Human variation in fetal long bones : an allometric approach

The object of this study is to describe absolute and allometric growth patterning in fetal postcranial long bones. The sample analyzed is from the DS Lamb fetal and perinatal skeletal collection at the National Museum of Natural History in Washington, D. C. Four groups containing roughly the same number of individuals were chosen for measurement and statistical analysis: White males (48), white females (29), black males (58), black females (39). Documented age is 24 to 40 weeks (age as a variable was omitted from this analysis). Long bones were measured to 0.01mm. Crown-rump lengths were documented. Statistical analysis consisted of two parts: absolute and allometric. Absolute measurements were found to be highly correlated among long bones and between long bones and crown-rump lengths for all groups. Regression statistics for absolute data described a rectilinear relationship for all long bone measurements with crown-rump length. Allometric analysis uses least squares regression of log transformed data (measurements and documented crown-rump lengths). Because the allometric coefficient equals the slope of the regression, testing for homogeneity of slope was done to determine any race/sex differences in relative growth rates. The GLM model for this test reflects an interaction between crown-rump length and a class variable RACESEX. For this study no significant differences (p<0.05) were found and the null hypothesis of equal slopes had to be accepted. A second model was performed without the interaction that tests for significant race/sex differences when adjusted for crown-rump length. Results from this model indicated significant differences for all long bone lengths and humerus distal widths. Pairwise comparisons of adjusted treatment means indicated significant race/sex differences for all four groups. Principal components analysis was used as an exploratory multivariate method to describe allometric growth patterning. Only long bone lengths were included because of their high correlation and the strength of the race/sex differences found in bivariate methods. Principal components analysis supported the patterning of long bone growth found in the regression model but was not used to determine race/sex differences. Results indicate that for any given crown-rump length, blacks have longer bone lengths and wider humerus distal widths than whites. This study anticipates similar race/sex patterning in adults where for any given stature, blacks have longer bone lengths than whites, particularly in the distal parts of each limb (radius and ulna in the upper; tibia in the lower limb).