liers 1988). For example, the benzodiazepine-GABA system
initiates brief analgesia in white-footed mice (Peromyscus
leucopus) after a brief exposure to an out-of-view short-
tailed weasel (Mustela erminea), but a longer predator ex-
posure induces a more tonic, endorphin-mediated analge-
sia. Activation of analgesia may allow continued defensive
action when an individual has sustained a painful injury,
either from a predator or a conspecific (Kavaliers 1988;
Rodgers and Randall 1986). These two analgesic mecha-
nisms may be parts of a broader defensive system sup-
ported by a midbrain structure called the periaqueductal
gray (Rodgers and Randall 1987; Dielenberg and McGre-
gor 2001; Castilho et al. 2002).
In summary, social and antipredator systems are typically
functionally distinct, but they generate similar demands on
the organism. These include the need to mobilize energy for
vigorous activity, focus attention and learning on emotion-
ally salient events and places, and mitigate pain and infec-
tion. These demands are met through adaptive properties of
the stress-response system.
Perceptual affordances
As conceived by Gibson (1979/1986), an affordanceis a
perceived behavioral implication of the physical properties
of some environmental feature in relation to the individual
and its context. For example, a physical property such as
brushy microhabitat has context-dependent meaning; it af-
fords safety from aerial predators for organisms of a given
size, but danger from ambushing terrestrial predators. Ro-
dents treat brush differently depending on the current source
of danger, preferring brushy habitat when threatened by
raptors, but not when endangered by terrestrial predators
(Sherman 1985; Kotler et al. 1993; Hanson and Coss 1997).
Similarly, burrows are refuges from aerial predators but not
necessarily from digging predators such as badgers (Balph
1961; Knopf and Balph 1969; MacWhirter 1992).
Conspecifics are social sources of affordances. Signals are
often designed through proximate and ultimate processes
to serve as cues for affordances; for example, the nepotistic
antipredator vocalizations of ground-dwelling sciurids are
associated with situations that afford danger and therefore
refuge-seeking and further assessment (Dunford 1977a;
Sherman 1977; Hoogland 1983; Owings et al. 1986). How-
ever, many socially generated affordances are not designed
as such, but are exploited through active assessment by
the perceiver (Otte 1974; Hennessy et al. 1981; Owings and
Hennessy 1984; note that active assessment extends be-
yond eavesdropping on signals intended for others by also
including extraction of cues from nonsignaling behavior).
An example of the latter is the ability to pick up cues about
an unfolding but unseen threat by using the perceptual ori-
entation and behavior of other individuals. For instance, an
individual typically directs its gaze to gather information
rather than to inform others, but others may use gaze di-
rection to localize the source of a disturbance. Belding’s
ground squirrels appear to locate sources of danger by ex-
ploiting such socially generated affordances, looking in
the same direction as an alarm-calling conspecific who is
watching a predator not visible to the gaze follower (Sher-
man 1977). The availability of such incidental affordances
is often cited as a source of benefits derived from social liv-
ing, an idea that has received support from the evidence
that detection of predators can be a socially collective pro-
cess (Hoogland 1981, 1995; discussed in more detail later).
Many socially generated affordances are available as by-
products of the fact that behavior tends to be matched to sit-
uational demands. California ground squirrels, for example,
adjust both their signaling and nonsignaling behavior to
the different levels of urgency of threat posed by raptorial,
mammalian, and reptilian predators. These adjustments in-
volve trading off between the conflicting demands of mini-
mizing conspicuousness to the predator and extracting cues
from the predator about the actual danger it poses. Raptors
present the most urgent threat, and squirrels minimize their
conspicuousness by whistling only once or a few times,
retreating to a burrow entrance, and remaining close to
the ground (Owings and Hennessy 1984; Sherman 1985).
Mammalian predators are intermediate in urgency of dan-
ger, and squirrels also retreat to a burrow, but then opt for
more assessment and less inconspicuousness by surveying
from a promontory or bipedal stance while chattering and
then calling repeatedly (Owings and Hennessy 1984; Sher-
man 1985). Snakes pose the least urgent threat, and squir-
rels behave much as described in our model episode, con-
fronting the snake in a way that augments their capacity to
pick up assessment cues and simultaneously increases their
conspicuousness to the snake (Owings and Hennessy 1984).
Humans can judge the level of urgency of danger from such
distinctive differences in behavior, and ground squirrels
also appear to be able to do so (Leger and Owings 1978;
Hersek and Owings 1993).Processes for use of space
Squirrel 9G’s behavior in our model episode provides a
good example of a connection between social and antipred-
ator domains with regard to the use of space. While dealing
with the snake, she also faced the social task of managing
her pups. And she rescued her young from the snake by
transferring them to the burrow of a male willing to receive
them (unpublished observations; this is not an uncommon
pattern, but we do not have data on how the male’s rela-
tionship with the female affects his receptivity to such bur-
row transfers). 9G’s direct movement to the male’s burrow
with her pups indicated that she already knew where thatSocial and Antipredator Systems: Intertwining Links in Multiple Time Frames 309