We studied the effects of habitat fragmentation on four species of North Am
erican tanagers (Piranga spp.) nt 1107 study sites in 1995. Volunteer resea
rchers used the standardized protocol of Project Tanager to select sites, c
ensus for breeding tanagers, and measure habitat and landscape features on
a continental scale In a principal components analysis, we used measures of
forest patch size, distance to nearest other forest patches (isolation), p
ercent forest cover and length of linear edge in a 1000-ha landscape, and e
levation. Our results showed that three independent habitat gradients affec
ted tanager occurence: (1) overall amount of fragmentation; (2) forest conf
iguration (the amount Of edge ill a landscape); and (3) elevation. Multiple
logistic regression indicated significant variation among the tanager spec
ies in how habitat gradients affected tanager occurence, presence of parasi
tic Brown-headed Cowbirds (Molothrus ater), and various nest predators. Sca
rlet, Western, and Summer Tanagers (P. olivacea, P. ludoviciana, and P. rub
ra, respectively) responded similarly to the fragmentation gradient with th
e probability of finding breeding tanagers dropping below 0.50 in highly fr
agmented sites. Response to the other habitat gradients was more variable a
nd effects on cowbirds and predators were usually opposite those of breedin
g tanagers (i.e. they were more common nt highly fragmented sites). Respons
e To habitat gradients also varied significantly among four regions within
the range of Scarlet Tanagers. The strongest response to fragmentation, by
breeding tanagers as well as by cowbirds and nest predators, was in the lar
gely deforested Midwest and Atlantic Coast regions. In the heavily forested
Northeast region, fragmentation had a significant Dirt less severe effect
on breeding tanagers. Effects were not significant in the Northern Forest r
egion, where tanager abundance was low. Results of our continent-wide study
suggest that (1) forest fragmentation can be described similarly across mo
st regions of North American: (2) three widespread tanager species exhibit
clear and similar negative response to habitat fragmentation; (3) local sen
sitivity to fragmentation varies geographically and may be lower In regions
with greater overall forest cover; and (4) results from single-species or
local studies cannot Dc extrapolated to other species or regions. Conservat
ion strategies to reverse the declines of forest bird populations will requ
ire knowledge of habitat requirements across the entire ranges of widesprea
d species as well as how demographic and landscape factors interact to crea
te population sources and sinks.