If is necessary to assess whether the management of protected areas is achieving the objectives set for them. In particular, changes and trends in wildlife populations should be documented. We compared the 1990 abundance of mammals in two Costa Rican protected areas Corcovado National Park (CNP) and Golfo Dulce Forest Reserve (GDFR), with similar environmental characteristics bat different hunting restrictions and levels of protection. We also monitored the abundance of mammals in CNP over a 4-year period. We also devised an inexpensive method of monitoring tropical rainforest mammal populations in a timely and efficient manner. The method is based on the use of mammal track records and arboreal mammal sightings and requires little effort. With this method 20 mammal species were recorded at CNP and 15 at GDFR. Species were consistently less abundant in GDFR than in CNP, principally those species preferred by hunters. Species were grouped according to whether or not they are used as food and whether they are locally or globally threatened or not threatened. The abundance of all these groups of species in GDFR was 6-28% the abundance in CNP. During 1990-1994, overall mammal abundance in CNP remained relatively stable, but there was considerable variability among species. As a group, both species used for food and globally threatened species declined in abundance during that period in CNP. The abundance of all groups of species declined from 1992 to 1994, coinciding with a reduction in hunting vigilance at CNP. Evidence suggests that the main factor differentiating the abundance of mammals in the two protected areas, and at CNP during the study period, was the level of hunting. Golfo Dulce Forest Reserve seems to be achieving only partial success in protecting wildlife, whereas Corcovado National Park seems to be considerably more effective, although not entirely successful.