624 
Journal of Agricultural Research 
Vol. XI, No. ii 
All correlation coefficients shown are positive, though this association 
in variation is never very close, only three of the correlations being as 
great as the smallest that Castle 1 found between skeletal dimensions in 
rabbits. The correlations at 14 days are both smaller and less uniform 
than at 280 days, indicating that animal proportions are more unstable 
at that immature age. 
The smallest correlation at 280 days is between hindleg length and 
trunk length. Jk is questionable whether the hindleg length is a satis¬ 
factory character for this purpose because of the unavoidable erorrs in 
measurement, owing to the tendency of sheep to hunch down when 
handled, thereby producing fluctuations in measurement of material 
vulue. Except this case, the 280-day correlations show a range between 
extremes which is approximately equal to the range in correlations 
between the skeletal dimensions in rabbits shown by Castle, the range 
shown by Castle being from 0.658 to 0.858, and the range shown here 
being from 0.537 to 0.700. It is very probable that the higher degree 
of association between the variations in dimensions shown by castle is 
due to the greater accuracy that can be obtained in measuring prepared 
single bones than by measuring characters in live animals, such as are 
represented in this work, most of which involves dimensions of several 
bones. 
The similarity in the ranges of correlations, however, indicates that 
it is true of sheep as Castle finds it of rabbits that 
to a large extent the factors which determine size are general factors affecting all 
parts of the skeleton simultaneously. 
The foreleg length shows the highest correlation with other measure¬ 
ments. Its coefficient with head length is 0.693 and with trunk length 
it is 0.700. 
The correlations between dimensions representing width are somewhat 
lower than this but fairly uniform. Since the growth curves (fig. 2) 
show that length dimensions attain full development somewhat earlier 
than dimensions of width, it is probable that, as width approaches 
maturity, its coefficients will more nearly approximate those of length, 
especially as the correlation does not seem to be as definite in the imma¬ 
ture as in those approaching maturity. 
1 MacDoweu., E. C. size inheritance in rabbits. With a preferatory note and appendix by W. E. 
Castle. Carnegie Inst. Washington. Pub. 196, p. 51. 1914. 
