[Illustration: _Figure 13.--Relative rates of predation on deer of various ages, based on comparisons of the ages of wolf-killed deer with those of a theoretical population (dashed line) and those of the hunter-killed population. See figure 7._]
These figures assume added significance when compared with a sample of deer killed by hunters in the same general area (fig. 1). Only 10 percent of the hunter-killed deer were 5-1/2 years old or older, and the percent killed in each year class dropped off suddenly from 13 percent aged 4-1/2 to 6 percent aged 5-1/2. If the age structure of the hunter-kill sample is reasonably representative of the age structure of the population at large, the wolf-kill data show that wolf predation in our study area during winter has a definite selective effect on the deer population.
There is no direct way of knowing that the age structure of the hunter-killed deer represents the age structure of the deer population at large. However, sampling hunter-kills is the most practical means available for gaining an index to the age structure of the existing herd. Further, there are three indirect pieces of evidence indicating that the hunter-kill sample represents the actual age structure of the population, just as Maguire and Severinghaus (1954) found in New York. First, our sample has the basic theoretical form expected of a stable deer herd; i.e., the youngest year class contained the most members, and each older cohort included fewer (fig. 7). Second, the age structure of our sample has the same form as most other deer age structures from widely diverse areas, (Ontario, Pimlott _et al._ 1969; southern Minnesota, Erickson _et al._ 1961; Massachusetts, Shaw 1951). Third, there is no reason to believe that in our area rifle hunting is especially selective for any particular age classes. In talking with large numbers of hunters, we have learned that most shoot at any and all deer they happen to see.
Even if the age structure of the hunter-kill sample did not approximate that of the actual herd, the comparison of the wolf-kill with the theoretical population dictates the same conclusion: the rate of kill of older deer by wolves was several times greater than that of younger deer, excluding fawns (fig. 13). In any case, if the actual deer population in our study area had an age structure similar to that of our sample of wolf-kills (which would be the only age structure that would contradict our conclusion), its numbers would be declining by orders of magnitude each year, and there would now be only a remnant population. Such obviously is not the case.