Phenotype matching
In this recognition mechanism, an individual learns some-
thing from its own phenotype or some of the features of the
kin with which it is reared, and acquires a kin template^1
that it later uses as a prototype against which other pheno-
types are compared. Kin templates are often acquired dur-
ing early development in a particular context, such as a
natal nest in which only siblings reside. Depending on the
match or overlap between the phenotype being assessed and
the individual’s template (Getz, 1981; Lacy and Sherman
1983), a conspecific may be recognized as a relative. Phe-
notype matching may explain recognition of unfamiliarkin
like paternal half-siblings, which has been documented in
ground squirrels (Holmes 1986a; Mateo 2002), laboratory
mice (Kareem and Barnard 1982), and some primates (Wu
et al. 1980; Widdig et al. 2001; Smith et al. 2003), father-
offspring recognition when males do not associate exclu-
sively with their young during early development (Buchan
et al. 2003) and recognition betweenfamiliaryoung that de-
velop together in litters comprising unequally related young
due to multiple mating by females or communal nesting
(Hauber and Sherman 2001).
Phenotype matching requires a detectable correlation
between genotypic and phenotypic similarity, and in many
rodents this correlation is reflected in the olfactory labels
shared by close relatives (Porter et al. 1983; Hepper 1987;
Sun and Müller-Schwarze 1997, 1998a). If the odors of
close kin are more similar than those of distant kin (Todrank
and Heth 2003) then olfactory investigation time should
vary with kinship. For instance, when juvenileS. beldingi
(fig. 19.2) were presented dorsal-gland and oral-gland odors
from their more and more distantly-related kin, juveniles
increased their investigation time (fig. 19.4), presumably
because the odors of more distant kin were perceived as
more unfamiliar (not matching the templates as well) than
those of their more closely related relatives (Mateo 2002).
(Across taxonomic groups and modalities, animals gener-
ally attend to novel stimuli longer or more strongly than
to familiar stimuli; Johnston 1981; Halpin 1986; Stoddard
1996). The genotypes of close kin become more similar if
inbreeding occurs, which, in group-living species, may make
it difficult for individuals to distinguish close kin from other
group members. For example, inbreeding is common (al-
though not exclusive [Braude 2000]) in colonies of naked
mole-rats (Heterocephalus glaber) and the mean coefficient
of relationship is 0.8 (Reeve et al. 1990). This may con-
tribute to why female naked mole-rats rely on familiarity to
recognize kin and do not appear to assess genetic similarity
by phenotype matching (Clarke and Faulkes 1999, based
on mate-choice preference tests (see also Ciszek 2000).
Both the prior association and phenotype-matching
mechanisms require that kin labels, the production compo-
nent of the recognition process, differ among classes of kin.
In ground squirrels, odors from oral and dorsal glands are
involved in social discrimination (Kivett et al. 1976; Harris
and Murie 1982), and in S. beldingioral and dorsal secre-
tions differ across at least four classes of kin, which makes
it possible to discriminate these kin labels in controlled tests
(Mateo 2003). Just as selection has molded the perception
component of the recognition process, selection is also ex-
pected to have molded the production component, so that
whether ground squirrels and other rodents can recognize a
particular category of kin may depend on discernable vari-
ation in the production component (Beecher 1991; Reeve
1989; Sherman et al. 1997). Thus like the perception com-
ponent (Mateo 2004), selection acting on the production
component may limit nepotistic behavior.
For example, S. beldingifemales do not treat distant
kin, like aunts or cousins, nepotistically (Sherman 1980a,
1981a). One possible functional explanation is that Hamil-
ton’s (1964) cost /benefit criteria are not met when mem-
bers of these classes interact (fig. 19.1). Alternatively, nep-
otism may not occur between distant kin because actors
cannot recognize them as relatives. However, we know that
distant female kin in S. beldingido produce olfactory kin
labels, and that other females discriminate among these la-
bels in controlled tests (fig. 19.4), which suggests that a fail-
ure to satisfy Hamilton’s criteria for nepotism rather than
limitations in the recognition component explains why dis-
tant kin are not recipients of nepotism (Mateo 2002). Thus
a system for nepotism and one for kin recognition can
evolve separately. That is, selection can act independently
on the recognition component(s) and the action component
(fig. 19.1), which calls into question the commonly held be-
lief that the absence of nepotism must be due to the absence
of a kin-recognition mechanism.Kin recognition in other ground-dwelling squirrels
Based on some of the early work on ground squirrel kin rec-
ognition, investigators began to seek parallels between inter-
specific differences in social organization and recognition
abilities (e.g., Michener 1983a; Holmes 1984b; Schwag-222 Chapter Nineteen
- The meaning of “template” is not well developed in the kin-recognition lit-
erature of behavioral ecology (although see Sherman et al. 1997), which may help
explain why investigators’ views differ about the relevance of templates to the
prior association and phenotype-matching mechanisms. In our treatment of phe-
notype matching, we use templatelike cognitive psychologists use prototypeto
refer to a set of features that are most commonly present in all members that
belong to a particular category (Smith and Medin 1981; Estes 1994; Hampton
1995). Prototypes arise when organisms experience individual instances within a
category and abstract the common attributes of those instances, storing them as
a generic representation of the common attributes of the category as a whole. In
contrast, exemplars arise when organisms experience individual instances within
a category and store those individual instances rather than abstracting the com-
monalities among them. In our view, templates, or more properly, prototypes, are
central to the phenotype-matching process, whereas exemplars are central to the
prior association mechanism (and individual recognition).