multiple matrilines but all workers acquire similar odors
upon eclosion (see Keller 1997). Nonolfactory cues may
also be used to detect kinship when animals do not produce
or perceive olfactory labels that are sufficiently variable
to support discrimination (e.g., visual similarities in zebra
finches; Burley and Bartels 1990). Future work should ad-
dress the potential for other nonolfactory modalities to
mediate recognition, such as ultrasonic vocalizations used
by bats in mother-offspring recognition (e.g., Balcombe
1990; Scherrer and Wilkinson 1993). In addition, most kin-
recognition research has focused on direct mechanisms of
recognition, yet there may be instances when indirect spa-
tial cues mediate differential treatment of kin. Mother-
offspring recognition, for example, is mediated by spatial
cues in many burrow-nesting rodents during early stages
of lactation when pups are confined to their natal burrow
(Holmes 1990), but as young become mobile spatial cues
will no longer correlate reliably with relatedness. It would
be valuable to investigate how mothers make the transition
from offspring recognition based on indirect spatial cues to
direct kin labels, including the question of when during
early ontogeny young begin to produce their distinctive ol-
factory kin labels and when mothers learn these labels and
begin to respond differentially to them.
Kin recognition and cooperative breeding
Cooperative breeding, which has been variously defined
(Solomon and French 1997), refers to a social structure in
which individuals help care for offspring produced by other
conspecifics, and is often explained at the ultimate level by
ecological constraints and Hamilton’s inclusive-fitness hy-
pothesis (Emlen 1991; Jennions and Macdonald 1994). Co-
operative breeding raises intriguing questions about kin
recognition (Lewis and Pusey 1997; Komdeur and Hatch-
well 1999; Griffin and West 2003), and we believe that this
is especially true for mammals when cooperative breeding
is characterized by communal nesting, which occurs when
two or more females raise their young together in a com-
mon nest or burrow. For example, prior association medi-
ates mother-offspring recognition in many species (Holmes
1990), yet in communal nesting if infants’ kin labels do not
develop until after litters mix, or if mothers do not learn
the kin labels of their own offspring before litters mix, then
mother-offspring recognition could be jeopardized, which
might lead to misdirected maternal care (König 1989b).
Communal nesting occurs in several species of rodents (re-
viewed by Solomon and Getz 1997; Hayes 2000), and we
suggest that it offers one of the most fruitful avenues for
future research on rodent kin recognition. For example,
communal nesting provides the ecological and social cir-
cumstances under which self-referent phenotype match-
ing rather than prior association might mediate mother-
offspring recognition (Holmes and Sherman 1982; Hauber
and Sherman 2001), and any study that revealed mother-
offspring discrimination in a communal nester would help
shed light on the functional significance of communal nest-
ing (see the following), which has remained elusive (Packer
et al. 1992; Lewis and Pusey 1997; Hayes 2000).
Perhaps the greatest paradox of mammalian communal
nesting is that communal nursing, the sharing of milk with
young produced by another female, often is associated with
it (König 1997; Hayes 2000). Given the physiological costs
of lactation (König et al. 1988), behavioral ecologists ex-
pect that under most but not all circumstances mothers
should avoid suckling offspring other than their own, which
we refer to as “alien offspring.” Communal nursing offers a
valuable empirical opportunity for proximate and ultimate
explanations to intersect and to enhance our understanding
of mammalian communal breeding (Lewis and Pusey 1997;
Griffin and West 2003). For example, Packer et al. (1992)
and Roulin (2002) summarized several functional hypothe-
ses to explain why mothers sometimes nurse aliens, includ-
ing kin selection and reciprocity, as well as a nonfunctional,
misdirected parental-care hypothesis. The last explanation
is invoked frequently, particularly when conesting females
are known to be unrelated, with the conclusion that the
costs of communal nursing were not large or predictable
enough to outweigh the benefits of enhanced offspring sur-
vival that sometimes occur in communal nesting (e.g., Man-
ning et al. 1995), or that the benefits of preventing milk
theft did not outweigh the costs of vigilance needed to
thwart it (Packer et al. 1992). Black-tailed prairie dogs nurse
their own infants exclusively until young first emerge, but
when juveniles from different litters begin to share burrows
after natal emergence mothers nurse communally. Investi-
gators suggest that mothers may not be able to discriminate
between their own and alien juveniles when both occupy a
common burrow, or that the cost of trying to discriminate
does not outweigh the cost of sharing milk (Hoogland et al.
1989). We suggest that results from mother-offspring dis-
crimination tests would immediately help narrow the field
of possible functional explanations for communal nursing
in a particular species. For example, if it were shown that
mothers wereable to recognize their young, or to discrimi-
nate among infants of varying relatedness, then one could
rule out the misdirected parental explanation and begin
to explore functional hypotheses for communal nursing.
In short, in the absence of detailed knowledge about the
costs and benefits of communal nursing and pup retrieval,
it would be wrong to conclude that mothers cannot recog-
nize their own offspring simply because they retrieve and
nurse alien young.
Evidence that communal nursing occurs is often used to226 Chapter Nineteen