Only natural selection can account for the extreme genetic diversity o
f genes of the major histocompatibility complex (MHC). Although the st
ructure and function of classic MHC genes is well understood at the mo
lecular and cellular levels, there is controversy about how MHC divers
ity is selectively maintained. The diversifying selection can be drive
n by pathogen interactions and inbreeding avoidance mechanisms. Pathog
en-driven selection can maintain MHC polymorphism based on heterozygot
e advantage or frequency-dependent selection due to pathogen evasion o
f MHC-dependent immune recognition. Empirical evidence demonstrates th
at specific MHC haplotypes are resistant to certain infectious agents,
while susceptible to others. These data are consistent with both hete
rozygote advantage and frequency-dependent models. Additional research
is needed to discriminate between these mechanisms. Infectious agents
can precipitate autoimmunity and can potentially contribute to MHC di
versity through molecular mimicry and by favoring immunodominance. MHC
-dependent abortion and mate choice, based on olfaction, can also main
tain MHC diversity and probably functions both to avoid genome-wide in
breeding and produce MHC-heterozygous offspring with increased immune
responsiveness. Although this diverse set of hypotheses are often trea
ted as competing alternatives, we believe that they all fit into a coh
erent, internally consistent thesis. It is likely that at least in som
e species, all of these mechanisms operate, leading to the extreme div
ersification found in MHC genes.