inspecting each individual, and stressed individuals were immediately released and monitored until recovery. Our procedures were approved by the Secretaría de Medio Ambiente y Recursos Naturales from Mexico (SEMARNAT), under the permit number SGPA/DGVS/02787/22. Collected ectoparasites were preserved in 70% ethanol for later identification and counting. In the laboratory, collected mites (Mesostigmata) and chiggers (Trombidiformes) were submerged in a chloral hydrate solution until translucid, and then individually mounted in Hoyer's medium (Moravvej et al., 2016). Fleas were submerged in KOH until transparent and mounted in Canada balsam (Wirth and Marston, 1968). All ectoparasite species were identified based on taxonomic keys specific for each ectoparasite order, family, or genus if necessary (Acosta and Morrone, 2003; Hoffmann, 1990; Keirans et al., 1989; Loomis, 1971; McDaniel, 1979; Pratt, 1956).

Data analyses. The diversity of rodent and ectoparasite communities was compared between disturbed and undisturbed sites using a Hutcheson t-test o compare the Shannon Diversity Index performed in the ecolTest R package (R Core Team, 2022; Salinas and Ramirez-Delgado, 2021). We compared ectoparasite species richness among site types using species accumulation curves inferred using the iNEXT package from R (Hsieh, Ma, and Chao, 2016; R Core Team, 2022). We also calculated the sample coverage, defined as a measure of sample completeness, giving the proportion of the total number of individuals in a community that belong to the species represented in the sample (Chao and Jost, 2012). We compared the relative abundance of ectoparasite orders between site types (i.e., Siphonaptera, Ixodida, Mesostigmata, and Trombidiformes) using a Mann-Whitney test in the R base package (R Core Team, 2022). Pielou's evenness index was used to compare ectoparasite communities between site types using a Mann-Whitney test (see F2 in Supplementary material S2) implanted in the R base package (R Core Team, 2022). We calculated the prevalence of taxonomic orders of ectoparasites on rodents per sampling event and compared them among disturbed and undisturbed sites using a Mann-Whitney test. To compare beta diversity of ectoparasite communities, we applied a non-metric multidimensional scaling (NMDS) using the 'metaMDS' function in vegan (Oksanen et al., 2022; R Core Team, 2022). We used the Bray-Curtis distance, and the treatment factor using 1,000 permutations. We then performed an analysis of similarity (ANOSIM) using the 'anosim' function from vegan under the same parameters.

We developed binomial generalised linear models to predict the presence of ectoparasites in rodents. We generated a global model in the MuMIn R package (Barton, 2018) to assess the influence of rodent life-history traits, rodent genus, rodent relative abundance, and habitat condition on the presence/absence of ectoparasites on rodents. Rodent host variables included activity patterns (nocturnal/crepuscular), habitat domain (arboreal/ground-dwelling), sex, body mass index, host genus, and rodent relative abundance per season for each site. The environmental variables included in the global model were season (wet/dry), shrub density, tree density, and site type (disturbed/undisturbed). As more than one model met the ΔAICc < 2 criterion, we applied a model full averaging to produce an average of the estimates with these models (Burnham and Anderson, 2002).

We calculated the spatial patterns of ectoparasite presence on rodent hosts through the segregation index (S) using the dixon2002 function from the ecespa R package (de la Cruz, Dixon, and Blanco-Moreno, 2023; R Core Team, 2022). This index describes the tendency of one species to be associated with itself or with other species (Dixon, 2002). S > 0 indicates that rodents hosting ectoparasites are segregated, whereas S < 0 indicates that rodents with ectoparasites are clustered and found near one another. S close to 0 suggests a random distribution of rodents with ectoparasites (de la Cruz et al., 2023).