This study addresses the quantitative genetic basis of phenotypic vari
ation and covariation for a series of meristic traits in the garter sn
ake Thamnophis sirtalis fitchi (six head scale counts: loreals, supra-
and infralabials, pre- and postoculars, temporals; three body scale c
ounts: ventrals, subcaudals, dorsal scale rows at midbody; two derived
traits: umbilical scar size and position). Each trait was scored on a
pproximately 540 offspring and their 47 dams captured in the wild whil
e gravid. Correlations of the meristic traits with body mass at birth,
dam's snout-vent length and body mass, litter size, and number of day
s each dam was held under laboratory conditions prior to giving birth
were removed by computing residuals from multiple regression equations
. Narrow-sense heritabilities (estimated by restricted maximum likelih
ood) of residuals were high for temporal scale counts (0.59), moderate
ly large for ventral (0.29) and subcaudal scale counts (0.41), and low
(in the range 0-0.12) for the other five traits. Probably as a conseq
uence of the low statistical power of significance testing under restr
icted maximum likelihood, only the heritability for temporal scales wa
s significantly different from zero. Phenotypic (r(p) = 0.25) and gene
tic (r(g) = 0.67) correlations between ventrals and subcaudals were po
sitive and significant. Phenotypic correlations between the head scale
counts were generally low; however, the genetic correlations were lar
ger, suggesting relatively tight integration at the genetic level. Phe
notypic correlations between the head and body scale counts were gener
ally low, but several genetic correlations were large (e.g., r(g) = 0.
59 for ventrals and infralabials, r(g) = 0.59 for subcaudals and supra
labials). These data indicate that scale counts from different regions
of the body are not evolutionarily independent characters, despite th
eir different spatial and temporal relationships during development. O
verall, genetic correlations were not strongly correlated with either
phenotypic (r = 0.42) or environmental correlations (r = 0.16).