infer something about kin recognition (e.g., mothers cannot
recognize their own offspring), but we caution that an ap-
propriate bioassay must be used to distinguish between ma-
ternal care, which may or may not be directed differentially
toward own and alien young, and maternal discrimination
abilities, which should be assessed with a neutral measure
such as differential olfactory investigation. For example,
in laboratory tests communally-nesting female house mice
(Mus musculus domesticus) retrieve age-matched own and
alien pups with equal frequencies (Manning et al. 1995),
which suggests that mothers may be unable to distinguish
between their own and alien young. Similarly, captive fe-
male S. beldingiretrieve alien pups into their own nests as
often as their own young, which suggests an inability to
discriminate. However, after bringing pups back into the
nests, mothers sniff and handle alien pups longer than their
own pups, indicating that mothers can discriminate be-
tween the two categories of pups and that mothers are mo-
tivated to tend to both (Holmes 1990). This means that to
test maternal discrimination abilities more directly, an op-
erant procedure or one like olfactory habituation (see the
preceding) should be used to reveal abilities that may be
masked by maternal motivation.
What we do know about mother-offspring recognition
abilities in communal nesting species is based on indirect
evidence rather than on studies designed explicitly to test
for recognition (Lewis and Pusey 1997). For example, we
know from field work that in maras (Dolichotis patagonum)
multiple females often share warrens, and that after young
come aboveground mothers reject nursing attempts by
aliens, but we know nothing about preemergent mother-
offspring recognition (Taber and MacDonald 1992a). To
draw ecologically valid inferences from laboratory studies
of communal breeding, females should be housed together
(e.g., in pairs) so that “nestmates” include young from mul-
tiple litters. In a mother-offspring recognition study of com-
munally-breeding degus (Octodon degus), for example, fe-
males lived together in pairs, and laboratory tests showed
that mothers could discriminate between the odors of their
own young and their co-nestmate’s young, despite mothers
being housed together since both of their litters were born
(Jesseau 2004). The paucity of empirical work on kin rec-
ognition in communal species stresses the need for research
on both the proximate and ultimate levels of analysis, to ex-
plain maternal care of alien young and the role of mother-
offspring recognition in this unusual form of parental care.
Earlier, we emphasized the distinction between differen-
tial treatment and preferential treatment of kin (fig. 19.1),
and we suggest that it is especially important to make this
distinction for communal breeders because, depending on
the cost and benefits experienced by females, mothers may
or may not nurse their own and alien young selectively,
which may or may not indirectly indicate whether mother-
offspring discrimination is possible.Does the absence of preferential treatment
indicate a lack of recognition?
Whether the proximate or ultimate aspects of kin recogni-
tion are addressed, the assay chosen to assess discrimina-
tion abilities is crucial. We suggest that there are two gen-
eral categories of kin-recognition studies based on the types
of assays chosen. A direct-inference studyis one in which a
(presumably) fitness-neutral assay like olfactory investiga-
tion or time-in-proximity (Smale et al. 1990; Todrank et al.
1998; Mateo 2002) is used to infer that discrimination oc-
curs. An indirect-inference studyis one in which discrimi-
nation abilities are inferred from behavior that is either nep-
otistic, behavior that reflects mate choice, or behavior that
is likely to have clear phenotypic costs (e.g., fighting or other
forms of agonism; Barnard and Fitzsimons 1989; Keane
1990b; Sera and Gaines 1994). Both kinds of studies can re-
veal kin-discrimination abilities. However, when differen-
tial treatment is not detected, the conclusion that discrimi-
nation is not possible is better supported (but not verified)
by results from a direct-inference study than an indirect-
inference study. This is because in an indirect-inference
study the phenotypic costs that actors might incur or the
benefits they might gain may be such that Hamilton’s rule
(1964) is not satisfied, so actors should not adjust their be-
havior based on differences in kinship (Beecher 1991).
For instance, despite the prevalence of studies demon-
strating phenotype matching in several rodents (Mateo
2003), most evidence for kin recognition in voles indicates
that prior association mediates sibling recognition in Mi-
crotusbecause only kin that are reared together are treated
as relatives (Boyd and Blaustein 1985; Ferkin and Rutka
1990; Sera and Gaines 1994; Berger et al. 1997; Paz y Miño
C. and Tang-Martínez 1999c). The absence of differential
treatment of unfamiliar relatives by voles during paired-
encounter tests has been interpreted by some to demon-
strate an inability to recognize unfamiliar kin (e.g., Gav-
ish et al. 1984; Paz y Miño C. and Tang-Martínez 1999b,
1999c). This may be a premature conclusion, as these stud-
ies used indirect-inference designs. The key questions are:
do voles produce cues that vary with relatedness (the pro-
duction component), and if so, are they able to use this
information to identify kin (the perception component)?
Direct-inference studies on voles’ discrimination abilities
would be valuable because they could reveal whether kin-
recognition abilities exist, despite not being expressed in all
contexts. When bank voles (Clethrionomys glareolus), for
instance, were presented two unfamiliar, anesthetized adult
males in olfactory preference tests, females spent more timeKin Recognition in Rodents: Issues and Evidence 227