The only other question that might arise from a comparison of the age structure of our wolf-killed deer with that of the hunter-killed deer concerns the area from which each sample was taken. Fifty of our wolf-kills came from a region almost inaccessible to hunters ([fig. 1]). However, the other 92 came from the same general area as the hunter-kills. Nevertheless, there was no statistically significant difference in age structure between the wolf-kills from the wilderness versus those from the hunted area ([table 1]). This fact also suggests that the human hunting in the area is relatively light and has little effect on the age structure of the deer population in the area.
Wolves may also be taking a disproportionately high number of fawns, although our data do not show this. Nevertheless, there may be a bias against fawns in our method. It is not unusual to discover the remains of a wolf-killed deer so completely eaten that there is no indication left of the animal's age. Because fawns often are only about half the size of adult deer, and their skeletons have not yet completely ossified, the chances are better that fawns will be more completely eaten. Pimlott et al. (1969) also recognized this possible bias, although their data did indicate that wolves were killing a higher percentage of fawns than occurred in the population.
Our study does support the other conclusion of Pimlott et al. (1969), based on a study of 331 kills, that wolf predation on deer during winter shows a definite selection for older animals. It does not agree with the tentative conclusion of Stenlund (1955) that wolves in the Superior National Forest do not prey disproportionately on old deer. However, Stenlund's conclusion was based on 36 kills and on the assumption that only deer at least 7 years old were "old." Deer 5 years old and older composed 33 percent of Stenlund's sample, a figure considerably higher than the 10 percent in these age classes in our hunter-kill sample ([table 4]). Thus Stenlund's data do not contradict our conclusion.
The age of 5 years seems to be the beginning of the period of vulnerability for adult deer. Although 5 years might not seem especially old, there are two aspects of significance concerning deer of this age and older. First, they are in the second half of the life span for most members of the species, and their alertness and ability to bolt quickly away might be expected to decline. It is of interest in this regard that Klein and Olson (1960, p. 87) believed 5 years of age to be "the upper limit of physiological efficiency" of black-tailed deer (Odocoileus hemionus) in Alaska. Second, up to the age of at least 4½ years, and perhaps beyond, the apparent weight-load-on-track of deer increases with age (Kelsall 1969). Thus older deer would sink farther into the snow than younger ones, and their escape might be slowed and hindered more. For further discussion of the effect of snow on the vulnerability of deer, see Mech et al. ([p. 51]).
Sex Ratio
Statistical tests comparing a number of subsamples of both wolf-killed deer and hunter-killed deer showed a series of significantly different sex ratios ([tables 1])-[3]. The ratio of males to females in the fawn cohort of the hunter-kill, which is probably the most representative of the actual fawn sex ratio, was even ([table 2]). With wolf-kills, however, a significantly higher percentage of females was taken in the fawn subsample (59 percent) than in the adult subsample (46 percent). These results compare favorably with those of Stenlund (1955), who found that from 1948 to 1953 in the same area as the present study 68 percent of 19 sexable fawn wolf-kills were females and 44 percent of 63 sexable adult wolf-kills were females.
If the sex ratio of fawns began even, and more females than males were killed by wolves, then a higher proportion of males would be left in the adult population, unless some other mortality factor kills more male fawns. Thus it is not surprising that in the wilderness area, where little or no hunting is done, the sex ratio of wolf-kills in the adult cohort is significantly heavy toward males (71 percent: 29 percent). This was also true of the wolf-kills in Algonquin Provincial Park, where males made up 57 percent of the total sexable wolf-kill (Pimlott et al. 1969). The latter figure may even have been higher if calculated for adults alone, for a preponderance of female fawns in the Algonquin Park data (such as occurred in our and Stenlund's samples) would tend to obscure the preponderance of males in the adult sample.
The adult subsample of hunter-kills also contained a higher percentage of males (66 percent : 34 percent). Although this might also reflect the influence of wolf predation on female fawns, it probably is more a result of the greater movement of bucks during the hunting season, which overlaps with the rutting season. Even the sex ratio of adult deer killed in wolf-free areas shows a preponderance of males (Erickson et al. 1961).
However, it appears that the higher harvest of bucks by human hunters does markedly affect the sex ratio of the deer population in the hunted area, for the wolf-kill of adults in that area contained a significantly higher percentage of does (56 percent) than did the wolf-kill of adults in the wilderness area (29 percent).
Evidently the hunter harvest is not heavy enough to affect the age structure of the deer population to any marked degree, for no significant difference in age structure was found between the wolf-kill in the hunted area and that in the wilderness area ([table 1]). This does not conflict with the conclusion that hunting affects the sex ratio of the deer herd, because it would take much less to influence a population characteristic having two classes (sex) than one having 14 (age).