Northern Prairie Wildlife Research Center
We oven-dried each scat for 48 hours at 70 C (Potvin et al., 1988; Corbett, 1989) and recorded the dry mass. Each scat was washed separately with water using two 36-mesh/cm² sieves, and residues were air dried (Greenwood, 1981). We identified residues of hair, teeth, eggshell fragments, feathers, scales, plant material and exoskeletal parts to the lowest taxon possible. Hair was identified primarily by micro structure (Moore et al., 1974); swift fox hair was excluded from analyses when it occurred in trace amounts because we assumed it was ingested during grooming. Bird eggshell fragments were separated based on color and thickness. We categorized eggshells ≤0.15 mm thick as passerine and other small species, >0.15 mm to <0.20 mm as medium-sized birds and ≥0.20 mm as ring-necked pheasants (Phasianus colchicus) and other large-sized birds (Greenwood, 1981; Mallory and Weatherhead, 1990). Seeds were identified using reference collections and manuals (Martin and Barkley, 1961; Davis, 1993). Seeds that occurred in trace amounts were excluded from analysis and were not presented because ingestion was likely incidental to consuming other foods (e.g., seed-eating prey items). Grass in swift fox scats was excluded from the calculation of overall plant frequency and all other analyses because the nutritional value of grass in a carnivore diet is questionable (Fichter et al., 1955) and most often grass occurred in trace amounts. We followed Zumbaugh et al. (1985) in defining "remains of animals too large to be killed by swift foxes" as carrion and other animals as prey. We visually estimated the percent volume of each food item for each scat (Kitchen et al., 1999). Estimates of volume were only used to help us evaluate the importance of plants, as food or if ingestion was incidental to consumption of other prey. For example, a single sunflower seed shell versus ingestion of many seeds indicated by a high volume of shells in a scat.
Scats were separated by season into spring (1 March-30 May 1996), summer (1 June-31 August 1996), fall (1 September-31 October 1993) and early winter (1 December-31 December 1993). Samples from fall and early winter were combined for analysis (hereafter fall) because sample sizes were small and the early winter weather conditions were mild and similar to the fall collection period (National Oceanic and Atmospheric Administration, 1993).
We used frequency of occurrence to present our results. Food items found were recorded as present or absent. Using chi-square analyses (FREQ PROC, SAS Inst., Inc. 1990), we compared frequencies of occurrence of prey items between cropland and rangeland by season. When expected values were <5 we used Fisher exact tests to evaluate variation in frequency of occurrence (Steel and Torrie, 1980). We limited our comparisons to prey items that occurred in ≥10% of scats in either area during a season.
In 1996 from May through September we conducted monthly spotlight surveys using methods similar to Ralls and Eberhardt (1997). Two observers conducted surveys between 2300 and 0300 h, driving 10-25 km/h along the same routes used for collection of scats. We recorded observations of swift foxes and other carnivores, but also Ord's kangaroo rats (Dipodomys ordii), cottontail rabbits (Sylvilagus spp.) and black-tailed jackrabbits (Lepus californicus). We calculated the rates of observations of these potential prey species by study area using average number of individuals observed per surveyed route.