In this paper, we address the question of what proportion of biodiversity is represented within protected areas. We assessed the effectiveness of different protected area types at multiple scales in representing primate biodiversity in the Brazilian Legal Amazon. We used point locality data and distribution data for primate species within 1°, 0.5°, and 0.25° spatial resolution grids, and computed the area of reserves within each cell. Four different approaches were used - no reserves (A), exclusively strict use reserves (B), strict and sustainable use reserves (C), and strict and sustainable use reserves and indigenous lands (D). We used the complementarity concept to select reserve networks. The proportions of cells that were classified as reserves at a grid resolution of 1° were 37%, 64%, and 88% for approaches B, C and D, respectively. Our comparison of these approaches clearly showed the effect of an increase in area on species representation. Representation was consistently higher at coarser resolutions, indicating the effect of grain size. The high number of irreplaceable cells for selected networks identified based on approach A could be attributed to the use of point locality occurrence data. Although the limited number of point occurrences for some species may have been due to a Wallacean shortfall, in some cases it may also be the result of an actual restricted geographic distribution. The existing reserve system cannot be ignored, as it has an established structure, legal protection status, and societal recognition, and undoubtedly represents important elements of biodiversity. However, we found that strict use reserves (which are exclusively dedicated to biodiversity conservation) did not effectively represent primate species. This finding may be related to historical criteria for selecting reserves based on political, economic, or social motives.