This investigation was concerned primarily with the estimation of fish, rabbit, and squirrel populations on various areas on the Virginia Polytechnic Institute College Farms at Blacksburg, Virginia. Particular attention was given to the capture-recapture method and other related methods of population estimation, and the evaluation, wherever feasible, of these methods by determining as accurately as possible the animal populations as determined by intensive trapping and pond drainage.

1. A total of 1,181 fish were marked during a period of 12 days. Fifty-nine rabbits and 200 squirrels were marked during a period of approximately 7,890 trap-nights.

2. Both the Krumholz method and the Schumacher and Eschmeyer method give a reasonably accurate and certainly usable estimate of fish pond populations.

3. The Krumholz method, as used in this study for a comparison with the Lincoln Index estimates of mammal populations, is not suited for use in calculating estimates when the number of captures and recaptures has been low for an area.

4. The number of animals seen per hour per acre varies greatly with the habitat, season, and the time of day; however, this technique could be a useful index to abundance when such records have been kept for a considerable period of time for the same area.

5. The efficiency and accuracy of population estimates based on short precensus and census periods are not constant because the number of individual captured and recaptured varies too greatly with the species and its environment. Table 11 is a summary of the population estimates of squirrels and rabbits on the three wooded study areas.

[see document for table]

1. The Lincoln Index method of population estimation, when applied to data obtained through four or more months of continuous squirrel trapping yielded an estimate that varied from 1.31 to 1.72 times the number of animals tagged on the area. The average variation was 1.537 with a standard deviation of 0.278. The Lincoln Index estimates of the rabbit populations varied from 1.78 to 1.96 times the number of animals tagged on the area. The average variation was 1.868 with a standard deviation of 0.0925. When this method of censusing is used, the variation should be kept in mind, and any population figure based on the Lincoln Index should also contain a statement as to whether or not the figure was adjusted.

2. The use of kill data from a squirrel harvest substituted for a census period of trapping yielded an estimate that was probably more inaccurate than that obtained by any other method. However, it is entirely possible that an insufficient number of animals had been marked prior to the harvest in order to obtain an acceptable estimate by this method.

3. The ratio of marked squirrels per leaf nest was fairly constant. If the four marked squirrels that were removed by hunting on Study Area III were not included in the total number of tagged animals on that area, the marked squirrel:leaf nest ratio would be 13:1 on Study Areas II and III and 13.1:1 on Study Area IV. The leaf nest counts were made in mid-April, after the nest had been subjected to the snow, sleet, and wind of winter. However, a leaf nest count made in mid-December on one study area gave the same number as the later count. The only disadvantage to using this method is that it should be used at a tie of year when there are no leaves on the trees in order to insure a complete count.