From Tables 1 and 2, (two) general conclusions can
be drawn. First, the relationship between basal GCs
and rank is highly variable. Variables that might
affect the strength and direction of this relationship
include the social system, domestication, phylogeny,
the behavioral traits that are associated with high and
low rank, and whether the study was conducted in the
wild or in captivity (Box 1). Second, although the data
are still sparse for the purposes of comparative
analysis, it is uncommon for subordinates of species
that live in permanent groups to experience
chronically elevated GCs (Fig. 1). Only one study of a
social species in the wild has found elevated basal GCs
in subordinates23, whereas five field studies have
found higher levels in dominants. These five cases are
phylogenetically well distributed, including a bird
(Florida scrub jay Aphelocoma coerulescens)22, a
rodent (alpine marmot Marmota marmota)24, a
primate (ring tailed lemur Lemur catta)25 and two
carnivores (dwarf mongoose Helogale parvula and
African wild dog Lycaon pictus)15,26.
The interpretation of differences among ranks in
GC levels is not entirely clear-cut. Apriori, the class of
individuals with higher basal GCs would generally be
considered more socially stressed. However, data from
captive common marmosets27,28 show that GC levels
do not increase upon attaining dominance: rather, GC
levels decrease among subordinates, in parallel with
estrogen levels. Field studies have not generally tested
whether the differences between ranks arise through
increased GC secretion in one class or reduced secretion
in the other class, although limited data from African
wild dogs and dwarf mongooses show that GC levels
increase upon attaining dominance. Considering the
impact of GCs on fitness, the difference in mechanisms
might be moot, if the social class with higher basal GC
levels is exposed to more of their harmful effects.
Implications and future research
If dominant animals do generally have elevated GC
levels among cooperative breeders, this would
represent a major shift in our view of social stress.
We must consider the stress of domination, as well as
the stress of subordination, and how these stressors
might affect social evolution. There are several
interesting questions to address. First, what
behavioral differences among species (Box 2)
predict whether stress will fall more heavily on
dominants or on subordinates (or perhaps on mid-
ranking animals, in some cases)? This is already an
active field of study25,29, and two possibilities are that
dominants have elevated GCs when a hierarchy is
unstable2 or in species in which dominants fight more
often than do subordinates26.
Second, what are the nonGC-mediated
mechanisms by which rank affects sexual behavior
and sex steroid levels? It is already clear that
reproductive suppression is not mediated by
chronic GC elevation for most of the cooperative
breeders that have been studied (Table 1, species
listed as ‘high skew’). However, many aspects of
reproductive behavior and sex-steroid secretion
are depressed in subordinates of these
species14,15,22,30. Considering the broad range of
pathologies that chronic GC elevations can cause2,4,
it is perhaps not surprising that reproductive
suppression is rarely GC mediated in species for
which reproductive suppression is a normal
feature of social organization. With such a social
system, stress-mediated mechanisms of
suppression would be evolutionarily invasible by a
direct mechanism of suppression that did not
involve GC elevation, and thus avoided the costs of
immune suppression and other pathologies. If this
line of reasoning is correct, stress-mediated
TRENDS in Ecology & Evolution
Vol.16 No.9 September 2001
http://tree.trends.com
495
Opinion
Although the terminology is cumbersome, there is an important
distinction between short-term (‘acute’) and long-term (‘chronic’)
stress responses. When exposed to a stressor, a typical individual
will mount a pronounced glucocorticoid (GC) response within
two to ten minutes. During this acute response, circulating GC
levels typically rise several fold, but drop back to basal levels
minutes (or perhaps a few hours) after the stressor is removed.
This short-term response is generally considered adaptive,
because it shifts energy and resources away from physiological
processes that can be curtailed briefly without harm, putting
these resources instead toward resolving the stressful condition.
By contrast, if an elevation of GCs persists for days or weeks,
then many of the short-term benefits become long-term pathologies,
including immune suppression, reproductive suppression, gastric
ulcers and muscle wasting. Sapolskyaprovides an excellent review
of the short-term benefits and long-term harms that elevations of
GCs usually provoke, and their mechanisms. Because of the
dichotomy in the effects of acute and chronic GC responses, studies
of social stress generally focus on chronic elevations of GCs
(elevated basal values). Some studies go further by also testing
for effects of social status on acute GC responsesb–d. When basal
GC concentrations are elevated, a common consequence is a
weakening of the acute GC responsed, probably through changes in
negative feedback mechanisms in the hypothalamic–adrenal–
pituitary axis. This is an endocrine vicious circle: a chronically
high baseline can provoke pathologies, and a weak acute
response is ineffective at handling short-term stressorse.
References
a Sapolsky, R.M. (1992) Neuroendocrinology of the stress response. In
Behavioral Endocrinology (Becker, J.B. et al., eds), pp. 287–324,
Massachusetts Institute of Technology Press
b Schoech, S. et al. (1991) Reproductive endocrinology and mechanisms of
breeding inhibition in cooperatively breeding Florida scrub jays
(Aphelocoma c. coerulescens). Condor93, 354–364
c Creel, S. et al. (1996) Social stress and dominance. Nature379, 212
d McEwen, B.S. and Schmenk, H.M. (1994) The Hostage Brain, Rockefeller
University Press
e Virgin, C.E. and Sapolsky, R.M. (1997) Styles of male social behavior and their
endocrine correlates among low-ranking baboons. Am. J. Primatol. 42, 25–39
Box 2.Chronic and acute stress