Meta-analyses of published correlation coefficients between multilocus
heterozygosity (MLH) and two fitness surrogates, growth rate and fluc
tuating asymmetry, suggested that the strength of these correlations a
re generally weak. A variety of plants and animals was included in the
meta-analyses. A statistically homogeneous group of MLH-growth rate c
orrelation coefficients that included both plants and animals yielded
a common correlation of r(z) = 0.133. A common correlation of r(z) = -
0.170 was estimated for correlations between MLH and fluctuating asymm
etry in three species of salmonid fishes. These results suggest that s
election, including overdominance, has at most a weak effect at allozy
me loci and cast some doubt on the widely held notion that heterozygos
ity and individual fitness are strongly correlated.