Northern Prairie Wildlife Research Center
Geographic Variations.—Numbers of ducks found at individual fox dens ranged from 0 to 67. The 67 ducks were found at a den excavated on J. Clark Salyer National Wildlife Refuge, North Dakota, during mid-June 1970; all were dabbling ducks and 59 (including 42 blue-winged teals) were hens, 4 were drakes, and 4 were of unidentified sex. Another den visited on the refuge at that time had remains of 36 ducks; remains of 42 ducks were taken from still another den on the refuge the previous year. These data demonstrate the capabilities of fox families to obtain and use large numbers of nesting ducks.
Several factors affected both the amount of food brought to individual dens and the amount and location of prey remains at dens. Foremost was the duration of den occupancy, number of pups, and age of pups. Red fox families in agricultural areas generally use several dens to rear a litter (Johnson and Sargeant 1977:10-11) and often, especially during the latter half of the denning season, split littermates among 2 or more dens (A. B. Sargeant and S. H. Allen, unpubl. data). Human and agricultural disturbances that occur on most lands in the midcontinent area cause foxes to move pups more often than would occur otherwise. Hence, both the number of rearing dens used by individual fox families and the amount of time pups spend at individual dens vary. Fox litters range from 1 to 12 pups but average about 4-6 pups (Storm et al. 1976:22, Pils and Martin 1978:13) and pups from 2 litters may communally occupy the same den (Sheldon 1950, Storm et al. 1976:23, Pils and Martin 1978:12, Tullar and Berchielli 1980). Age of pups at the time of our den visits varied because of asynchrony in fox whelping dates (Storm et al. 1976:19, Sargeant et al. 1981) and differences in dates observers visited dens (Fig. 5). Because of these variables, meaningful comparisons of numbers of ducks found at dens had to be based on averages from relatively large numbers of dens with similar pup age and den visit characteristics.
Fox predation rate indices (average number of ducks/den from surface food remains) for each nesting ground sample area (except Manitoba where 8 dens were visited) ranged from 0.01 duck/den in Iowa to 1.80 ducks/den in east river North Dakota (Table 7). The index values are directly comparable although subject to bias from latitude caused asynchrony of the fox denning and duck nesting seasons. This bias probably resulted in underestimation of predation on nesting ducks in Iowa and South Dakota relative to that in North Dakota. We believe differences in predation rate indices from areas with similar mean fox whelping dates reflected differences in magnitude of predation on ducks by individual fox families in each area. Hence, the average fox family in east river North Dakota took about 7 times more ducks than the average fox family in west river North Dakota, whereas the average fox family in east river South Dakota took about 1.1 times more ducks than the average family in west river South Dakota.
It is likely that environmental factors contributed to the relatively high fox predation rates on ducks in eastern North Dakota. For example 78% of the land on our 2 Drift Plain detailed study townships was cultivated annually (Sergeant et al. 1975), and in other nearby areas annual tillage was even greater (Higgins 1977). Most cultivated lands are tilled in late summer or fall, left bare through winter, and tilled again in spring for planting. Populations of many prey species in such areas are greatly reduced. Because many wetland basins remain, large numbers of ducks and other migratory marsh birds occupy this habitat each spring. Hens of most upland nesting ducks in the area select limited remaining untilled habitats for nesting (Higgins 1977). Hence, as agriculture intensifies, nesting ducks (and duck eggs) gain increasing prominence as one of the most abundant, vulnerable, and desirable prey types available to foxes. Although this principle applies to all areas, the effect is likely greatest in eastern North Dakota or further north where climatic factors more severely limit agriculture and prey diversity.
The predation rate indices were closely correlated with the percentage of dens with surface ducks (r = 0.934, P < 0.05). Thus, each increase in the index was accompanied by an increase in the proportion of foxes involved in the predation. The fact that duck remains were found above ground at 63% of the east river North Dakota dens visited (Table 7) even though some dens had no food, were used for short periods, or were in areas with few ducks indicated that nearly all fox families exposed to nesting ducks in that area preyed on ducks.
Influence of Duck Availability and Environment.—Data from the S-county intensive study area in eastern North Dakota show a strong correlation between the duck availability index and both the fox predation rate index (r = 0.973, P < 0.0001) (Fig. 7) and the percentage of dens with surface ducks (r = 0.938, P < 0.0002) (Fig. 8). From these data it is evident that both the number of ducks taken by fox families and the proportion of fox families engaged in predation on ducks increased as the availability of nesting ducks increased. The predation rate index, however, increased exponentially relative to increases in the duck availability index (Fig. 7). If our duck availability index accurately reflects availability of dabbling ducks to foxes, then, within the range of duck densities present in the intensive study area, the probability of an individual duck being captured by foxes was highest when nesting ducks were most available.
Fig. 7. Relation between number of surface ducks found at dens (predation rate) and an index of availability of ducks within 3.2 km of dens (r=0.973, P< 0.0001). Fit based on logarithmic transformation of predation rates; data were weighted for numbers of dens (in parentheses) visited in each duck abundance category. See Methods for calculation of duck availability index.
Fig. 8. Relation between percentage of red fox rearing dens with duck remains found in and around den entrances (surface ducks) and an index of duck availability within 3.2 km of dens (r=0.938, P < 0.0002). Correlation weighted for numbers of dens (in parentheses) in each duck abundance category.
Predation rate indices for all dabbling ducks combined averaged slightly lower but were more variable on the Drift Plain than on the Missouri Coteau (Table 8) both the highest and lowest predation rate indices occurred on the Drift Plain. Although separate May pond and dabbling duck population indices were not available for each physiographic region, the differences in predation rate indices were likely caused by differences in wetlands. Wetland water levels, and hence duck densities, were less stable on the Drift Plain than on the Missouri Coteau.
Species Vulnerability.—Differences in responses by duck species to wetland conditions are clearly reflected in comparisons of fox predation rate indices in the 3-county intensive study area to dabbling duck densities in Stratum 46 (Table 8). Comparison of coefficients of variation (CV) for changes in populations of the principal dabbling duck species showed that the mallard population was relatively stable during the 5-year period and varied less than populations of the other species. Similarly, CV's for predation rates on mallards were more constant than those for other species. However, the predation rate CV for mallards varied nearly twice as much as the population CV, and the predation rate index was much more strongly correlated with the May pond index than was the population index. Thus, though size of the mallard breeding population did not vary greatly among years, the effects of variable nesting conditions on the predation rate indices were clearly evident.
The northern pintail population varied more than populations of the other principal species, and the variation was strongly correlated with changes in the May pond index (Table 8). During wet springs, as in 1969, large numbers of northern pintails nest in the northern plains, but during dry springs, as in 1971 and 1973, many overfly the region (Calverley and Boag 1977, Derksen and Eldridge 1980) and consequently are unavailable to midcontinent foxes. Therefore, northern pintail densities reflected nesting conditions and northern pintail sensitivity to drought. The fox predation rate index for northern pintails varied greatly among years and was strongly correlated with both the May pond and northern pintail population indices.
Annual predation rate indices for northern shovelers were the most variable of the principal species, and differences between CV's for the population and predation rate indices were the greatest recorded. Both indices were closely correlated with the May pond index although the population changed much slower than did the predation rate. For example, from 1969 (wettest year) to 1973 (driest year) the population declined 58% whereas the predation rate index declined 86%. This would have occurred if drought affected nesting effort more than it affected population size and if predation occurred primarily at nests. Although all 5 species were affected in this manner, the effects were most pronounced with the northern shoveler. This likely reflected the northern shoveler's greater dependence on drought sensitive aquatic invertebrates, especially crustacea, for food (Swanson et al. 1979).
Both the population and predation rate indices for blue-winged teals fluctuated greatly, but neither index was closely correlated with the May pond index. From 1969 to 1970 the population and predation rate indices declined 48 and 57%, respectively, though the May pond index declined only 11%. From 1970 to 1971 the population index increased 61% and the predation rate index remained almost unchanged, but the May pond index declined 43%. There was, however, a reasonably close correlation between the predation rate and population indices.
The gadwall population was relatively stable, and annual population changes were closely correlated with the May pond index. The predation rate index for gadwalls was also relatively stable. The predation rate index, however, was imprecisely linked to both the May pond and population indices. Nevertheless, the effects of the 2 drought years, 1971 and 1973, on the predation rate index were evident. The relative insensitivity of the predation rate index to changes in May pond numbers and to population size probably reflected the late but apparently persistent nesting habits of this species.
Relative vulnerability indices for each dabbling duck species to foxes were determined by dividing the average predation rate index for each species by the respective population index and multiplying the results by 1,000 (Table 9). The results are density independent and directly comparable among species and years, but relate only to ducks living in areas frequented by red foxes. During most study years nearly all ducks in the 3-county intensive study area lived in such areas.
Among the 5 principal dabbling duck species, northern pintails were most vulnerable to foxes, and gadwalls and blue-winged teals were least vulnerable. Mallards and northern shovelers were slightly less vulnerable than northern pintails. Dabbling ducks were most vulnerable to foxes during wet springs when nesting effort was strongest. Vulnerability indices during the 3 wettest springs (1969, 1970, and 1972) averaged 1.4 times greater than those of the 2 drought springs (1971 and 1973). In general, early nesting species were much more vulnerable to fox predation than late nesting species.