of which could influence subsequent recognition abilities
(Mateo and Holmes 2004).
For example, there is considerable interest in whether an
individual can use its own phenotype as a standard against
which other phenotypes are compared to assess kinship
(Hauber and Sherman 2001; Todrank and Heth 2003; Tsu-
tsui 2004). If cross-fostering is used to test a self-matching
hypothesis then only one member of a litter can be trans-
ferred to a foster litter, and none of the individuals in the
foster litter, including the foster mother, can be related to
the transferee whose self-matching ability is being studied.
This transfer regimen ensures that any recognition of unfa-
miliar kin by the transferee is based on self-inspection rather
than on comparison with the foster family’s phenotypes
(Mateo and Johnston 2000a). Moreover, cross-fostering
does not address whether in utero experience influences the
ontogeny of kin recognition (Hepper 1987; Robinson and
Smotherman 1991), although embryo transfer can be used
to address in utero effects on discrimination abilities (e.g.,
Isles et al. 2001).
Many kin-recognition studies on rodents have used ei-
ther a paired-encounter design, in which the social inter-
actions of two conspecifics are observed for a short period
of time in an unfamiliar environment (e.g., Michener 1974;
Grau 1982; Ferkin and Rutka 1990), or a single-subject de-
sign, in which a phenotypic cue from a conspecific (e.g., an
odor) is presented to a subject and its response is assessed
(e.g., time spent sniffing; Hepper 1987; Murdock and Ran-
dall 2001; Mateo 2002). A benefit of the paired-encounter
design is that it simulates situations in nature when an ac-
tor encounters another conspecific and engages in recogni-
tion behavior. A weakness of the design is that it may not
be possible to separate one individual’s discrimination abil-
ities from another’s (Holmes 1986b). If, for example, A can
recognize B as a sibling then A might rarely behave agonis-
tically toward B. However, if B fails to recognize A as a sib-
ling then B might routinely direct agonism toward A, which
could elicit more agonism from A, thus hiding from the in-
vestigator A’s ability to recognize B. Asymmetrical recogni-
tion abilities may be common in parent-offspring interac-
tions at some points in development (Holmes 1990; Insley
2001), although the absence of observable behavioral dis-
crimination on the part of young may be due to motor lim-
itations rather than recognition abilities. Because the adap-
tive problems faced by parents and young differ, it would be
valuable to determine if recognition is symmetric or if only
one member of the parent-offspring dyad can recognize its
relative (Gustin and McCracken 1987; Beecher 1991; Ins-
ley 2001), which may require use of single-subject designs.
The single-subject design is especially useful for examin-
ing the sensory basis of kin recognition. This is because cues
from one modality can be presented in isolation from oth-
ers (Todrank et al. 1998; Beecher et al. 1989; Parr and de
Waal 1999). In rodents, olfactory cues seem to be the pri-
mary mediator of kin recognition (Halpin 1986; table 1 in
Mateo 2003). Besides standard odor-preference tests (e.g.,
Ferkin and Rutka 1990; Clarke and Faulkes 1999), the
habituation-discrimination technique has helped answer
questions about odor-based kin recognition. The basic pro-
cedure entails recording a test animal’s response (e.g., sniff-
ing time) during repeated presentations of an odor — the
habituation stimulus — followed by the presentation of a
new odor, the test stimulus. Investigation time typically de-
creases across repeated presentations (habituation) and in-
creases when a new odor is presented, if the actor can dis-
criminate between the original and the new odor (Johnston
1993; Mateo 2002; Todrank and Heth 2003). The habitu-
ation technique has been valuable in studies of the produc-
tion component of kin labels (e.g., Todrank et al. 1998) be-
cause it reveals whether test animals perceive odors from
related individuals as being similar, which could facilitate
recognition by phenotype matching (see the following). The
technique also can demonstrate whether animals perceive
odors of their distant kin (e.g., cousins) as different from
nonkin (Todrank and Heth 2003), which could help explain
why nepotistic treatment is (or is not) limited to certain
classes of kin (Mateo 2002). When using the habituation
technique, it is important to appreciate that a test animal’s
experience prior to the study itself may influence its re-
sponse. For example, an individual that discriminates be-
tween odors of its full-siblings and paternal half-siblings
might not be distinguishing differences in odor that corre-
late with kinship per se, but rather between familiar and un-
familiar odors, if the individual was reared with siblings
and had never encountered paternal half-siblings.
Laboratory studies enable experimental control, but they
usually lack the contextual features that may affect discrim-
ination performance. For example, when juvenile (young-
of-the-year) Belding’s ground squirrels (S. beldingi,fig. 19.2)
were tested in the laboratory, they appeared unable to dis-
criminate between littermates and nonlittermates during
(1) paired-encounter tests when agonistic interactions were
assessed, (2) food-sharing tests, and (3) sleeping-partner
preference tests (Holmes, unpublished data). However,
when another set of juveniles was observed in large, out-
door enclosures, complete with natural vegetation and bur-
ied nestboxes, juveniles readily distinguished between lit-
termates and nonlittermates (Holmes 1994).
In rodents, social experience during early development,
often at specific times, is crucial to the ontogeny of kin rec-
ognition. For example, juvenile Columbian ground squir-
rels interact differentially with their littermates and non-
littermates during the first few weeks after they emerge
aboveground from their natal nest (Waterman 1986, 1988).Kin Recognition in Rodents: Issues and Evidence 219