dorphin release (Keverne et al. 1989; Spruijt et al. 1992).
The resulting capacity of grooming to produce more posi-
tive psychological states appears to have led to its use for
arousal regulation in agonistic encounters; such use, com-
bined with the fact that tonic grooming is a low-priority ac-
tivity, has made tonic grooming a reliable perceptual cue
that the groomer is reluctant to escalate (Spruijt et al. 1992).
Groomers’ adversaries may have responded to this cue by
reducing escalation, a type of exchange that could have been
the source of selection leading to ritualization of an agonis-
tic display (Tinbergen 1952), now seen as highly stereotyped
cephalocaudal self-grooming in California ground squirrels
(Bursten et al. 2000). A similar de-escalation process, fol-
lowed by ritualization, may have operated with rodent gaze
aversion (Grant and Mackintosh 1963; see also Coss et al.
2002).
Such evolutionary proliferation of the functions of self-
grooming illustrates how cascading change in biological sys-
tems might be driven by active assessment and management
(Owings and Morton 1997, 1998). Assessment is the active
quest for useful cues, such as those provided by the groom-
ing activity of an adversary when this activity correlates
with a reluctance to escalate. Management is synonymous
with regulation; animals regulate themselves and others in
their own interest. Self-interested assessment and manage-
ment give behavioral systems an opportunistic quality that
makes different functional systems potential resources for
each other. In the case of ground squirrel self-grooming, the
expropriation of grooming for the management of arousal
in agonistic contexts led to the availability of grooming as
an assessment cue that escalation was unlikely. Use of this
cue in assessment by adversaries may in turn have favored
the specialization of grooming for a management (signal)
function. The functions of this self-grooming pattern ap-
pear to have proliferated even further, becoming a part of
olfactory communication systems. Male meadow voles and
prairie voles (Microtus pennsylvanicusand M. ochrogas-
ter,respectively) use grooming as a means for broadcasting
scent to potential sexual partners (Ferkin et al. 1996; Wolff
et al. 2002). And some ground-dwelling sciurids use the
latter portions of the cephalocaudal grooming pattern to
anoint their flanks and tails with snake scent (Kobayashi
and Watanabe 1986; Owings et al. 2001), an activity that
may serve an antipredator function (Kobayashi and Wata-
nabe 1986).
Sociodevelopmental sources of antipredator inferences
Early in ontogeny, social living can generate experiences
that foster development of sophisticated inferences about
predators. These can take at least three different forms in
which information is connected to affordances via specific
perceptual cues: (1) generalizing from conspecific intention
cues to predator intention cues; (2) using social cues picked
up early from social interactions to detect or locate preda-
tors; and (3) recognizing the distinctiveness of nonmam-
malian predators, such as snakes and raptors. All of these
opportunities should be readily available from interactions
with littermates and adults during early development.Generalizing from conspecific to predator-intention cues
Interactions with parents and peers can produce experi-
ences that foster development of skills useful in anticipating
the behavior of others, some of which are probably generic
enough to be used during interactions with predators. For
instance, the schema of two facing eyes is a widespread cue
used by animals to detect when another animal is looking
at them; this cue is used during interactions with both con-
specifics and heterospecifics (Coss 1978; Topal and Csanyi
1994). Use of this cue for assessing intentions may have fa-
vored the use of directed gaze as a signal (Coss et al. 2002).
In fact, staring is an aggressive act in California ground
squirrels (Owings et al. 1977). Aggressive use of staring
may subsequently have led to signal-enhancing devices such
as the eye rings of sciurids, and even mimicry of the pat-
tern of two facing eyes in other vertebrates (Coss and Gold-
thwaite 1995). Additional examples that could be derived
from social experience include anticipating the actions of
predators on the basis of their direction of orientation, pat-
terns of movement and immobility, and postures, as well as
discovering a predator’s intentions by noting that its behav-
ior is contingent on the prey’s behavior (e.g., Levin 1997).Using social cues to detect or locate predators
As noted earlier, Belding’s ground squirrels appear to ex-
ploit the gaze direction of alarm-calling conspecifics as a
source of information about predator location (Sherman
1977). Such an ability might have its developmental roots
in learning about the significance of a littermate’s focus of
attention on a third littermate, perhaps in the context of the
third individual’s joining a play bout. Interactions of this
sort can provide a context for learning correlations between
an individual’s current behavior and subsequent meaning-
ful events, such as initiation of an interaction.Does early social learning enhance the distinctiveness
of snakes and raptors?
Norway rats undergo the functional equivalent of filial im-
printing, developing an attachment to their own mother
through olfactory learning prior to birth, and extension of
that learning to tactile cues during the early postnatal pe-
riod (Polan and Hofer 1999). Such early learning contrib-
utes to the development of a same-species preference based
on odor cues in a number of rodent species (see VasilievaSocial and Antipredator Systems: Intertwining Links in Multiple Time Frames 313