veys antipredator benefits. First, prairie dogs in larger social
groups respond with shorter latencies to simulated preda-
tory attacks by badgers (Taxidea taxus), and black-tails de-
tect predators more quickly than white-tails. Second, this
advantage of group size and /or density holds even though
individuals in larger groups devote proportionately less
time to antipredator vigilance than those in smaller groups,
and black-tails are less vigilant per capita than white-tails.
Third, both breeding synchrony and center-edge differences
in individual alertness indicate that social living provides
selfish-herd effects. These social differences may have origi-
nated for antipredator reasons, as white-tail habitats con-
tain significantly more protective cover than black-tail
habitats, and black-tails enhance the visibility of their sur-
roundings by cutting down vegetation near their burrows.
Relational Similarities Generate Systems Convergence
When relations between predator and prey are similar to
those among conspecifics, social and antipredator systems
are likely to converge. Here we explore two such relational
similarities, temporally extended (tonic) proximity between
individuals and symmetry in power, and corresponding con-
vergent properties of social and antipredator systems.
Temporally extended (tonic) proximity
Ground squirrels and rattlesnakes share the same vicinity
for multiple days (see introductory episode), as conspecifics
often do. This is because rattlesnakes are ambush hunters
that choose a site and often remain there for several days,
awaiting the opportunity to surprise their prey (Hersek
1990; Greene 1997). This can generate tonic states of alert-
ness in the squirrels (Hersek and Owings 1993; Hersek and
Owings 1994). Such sustained associations have apparently
been a source of selection on ground squirrels, favoring
tonic features of antisnake behavior systems similar to the
tonic features of behavior systems used in sporadic inter-
actions with conspecifics (e.g., the cumulative effects of re-
peated interactions, as in Hinde 1974, and the design of
signaling systems to produce tonic effects, as in Schleidt
1973). Tonic features of antisnake behavior were evident in
tail-flagging, a visual signal that is specific to snake contexts
(Hennessy et al. 1981). Although tail flagging is snake spe-
cific, 90% of the tail-flagging episodes recorded in the field
were emitted outside the context of direct encounters with
snakes. Nevertheless, this tonic tail flagging was linked to
indirect tonic sources of variation in squirrel vulnerability
to snakes, such as the presence of a rattlesnake in the col-
ony that day, or developmental changes in pup vulnerability
to snake predation (Hersek and Owings 1993). The tonic
dimension of behavior varies adaptively with temporal ex-
tension of predatory threat. Rock squirrels (Spermophilus
variegatus) persist in their alertness about snakes for longer
after encounters with rattlesnakes than with gopher snakes
(Pituophis melanoleucus), who are less likely than rattle-
snakes to remain for extended periods in one area (Hanson
2003). This difference in persistence is independent of the
amountof danger posed by the snakes’ presence, as inferred
from squirrel behavior toward the snakes during actual en-
counters (Hanson, unpublished analysis).Symmetry in power
As illustrated by our opening squirrel-snake scenario, the
relationship between California ground squirrels and their
principal predator, northern Pacific rattlesnakes, involves
symmetry in power. Rattlesnakes are the major source of
pup mortality (Fitch 1948a, 1949), but adult ground squir-
rels can be formidable opponents of rattlesnakes, protected
by both skillful antisnake behavior (Owings and Coss 1977;
Hennessy and Owings 1988), and blood serum proteins that
neutralize rattlesnake venom (Poran et al. 1987; Poran and
Coss 1990; Biardi et al. 2000). While rattlesnakes normally
cannot kill adult ground squirrels, the lower serum volume
in pups falls short of that needed to prevent death (Poran
and Coss 1990). This combination of rattlesnake preda-
tion on pups and adult capacity to mount an effective de-
fense has set the stage for extended bouts of conflict be-
tween adult squirrels and snakes that can involve injury and
even death to both parties (Hennessy and Owings 1988;
Coss and Owings, 1989; Hersek 1990; Owings 2002). The
repertoire of maneuvers that each party brings to these con-
flicts is illustrated by our opening episode (Owings and
Coss 1977; Hennessy and Owings 1988; Rowe and Owings
1990).
Due to these commonalities, many of the concepts and
phenomena of intraspecies conflict also characterize inter-
actions between adult squirrels and snakes (Swaisgood,
Owings, and Rowe 1999). In both systems, fatal injuries
are rare but can occur for either participant (Fitch 1949;
Hersek 1990). Further, ground squirrels use “conventional
fighting methods” (Parker 1974) to probe and assess the
danger posed by rattlesnakes, and vary their antisnake be-
havior appropriately. These squirrels evoke rattling by con-
fronting rattlesnakes (Rowe and Owings 1978), and use
the resulting acoustic feedback to adjust their assertiveness
downward when rattle structure is a cue of high risk (rat-
tling from a large or warm snake) or upward with acous-
tic cues of low risk (rattling from a small or cold snake;
Rowe and Owings 1978; Swaisgood, Owings, and Rowe
1999, Swaisgood, Rowe, and Owings 1999; Swaisgood
et al. 2003). Ground squirrel antisnake behavior is also sen-Social and Antipredator Systems: Intertwining Links in Multiple Time Frames 311