The stability of radiogrammetrically determined Metacarpal II (M II) b
one dimensions was determined in 184 Belgian boys between 13 and 18 ye
ars of age. Stability was assessed by Fisher z transformations of Pear
son correlations between discrete measurement intervals and by structu
red covariance analysis incorporating compound, banded, and autoregres
sive models. Adjacent-age stability (1-year measurement interval) was
high (0.82 less than or equal to r less than or equal to 0.97) for all
metacarpal dimensions for all approaches, but stability coefficients
decreased with increasing measurement interval for the correlational,
banded, and autoregressive analyses. Despite the interval-related decl
ine, correlations remained moderate to high, ranging between 0.65 and
0.97 for the correlational analyses and 0.58 and 0.91 for the covarian
ce analyses, indicating a moderate to high degree of stability of M II
dimensions in males during adolescence. Periosteal diameter demonstra
ted the greatest stability of all metacarpal variables for all statist
ical analyses. Of the covariance models, the compound model provided t
he poorest fit to the data, and only for metacarpal length and periost
eal diameter did the more complex (6 parameters) banded model provide
a significantly better fit than the simpler (2 parameters) autoregress
ive model. These results indicate that M II bone dimenions are fairly
stable in males between 13 and 18 years of age and that the autoregres
sive model, with the exception of metacarpal length and periosteal dia
meter, provides as good a fit to the metacarpal data as the more compl
ex banded model. (C) 1994 Wiley-Liss, Inc.