A major obstacle to conserving tropical biodiversity is the lack of in
formation as to where efforts should be concentrated, One potential so
lution is to focus on readily assessed indicator groups, whose distrib
ution predicts the overall importance of the biodiversity of candidate
areas(1,2). Here we test this idea, using the most extensive data set
on patterns of diversity assembled so far for any part of the tropics
. As in studies of temperate regions(2-8) we found little spatial cong
ruence in the species richness bf woody plants, large moths, butterfli
es, birds and small mammals across 50 Ugandan forests. Despite this la
ck of congruence, sets of priority forests selected using data on sing
le taxa only often captured species richness in other groups with the
same efficiency as using information on all taxa at once. This is beca
use efficient conservation networks incorporate not only species-rich
sites, but also those whose biotas best complement those of other area
sg(9-11). In Uganda, different taxa exhibit similar biogeography, so p
riority forests for one taxon collectively represent the important for
est types for other taxa as well. Our results highlight the need, when
evaluating potential indicators for reserve selection, to consider cr
oss-taxon congruence in complementarity as well as species richness.