Stable isotopes reveal seasonal dietary responses to agroforestry in a venomous mammal, the Hispaniolan solenodon (Solenodon paradoxus)

Abstract

While trends in tropical deforestation are alarming, conservation biologists are increasingly recognizing the potential for species survival in human-modified landscapes. Identifying the factors underlying such persistence, however, requires basic ecological knowledge of a species' resource use. Here, we generate such data to guide conservation of an understudied venomous mammal, the Hispaniolan solenodon (Solenodon paradoxus), that occupies a mosaic landscape of agriculture and forest fragments in the western Dominican Republic. Using feces collected in both wet and dry seasons, we found significant differences in the stable isotope values of carbon (δ¹³C) between pasture (-24.63 ± 2.31‰, Las Mercedes) and agroforestry (-28.07 ± 2.10‰, Mencia). Solenodon populations in agricultural areas occupied wider isotopic niche spaces, which may be explained by more diverse resource within these patches or individuals combining resources across habitats. We detected elevated δ¹⁵N values in the dry season of pasture areas (8.22 ± 2.30‰) as compared to the wet season (5.26 ± 2.44‰) and overall narrower isotopic niche widths in the dry season, suggestive of the impacts of aridity on foraging behavior. Our work highlights the importance of considering a more nuanced view of variations in 'modified' or "agricultural" landscapes as compared with strictly protected national parks. We suggest that seasonal differences in foraging should be considered as they intersect with landscape modification by landowners for maintaining resources for focal consumers. This work adds to a growing body of literature highlighting that fecal stable isotopes are a non-invasive and cost-effective monitoring tool that is particularly well-suited for cryptic small mammal species, ensuring actionable and evidenced-based conservation practices in the tropic's rapidly changing landscapes.

Keywords: Caribbean; agroforestry; insectivory; museum collections; small mammal; stable isotope analysis.

FIGURE 1

(a) An extremely fresh solenodon fecal sample. Typical signs of solenodon foraging include digging in leaf litter (b) and rotten logs (c) with their probing, ball‐and‐socket jointed nose and robust and clawed forelimbs. (d) A Cuban almiqui (Atopogale cubanus) at the Smithsonian National Museum of Natural History—this species is extremely rare in museum collections

FIGURE 2

(a) The Caribbean island of Hispaniola. Extant subspecies ranges of Solenodon paradoxus and color‐coded by subspecies: dark gray is S. paradoxus haitiensis, light gray is S. paradoxus paradoxus, and white is S. paradoxus woodi (redrawn from Turvey et al., 2017). Approximate hair sample locations are shown in purple (Haiti: 1. Sud; Dominican Republic: 2. Cordillera Central, 3. Santiago de los Caballeros, 4. La Vega, 5. Samana, 6. Sabana del Mar). The region of fecal sample collection is encircled by a dashed line. (b) The area of fecal sample collection showing Mencia and Las Mercedes; orange = dry season collection, blue = wet season collection. Close up of (c) Mencia and (d) Las Mercedes, with dashed lines indicating different landowners and local place names. Maps produced using ESRI ArcGIS, map credits: CGIAR, Earthstar Geographics, Garmin, USGS, METI/NASA, HERE, FAO

FIGURE 3

Values of (a) δ¹⁵N (‰) and (b) δ¹³C (‰) from Las Mercedes across wet (blues) and dry (oranges) seasons. Comparison of the (c) total area (TA) of the convex hulls and 95% confidence intervals around the bivariate means of different groupings (dry forest (dark orange), dry agriculture (light orange), wet forest (dark blue), and wet agriculture (light blue)) and (d) the Bayesian standard ellipse area (SEA_b), with white Xs indicating the maximum likelihood estimates of SEA_c. Sample sizes are in Table 2

FIGURE 4

Values of (a) δ¹⁵N (‰) and (b) δ¹³C (‰) from Mencia and Las Mercedes across habitat types (forest = greens, agriculture = browns). Comparison of the (c) total area (TA) of the convex hulls and 95% confidence intervals around the bivariate means of different groupings (Mencia forest (light green), Mencia agriculture (light brown), Las Mercedes forest (dark green), and Las Mercedes agriculture (dark brown)) and (d) the Bayesian standard ellipse area (SEA_b), with white Xs indicating the maximum likelihood estimates of SEA_c. Sample sizes are in Table 2