these immunologically critical loci. Second, we describe an
inbreeding study that shows that the magnitude of inbreed-
ing depression is far more severe than previous laboratory
measures have suggested. This work emphasizes the insight
to be gained by extending any measurement of a fitness cost
to include lifetime fitness. Third, we describe a partial solu-
tion to the paradoxically low frequencies of the tgene com-
plex found in nature. Despite the ability to distort genetic
transmission wildly in its own favor, mice heterozygous for
this selfish complex of closely linked genes suffer reduced
reproductive success and increased mortality, effectively
crippling the spread of the tcomplex in nature. Excellent re-
views have been written on each of these three topics in re-
cent years, and our purpose is not to provide an in-depth
synthesis of relevant literature. Rather, we illustrate solu-
tions and commonalities in the study of these three distinct
evolutionary genetic problems, showing how social ecology
was used to ascertain fitness differences and resolve these
biological puzzles.
Gene Function Is Fitness!
The ultimate function of all genes is how
they contribute to fitness in nature
Mutations that decrease the fitness (reproductive success)
of their bearers tend to be eliminated by natural selection,
and those that increase fitness tend to become fixed. A gene
with no known discernable morphological or physiologi-
cal effect in the laboratory (producing a cryptic phenotype)
may nevertheless be vital to fitness and must be analyzed
in a context appropriate for measuring the correct compo-
nents of fitness (see the following inbreeding example). Al-
ternatively, genes with well-characterized phenotypes may
have additional cryptic phenotypes (see the t-haplotype ex-
ample in the following). Consequently, characterization of
gene function will always be incomplete without fitness mea-
surements in an ecological context. Although fitness is the
ultimate target of selection, it will always be the product of
particular molecular, cellular, and physiological processes.
Understanding these mechanistic pathways is the focus of
the mainstream molecular research program. However, this
research program reaches a dead end in the absence of phe-
notypes. Fortunately, ecological approaches can help reveal
many phenotypes that are cryptic under laboratory condi-
tions, allowing the molecular characterization of gene func-
tion to proceed. An integrative approach combining fitness
assays followed by molecular studies promises to be a pow-
erful procedure for the discovery of gene function, particu-
larly for genes with cryptic phenotypes.
The practical problem is how best to measure fitness. Se-
lection is difficult to measure in the laboratory. The artifi-
cial conditions of caged breedings often produce inconsis-
tent results and have had limited success overall in defining
the mechanisms underlying reproductive differences among
genotypes. Yet studies performed in the wild have their
own sets of problems. Stochastic environmental conditions
(weather, food, shelter, etc.) increase the variance of tested
effects, adding noise to already statistically complex data
sets, and the loss of subjects to dispersal and various
sources of mortality confounds lifetime measures of fitness.
House mouse communities are socially complex, and mat-
ing success depends on a suite of ecological and social cues
for information on territory, mate, and rival quality. By
providing mice with a means by which to display and uti-
lize these socially and ecologically relevant cues while con-
trolling extrinsic environmental factors, tests of genetic hy-
potheses become quite manageable. The house mouse is
already the leading mammalian model organism for mo-
lecular and physiological studies. However, the ability to
simulate an ecological field study within a controlled envi-
ronment makes the house mouse extremely tractable for so-
cial ecology studies as well. Combining these two strengths
makes the house mouse the vertebrate organism of choice
for studying questions that require molecular, physiologi-
cal, and ecological approaches. As we hope to demonstrate
in this chapter, the massive enterprise of determining gene
function demands such broad approaches.Seminatural Populations in MusThe success of seminatural Muspopulation studies is likely
owed to the 10,000-year-long commensal relationship be-
tween house mice and humans that exists to this day. The
highly social house mouse (Mus musculus/Mus domesti-
cus) is quite willing, even enthusiastic to set up complex
societies within the wire, steel, and concrete labyrinth of
racquetball-court-sized artificial mouse enclosures. Our at-
tempts to provide mice with environmental complexity
without obscuring our view of them involves a consider-
able amount of strategically placed wire mesh, which seems
to provide an important element of spatial complexity
(fig. 5.1). Wire mesh partitions subdivide each enclosure into
six or eight equal subsections. The same material is placed
in cylinders or spirals around food corrals, and made into
platforms with multiple tiers for climbing. Mice can easily
scale the wire mesh and learn to maneuver quickly around
and over it, but the partitioned subdivisions tend to provide
dominant males with convenient territorial boundaries,
and all mice have to negotiate a texturally complex three-
dimensional space to cover any ground. In addition to these
steel structures, mice are given unlimited food and water,58 Chapter Five