Virtual population analysis (cohort analysis) was used to reconstruct past
dynamics of a harvested population of martens Martes americana in the Brace
bridge District of southern Ontario. Harvests in the Bracebridge District w
ere managed using a quota system set by regional authorities. Quotas change
d from year to year, apparently on the basis of past trapping success and v
ariation in the proportion of young-of-the-year among harvested animals. Th
e proportion of young in the harvest was a sensitive indicator of the annua
l rate of increase, whereas trapping success tended to be linked most stron
gly, in inverse fashion, with marten harvesting quotas. The proportion of m
artens harvested each year was constant, averaging 34%, despite 3-fold vari
ation in marten abundance. This proportion was very close to the maximum su
stainable yield (36%) for the population, suggesting that the management po
licy in the administrative unit was effective in the past in sustaining the
source population as well as yielding high trapping returns. Monte Carlo s
imulation showed that proportionate harvesting, such as the policy in the B
racebridge District during 1972-1991, should be considerably less likely to
lead to overharvesting than a constant quota policy, particularly at high
average yields.