indirect(kin) recognition, which depends on contextual
cues like the location of a nest or territory (Waldman 1987).
Likewise, Hepper (1986) distinguished between conspecific
cues and environmental cues. Some authors (Barnard 1990;
Tang-Martínez 2001) have argued that recognition based
on indirect contextual cues is not kin recognition, despite
Hamilton’s (1964) observation that spatial cues can be cor-
relates of relatedness and mediate differential treatment of
kin. When behavioral ecologists are primarily interested in
the fitness consequences of behavior, then kin recognition
will include differential treatment mediated by direct and in-
direct contextual cues (Sherman et al. 1997), but when more
proximate neurophysiological or cognitive questions are be-
ing addressed then kin-recognition studies are more likely
to focus on cues borne by individuals (direct recognition).
In a provocative and controversial paper entitled “Do
Animals Really Recognize Kin?”, Grafen (1990) argued
that “kin recognition” should only be used when the mech-
anism mediating recognition had been specifically selected
to assess genetic similarity and that, under this definition,
many putative examples of kin recognition were mediated
by mechanisms that had evolved to solve other recognition
problems, such as recognizing group members or conspecif-
ics. In many group-living organisms, however, group mem-
bers are close kin, so that kin-selected recognition abilities
would operate in these groups (Sherman et al. 1997). In
contrast, Grafen’s argument may protect against the erro-
neous inference that if kin treat each other differentially
then they must have evolved to do so because of nepotis-
tic benefits. However, it is unusual in behavioral ecology to
define terms based on both their proximate controls and
their functional outcomes. Some investigators have accepted
Grafen’s (1990) arguments (e.g., Barnard 1990; Hurst and
Barnard 1995), but his views have been challenged (Blau-
stein et al. 1991; Byers and Beckoff 1991; Stuart 1991;
Sherman et al. 1997), and most empiricists studying kin
recognition today operationalize kin discrimination as dif-
ferential treatment based on correlates of genetic related-
ness, as the term was used in the 1970s and 1980s, a tradi-
tion that we follow in this chapter.
Components of the recognition process
Kin recognition requires (1) an actorwho tries to make a
discrimination, (2) a kin templateor internal representa-
tion against which an actor compares an “unrecognized”
conspecific’s phenotype, and (3) a recipient,the unrecog-
nized individual whose identity an actor seeks to establish
(Holmes and Sherman 1983; Liebert and Starks [2004] re-
view these and other terms that have been central in the kin
recognition literature). Some conceptual clarity was brought
to the kin-recognition process when investigators acknowl-
edged explicitly that it comprises two (Beecher 1982; Sher-
man and Holmes 1985; Gamboa et al. 1986) or three (Wald-
man et al. 1988; Reeve 1989; Sherman et al. 1997) distinct
components. First, the productionor expression compo-
nentrefers to the nature of recipients’ kin labels that make
them distinguishable by actors, including when and how re-
cipients develop these labels (Tsutsui 2004). The proximate
source of kin labels may be endogenous, produced by in-
dividuals themselves, and /or exogenous, acquired from the
environment, and both endogenous and exogenous labels
may be inheritable (Gamboa et al. 1986). Second, the per-
ception componentfocuses on when and how an actor
acquires a kin template (reviewed in Mateo 2004), which
we define below. The perception component also includes
an actor’s sensory detection of a recipient’s kin labels and
the subsequent comparison that an actor makes between
its kin template and a recipient’s labels. When people re-
fer to a “kin recognition mechanism” they are often allud-
ing to the perception component. Third, the action com-
ponentwas introduced in a seminal paper by Reeve, who
defined it as “the determinants of the action taken by an
actor that has calculated a particular degree of similarity
between its template and a recipient’s phenotype.” (Reeve
1989, p. 408).
By introducing the action component (reviewed in Lie-
bert and Starks 2004), Reeve sought to make explicit that
an actor that detects a particular match between its tem-
plate and a recipient’s phenotype might behave differently
depending on the social and ecological context, tolerating
a recipient in one situation and rejecting the same recipi-
ent in another. In our view, the production and perception
components encapsulate the proximate basis of the kin-
recognition process, whereas the action component empha-
sizes the adaptive response that may follow aftera discrim-
ination has been made (fig. 19.1). If Hamilton’s (1964) rule
(r b c) is satisfied, nepotism may follow after kin are
recognized, but if the rule is not satisfied nepotism will not
occur, even if conspecifics are recognized as kin. In other
words, we suggest that it is important to separate two cate-
gories of behavior: first, the differential treatmentof kin
(e.g., fitness-neutral behaviors such as proximity during
foraging), which emerges from the production and percep-
tion components and comprise the recognition process, and
second, the preferential treatmentof close kin or agonistic
treatmentof distant kin or nonkin, which emerges from
the action component (fig. 19.1). A fitness-neutral bioassay,
like differential olfactory investigation of kin and nonkin,
might reveal that kin discrimination occurs, whereas a bio-
assay like food sharing (Porter et al. 1980) or infanticide
(Sherman 1981b; Hoogland 1985) might or might not re-Kin Recognition in Rodents: Issues and Evidence 217