veal kin discrimination, depending on the inclusive fitness
costs and benefits of the behavior (Beecher 1991). If food
sharing or infanticide were viewed as the action component
then the absence of differential treatment of kin could mean
that the actor was unable to discriminate kin from nonkin
orthat the actor’s fitness interests were better served by fail-
ing to treat kin and nonkin differently. That is, a behavioral
assay that is likely to affect an actor’s inclusive fitness may
obfuscate studying differential treatment (kin discrimina-
tion) separately from preferential treatment (nepotism). We
believe that including the action component as part of the
kin-recognition process entangles a proximate mechanism
and the functional behavior it generates (see Griffin and
West 2003). However, we certainly believe that complete
explanations for kin recognition require analyses of prox-
imate andultimate factors (Holmes and Sherman 1982;
Blaustein et al. 1991; Mateo 2002).
Experimental methods and research designs
Experimental studies of kin recognition raise several meth-
odological issues (Gamboa et al. 1991; Todrank and Heth
2001) and here we highlight a few. Kin-recognition abilities
are often inferred from field data that describe variation
in social interactions with kinship (Hoogland 1995; Che-
ney and Seyfarth 1999; Smith et al. 2003; Griffin and West
2003). However, field studies that assess kin recognition
with fitness-neutral assays (i.e., not involving kin favorit-
ism) are rare (Michener 1973a; Waldman 1982; Sun and
Müller-Schwarze 1997), and most experimental studies have
been conducted on laboratory-born, reared, and tested an-
imals (reviewed in Dewsbury 1988; Mateo 2003), which
raises some questions about ecological validity. For ex-
ample, if kin labels depend, in part, on diet, then maintain-
ing captive animals on the same diet may minimize differ-
ences in labels that might be important to discrimination
in nature, leading to false negatives. In spiny mice (Acomys
cahirinus), for instance, both genotype and diet contrib-
ute to kin labels, and if diets are changed so that members
of a sibship are maintained on different diets, then wean-
lings’ preference to huddle with their siblings disappears (al-
though their ability to discriminate may remain; Porter et al.
1989). If captive rearing interferes with the developmental
process that produces species-typical kin-recognition abili-
ties, then one solution is to livetrap field-reared animals and
study their discrimination during brief tests in captivity, as
Hare (1992; 1994) did to study littermate recognition in
Columbian ground squirrels (Spermophilus columbianus).
Another alternative is to conduct parallel studies in the lab-
oratory and field, comparing the recognition abilities that
are displayed in the two environments (Holmes and Sher-
man 1982; Gamboa et al. 1986; Blaustein and Waldman
1992).
Understanding kin-recognition mechanisms at the prox-
imate level lends itself to experimental study because the
two factors that often mediate recognition can be manipu-
lated by cross-fostering. These two factors include prior as-
sociation,direct interactions between individuals that re-
sult in learned familiarity with each other’s phenotypes, and
genetic relatedness,the factors besides prior association that
correlate reliably with genes identical by descent, like phe-
notypic similarity among kin. Various cross-fostering de-
signs have been used to study how prior association and
genetic relatedness mediate kin recognition (Todrank and
Heth 2001; Mateo and Holmes 2004). Typically, an infant
is taken from its genetic mother at birth and transferred to
a foster mother that is unrelated to the infant. One or more
infants may be fostered and infants are often exchanged re-
ciprocally between two litters. The effects of cross-fostering
on recognition abilities can be complex (Todrank and Heth
2001) because when more than one member of a litter is
transferred to a foster litter the transferee will be exposed
to (1) a foster parent, (2) foster siblings, (3) a genetic sibling
that was transferred with it, and (4) its own phenotype, all218 Chapter Nineteen
Figure 19.1 Kin recognition requires (1) the production of unique kin labels
(the production component) and (2) the ability to perceive such labels and com-
pare them with an internal representation according to some sort of matching
rule (the perception component). These two requirements comprise the recogni-
tion component of the kin-recognition process. Kin recognition mechanism can
facilitate nepotism, the action component, but only if Hamilton’s (1964) rule is
satisfied. (An X in a box indicates the absence of a functioning component.) It is
thus useful to distinguish between differential treatment of kin (kin recognition)
and preferential treatment of kin (nepotism), because the former will not inevi-
tably result in the latter. Kin recognition may also mediate mate choice, which is
not considered in this diagram.