No island-effect on glucocorticoid levels for a rodent from a near-shore archipelago

Abstract

Island rodents are often larger and live at higher population densities than their mainland counterparts, characteristics that have been referred to as "island syndrome". Island syndrome has been well studied, but few studies have tested for island-mainland differences in stress physiology. We evaluated island syndrome within the context of stress physiology of white-footed mice (Peromyscus leucopus) captured from 11 islands and five mainland sites in Thousand Islands National Park, Ontario, Canada. Stress physiology was evaluated by quantifying corticosterone (a stress biomarker), the primary glucocorticoid in mice, from hair and its related metabolites from fecal samples. White-footed mice captured in this near-shore archipelago did not display characteristics of island syndrome, nor differences in levels of hair corticosterone or fecal corticosterone metabolites compared with mainland mice. We suggest that island white-footed mice experience similar degrees of stress in the Thousand Islands compared with the mainland. Although we did not find evidence of island syndrome or differences in glucocorticoid levels, we identified relationships between internal (sex, body mass) and external (season) factors and our hormonal indices of stress in white-footed mice.

Keywords: Biogeography; Corticosterone; Glucocorticoid; Islands; Physiology; Predation; Stress; Stressor.

Figure 1. Trapping locations of white-footed mice (Peromyscus leucopus) in Thousand Islands National Park, Ontario, Canada (USGS, 2016).

Islands on which trapping occurred are shaded in dark grey. Abbreviations: AI, Aubrey Island; BI, Beau Rivage Island; CE, Constance Island; CI, Camelot Island; EP, Escot Property; GA, Georgina Island; GI, Grenadier Island; HI, Hill Island, JC1, Jones Creek 1; JC2, Jones Creek 2; LB, Landon Bay; LI, Lindsay Island; MD, McDonald Island; MI, Mermaid Island; MT, Mallorytown; and TI, Thwartway Island.

Figure 2. Correlation between years in abundance of white-footed mouse (Peromyscus leucopus) at multiple trapping locations.

All trapping occurred in Thousand Islands National Park, Canada. The solid black line represents the linear relationship in abundance between the two years and the dashed line represents the predicted line if abundances were equal between years. Points that fall on the dotted line represent sites where abundance was nearly equal for white-footed mice between the two years. Abundance was measured in catch-per-unit-effort (CPUE, captures per 100 trap nights), which was corrected for tripped traps. Abbreviations represent individual sampling locations and correspond with those in Fig. 1.

Figure 3. Body mass (A), hair corticosterone (B), and fecal corticosterone metabolites (C) of white-footed mice captured during the summer (July–August) in two consecutive years.

Body mass was measured in grams, and both measures of CORT were measured in ng/g. All response variables were ln-transformed.

Figure 4. Hair corticosterone (CORT_hair; measured in ng/g and ln-transformed) levels increased with body mass in white food mice captured over two years in the Thousand Islands National Park, Canada.

The solid line shows a simple regression line; however, the relationship was tested within a linear mixed effects model. Body mass was measured in grams, and CORT_hair was measured in ng/g. Both measures were ln-transformed.

Figure 5. Seasonal variation in (A) hair corticosterone (CORT_hair), and (B) fecal corticosterone metabolites (CORT_fecal) from white-footed mice captured in spring and summer, 2016.

Females had lower hair corticosterone levels in the spring (May-June) than in the summer (July-August) (sex*season interaction; p < 0.0009; A), while both sexes had lower fecal corticosterone metabolites in spring than summer (p < 0.0001; B). Both CORT measures were quantified in ng/g and ln-transformed.

Figure 6. Correlation between hair corticosterone levels and fecal corticosterone metabolites for white-footed mice (n = 180) captured in the Thousand Islands National Park, Canada.

Both CORT measures were quantified in ng/g and ln-transformed.