the phenomenon for which MHC had first been discovered
(See Klein 1986 for an MHC historical account). However,
only in the 1980s did the central role of MHC in immunity
begin to reveal the relationship between MHC and mate
choice. During this time, a reemerging interest in Darwin-
ian theories of sexual selection, including Hamilton and
Zuk’s fusion of immunocompetence and good genes sexual
selection theories (Hamilton and Zuk 1982), pinpointed a
crucial link between immune function and reproductive be-
havior. These authors suggested that pathogens and par-
asites might help to maintain genetic variance among the
secondary sex characteristics that are exploited (generally
by females) for mate choice. Among the various benefits of
MHC-based sexual selection, it became apparent that mate
choice based on these genes could have the immediate con-
sequence of better protecting one’s offspring against infec-
tious disease.
Rather than long tail feathers, bright plumage, or spec-
tacular behavioral displays, secondary sex characteristics
arising from the MHC are olfactorily perceived (although
see Von-Schantz et al. 1996). Leinders-Zufall and co-
workers (2004) have recently identified the long, mysteri-
ous source of MHC-based odorants, which arise from pep-
tides bound to MHC molecules. These peptides are then
detected by the vomeronasal organ (VNO), an olfactory ac-
cessory organ usually thought to detect soluble phero-
mones. Leinders-Zufall et al. (2004) applied peptides to the
VNO and found sensory neurons that detected peptides
with a specificity similar to that of MHC molecules them-
selves. If amino acids in anchor positions of the peptide
were changed, the peptide would not be recognized and the
sensory neurons would not fire. However, if non-anchor
amino acids were mutated, the changes were ignored and
the peptide was readily recognized by sensory neurons —
just like MHC! This explains how mice can detect odor dif-
ferences based on only a few amino acid differences at a
single antigen-presenting locus (Yamazaki et al. 1983; Ya-
mazaki et al. 1990; Carroll et al. 2002). Moreover, these
highly polymorphic antigen-presenting genes, with up to
450 alleles at each locus (Adams and Parham 2001) provide
tremendous underlying genetic variation, upon which sex-
ual selection and kin selection may act. Either by choosing
mates that can provide superior MHC genotypes for off-
spring, or by avoiding inbreeding with mates that share fa-
milial MHC odors, the extreme genetic diversity of MHC
genes, coupled with allele-specific odors, provide a com-
pelling system for allowing individuals to select among an
archive of “good genes.” No other candidate genes within
the MHC region are quite so convincing as potential medi-
ators of MHC-based reproductive behavior.
The origin and maintenance of MHC polymorphisms
remains a puzzle. In fact, much recent experimental and
theoretical work has been aimed at understanding the na-
ture of selective forces that must operate to maintain MHC
polymorphisms, which are orders of magnitude greater than
polymorphisms at the majority of other vertebrate loci.
Most models have focused on the role of pathogens in driv-
ing MHC diversity. Since different MHC molecules vary
greatly in their binding capacities for billions of different
possible protein antigens, the resulting allele-specific patho-
gen resistance provides one potential and powerful source
of selection for diverse alleles (Doherty and Zinkernagel
1975). However, MHC-dependent mating preferences may
also prove to be a powerful mechanism for driving MHC
diversity. MHC-based mating preferences are not only con-
sistent with pathogen-driven models of MHC diversity, but
will tend to reinforce them, maintaining diverse alleles in
the population as well as selecting for rare and novel alleles
(Penn and Potts 1999).
Surprisingly, pathogen-mediated selection mechanisms
have been exceedingly difficult to detect except in the pres-
ence of multiple pathogens or strains of pathogens (Thursz
et al. 1997; Carrington et al. 1999; Penn et al. 2002; Mc-
Clelland et al. 2003). In contrast, MHC-based mating pref-
erences have been abundantly demonstrated in mice since
Yamazaki’s preliminary observations (Jordan and Bruford
1998; Penn and Potts 1999). However, laboratory studies
using inbred strains have produced varied results that are
somewhat difficult to interpret with respect to sexual se-
lection theory (Manning, Potts, Wakeland, and Dewsbury,
1992). For instance, two out of ten mouse strains tested
showed mating preferences for MHC-similar (assortative)
rather than dissimilar (disassortative) animals (See table 1
in Jordan and Bruford 1998), somewhat inconsistent with
good genes explanations of MHC-based mate choice. More-
over, most laboratory studies tested male preferences, leav-
ing open the question of whether females, predicted to be
the choosier sex due to their greater investment in offspring,
also mated disassortatively.Sexual selection in seminatural enclosures
In the first MHC study performed in seminatural enclosures,
Potts et al. (1991) established three crucial details lacking
from previous laboratory studies. First, MHC-based mat-
ing preferences took place in the context of Muspopula-
tions, laying to rest the suggestion that MHC-based mat-
ing preferences were merely a laboratory artifact. Second,
as the mice were recently derived from wild populations,
MHC-based mating behaviors prevailed despite the in-
creased complexity of social cues resulting from randomly
assorting wild genetic backgrounds. And third, the selec-
tion coefficient arising from nonrandom mating was clearly
strong enough to maintain the allelic diversity found in sur-60 Chapter Five