Many years of continuous collections of deer kill data on state game management areas have largely been poorly utilized by deer biologists. Recent computer modeling developments have provided a vehicle for the examination of these collections of data. This thesis represents an attempt to demonstrate the usefulness of minimum population reconstruction in defining those populations that best contrast different harvest strategies. Buck-only deer hunting is compared to either-sex hunting on two similar areas.

To determine if variables such as weather are acting to alter hunter participation and harvest in southeastern deer herds, 25 independent variables were tested. Correlation coefficients and maximum R² improvements provided the statistical basis for drawing conclusions concerning the relative "strength" of these variables. Higher than normal temperatures before and during the hunt were found to be positive contributors to harvest, as was precipitation before the hunt. Hunt days that followed nights of full moon also demonstrated a positive, but generally weak association with kill.

The size of the deer population on an area, and the number of hunters that participate in a hunt were found to be the most powerful contributors to overall deer harvest. An evaluation of vegetative density indicated that buck hunting is most successful on areas that are more open, whereas doe harvest was related more to a denser vegetative classification.

Buck-only hunting, when promoted over long periods of time, was found to drastically distort male age structure in favor of young age classes. Heavy either-sex harvests, however, provided not only a greater harvest of deer, but sex and age classes remained relatively stable in the herd. Continual monitoring of deer kill and careful analysis of reliable data through modeling is a most efficient method for not only detecting potential problems in herd welfare, but also in the establishment of forthcoming hunt regulations.