Northern Prairie Wildlife Research Center
We eartagged, weighed, sexed, and noted reproductive condition of all captured squirrels. We implanted a 20-g, 30-MHz radio transmitter into the abdominal cavity of selected ground squirrels (Eagle et al. 1984). Transmitters had a 120-day life expectancy. We released radio-equipped ground squirrels at the site of capture within 24 hours after capture.
We obtained most radio locations by triangulation from a vehicle with a roof-mounted antenna. A hand-held receiver and antenna (Cedar Creek Bioelectronics, Bethel, Minn.) were used to locate burrows. We verified locations by visiting occupied burrows after ground squirrels entered them at night. Ground squirrels were recorded as active if their radio signal modulated during a 30-second period while each location was being taken and as inactive if no modulation occurred. Activity recorders (Sunquist and Montgomery 1973) were placed near selected burrows to determine if the ground squirrels left their burrows at night and to verify our interpretation of signal modulations for activity data.
Each field season consisted of an intensive and nonintensive radio-tracking period. The intensive period began in late April and lasted through midAugust and was divided into 8 consecutive 2-week sample units. During these periods we attempted to locate all radio-equipped ground squirrels at 45-150 minute intervals during 6 daytime tracking sessions spaced throughout each sample unit, and once a day on 4 additional days during each unit. The 4-15-hour tracking sessions rotated among 3 overlapping time periods: sunrise-sunset, sunrise-1400 hours, and 1000 hours-sunset. Early morning tracking began and evening tracking ended by locating the burrows in which the radio-equipped ground squirrels were resting. The nonintensive tracking period began in mid-August and ended when all radio-equipped ground squirrels had entered hibernation or their transmitters failed. During this period we located ground squirrels at least once a week, but did not attempt to follow their daily movements.
We used the computer program TELEM (Koeln 1980) to determine home-range sizes by the minimum convex polygon method (Mohr 1947), geometric centers of activity (Hayne 1949), and mean activity radii (Dice and Clark 1953). In addition, we calculated total distance traveled daily for each ground squirrel for days when the animal was tracked from sunrise to sunset and ≥6 radio locations were obtained. Total distance traveled daily was the sum of the straight-line distances between consecutive radio locations. We calculated biweekly home-range sizes for each ground squirrel for each 2-week sample unit in which it was located ≥14 times. Annual home-range sizes were calculated for each ground squirrel located ≥70 times and on ≥20 days distributed throughout the tracking period. We used Student's t-test to determine intersexual differences in home-range size and movement parameters, and analysis of variance (ANOVA) for between-year differences.
We described vegetation height and density by visual obstruction measurements (Robel et al. 1970) taken systematically in each field 3 times annually at 7-week intervals beginning in mid-April. Litter depth was measured at each sampling site during mid-April each year. Differences in vegetation parameters were evaluated by ANOVA. We combined ground squirrel radio locations from both years to analyze differences in use of habitat blocks on the WPA. We assumed that the radio-equipped individuals represented the population. We plotted ground squirrel locations on gridded maps and tallied all 0.5- × 0.5-m grid cells with ≥1 radio location to portray habitat use. We calculated electivity indexes (Ivlev 1961:45) to express habitat selection. Electivity indexes range from -1 to +1 indicating habitat use less or greater than habitat availability, respectively.