may be linked to the phenology of a smaller range of tree
species, and this species may limit reproductive activities
to relatively short periods of time when preferred foods
are available. Specifically, gestation and lactation, when
mothers experience high energy demands, may be timed
to match the first dry season (February to mid-April)
when photosynthetic activity (i.e. leaf growth) peaks in
the tropics to coincide with high solar radiation and low
insect herbivory (Windsor 1990; Schaik, Terborgh &
Wright 1993). Rearing young during the dry season may
also provide thermoregulatory advantages. Similarly,
maned sloths appear to time the production of their
young with the transition from the wet to dry season, presumably
to match energetically demanding periods with
favourable environmental conditions (Dias et al. 2009).
Our results suggest that annual survival is high in adult
males of both species (>090). While we were unable to
estimate annual survival rates for adult females using
mark–recapture models, colour-marked female B. variegatus
were observed surviving 23 of 26 survival intervals.
This observation yields a na€ıve, and minimum, annual
survival estimate of 0906 (SE = 0063), suggesting that
survival is high for both sexes in adult B. variegatus. High
adult survival rates are perhaps not surprising given that
two of the sloths’ primary historic predators, jaguars Panthera
onca and harpy eagles Harpia harpyja, are no longer
present in the farm or the surrounding landscape. Nevertheless,
predation, primarily by coyotes Canis latrans and
domestic dogs Canis lupus familiaris, was the leading
cause of mortality for adults (Table 2). Whether predation
by coyotes and dogs has replaced predation by historic
predators is unknown, but it seems likely that predation
on adults is currently the most important factor limiting
this population given the sensitivity of population growth
in long-lived species, with low fecundity is most sensitive
to adult survival. Montgomery & Sunquist (1978) suggested
that sloth populations on Barro Colorado Island,
Panama, were limited by bottom-up processes because
dead subadults and adults were often found during the
rainy season, presumably having succumbed through
nutritional stress. However, only two of 16 adults in our
sample died from nutritional stress or disease, suggesting
that the agro-ecosystem may provide adequate foraging
habitat for the survival of adult sloths. Nevertheless, we
cannot exclude the possibility that nutritional stress could
have contributed to predation on adults in our study.
Information on juvenile and subadult mortality factors
was sparse as we did not radiomark these age classes, but
nutritional stress or disease was implicated in all six
observed mortalities (Table 2).
ROLE OF CACAO IN THE CONSERVATION OF
BIODIVERSITY
The development of shade-grown agriculture systems presents
both opportunities and challenges for biodiversity
protection. Agro-ecosystems generally harbour more
biodiversity than crops grown in monocultures and can
retain a significant component of the biodiversity that
occurs in adjacent, ecologically similar reserves. However,
cacao farms without necessary habitat elements can be
depauperate of vertebrate biodiversity, particularly arboreal
mammals such as sloths and primates (Rolim & Chiarello
2004). Moreover, the continued clearing of original
forests for agriculture, even when grown underneath an
overstorey of native trees, continues to erode tropical biodiversity.
In such cases, remnant animal communities are
typically biased toward generalist species at the expense of
species with specialized habitat or foraging requirements
(Harvey, Gonzalez & Somarriba 2006). Nevertheless, our
results indicate that some species with specialized habitat
requirements, not to mention limited dispersal abilities,
such as C. hoffmanni, can be self-sustaining within an agroecosystem
of modest size (c. 4 km2) given that required
habitat elements are retained. Similarly, maned sloths in
the Brazilian Atlantic forest have been found to use biologically
rich cacao farms for foraging and can produce offspring
in this habitat, although they may require adjacent
tracts of intact forests to persist (Cassano, Kierulff & Chiarello
2011). In contrast to C. hoffmanni, our results indicated
that B. variegatus did not appear to be self-sustaining
in the absence of immigration, perhaps due to its more
specialized dietary needs and sedentary behaviour. The
differences in demography between these two ecologically
similar species emphasizes that even subtle variation in lifehistory
strategies can interact with environmental conditions
present in an agro-ecosystem and lead to different
levels of vulnerabilities among species. Clearly, features
within agricultural systems required to maintain sensitive
tropical species will vary according to the life-history strategy
of the species in question. More broadly, the value of
human-modified landscapes for biodiversi