Scat Collection and Dissection
We conducted this study over a 1.5-year period with scat collection beginning on May 11th, 2021, and ending on November 9th, 2022. We tracked wild gopher tortoises at the Richmond Tract twice weekly via radio telemetry. If an individual defecated during handling, we collected the sample labeled it with the tortoise’s ID number, location, and the date. Sometimes, we encountered fresh fecal samples belonging to unknown tortoises and thus collected them. For these samples, we GPS marked, dated, and labeled them with a unique identifier that indicated it came from an unknown individual.
Fecal samples were either dissected on the same day of collection or frozen for dissection at a future date. We performed scat dissections with forceps over laminated graph paper (29.59cm x 21.01cm) containing 5mm x 5mm grids as a static background reference, and we used this background reference to compare the relative contributions of food items to the total fecal volume. All food items recovered from fecal samples were identified to the lowest taxonomic unit or qualified as their own category (e.g., fur/hair was recovered from multiple species and subsequently categorized as mammal fur).
After dissecting fecal samples and separating contents over the laminated graph paper, we then visually estimated the relative contributions of each food category to the total scat volume was; this approach has been compared with other methods resulting in its acceptance for dietary studies (Klare et al. 2011). We quantified the contribution of five functional food categories – Grasses, Legumes, Other Plants, Fruit, and Animals – to the total fecal volume as either 0.01, 0.05, or in increments of 0.05 all the way to the total scat volume proportion of 1.00. If a value less than 0.05 remained after quantifying the relative contribution of all food items, we would allocate the amount to the most abundant food category in the sample.
Additionally, we identified all seeds extracted from fecal samples to the lowest taxonomic unit using dichotomous keys, online references, and consultations with local botanists (Gann et al. 2001, 2002, Wunderlin et al. 2016). Seeds were then counted and the dispersal syndromes were recorded using criteria from Ridley (1930) and Van der Pijl (1982) to distinguish species with the endozoochory syndrome from all other syndromes (Table 1).
Quantifying the Strength of Frugivory through Time
While other studies have used activity budgets to quantify the strength of frugivory (Pavelka and Knopff 2004), we directly measured the proportion of total fecal volume comprised of fruit. This measure was used as the response variable in all subsequent analyses on frugivory. As our predictor variables, we used indices of time in the form of calendar days to cover the whole calendar year and in a separate analysis, we used an interaction term between the season in which the sample was collected with the number of days elapsed within that season.