D. H. DE Souza 
489 
and length of the hip bone and the breadth of the sacrum have coefficients not 
very different from those of the transverse and inter-cristal diameters, and the 
height of the sacrum has a variability just about double that of any other bone 
measurement. The results are not surprising. Apart from the variation in a 
measurement due to individual differences, there are other causes for variation 
in the case of these measurements taken on old pelves. It is not always easy to 
obtain accurate measurements, especially of the diameters, and after death parts 
of the bone get worn away. The wearing away of the pubic crest, for example, 
and the difficulty which may be experienced in determining the lowest point of 
the symphysis, must be accountable to a great extent for the large variability 
of the pubic height. The height of the sacrum, being dependent on the number 
of vertebrae united to form that bone and on the integrity of the tip, which is apt 
to get worn, is more variable than the breadth of that bone and than the height 
and breadth of the hip bone. The extremities of these last are rounded and less 
likely to be worn. The breadth is a little more variable than the height, and 
experience proves that, of the two, it is the more difficult to measure accurately. 
The diameters must depend on the same factors as the bones, the individual 
differences, wear and tear, and difficulty of getting accurate measurements. The 
last especially must be of importance here. The obstetric conjugate, for example, 
depends upon the position of the sacral promontory, and therefore on the number 
of vertebrae forming the sacrum and on the proper fitting together of the bones, 
on the integrit} 7 of the upper and posterior part of the pubic bone, and on the 
ease or difficulty with which its anterior extremity can be determined. How 
much of the variation is due to each of these factors it is impossible to say. In 
such a diameter as the transverse there is little likelihood that the parts of the 
bone forming its extremities will be much worn away, but it would be unwise to 
go beyond a general statement of this kind. 
The pubic height and the height of the sacrum vary more than any other 
measurement of a single bone with which we are acquainted. The measurements 
of the mandible come next in order. Their coefficients of variation, as given 
by C. D. Fawcett*, are, for the greatest height 9'93, for the greatest width at 
the condyles 7 - 46, and for the greatest width at the angles 7'62. The variations 
in the lengths of the long bones in different races (found for the French and 
Aino by Lee and Pearson f, for the Naqada by Warren J, and for other races by 
Pearson §) have been compared with those of our measurements. The coefficients 
have as their limits 41 7 and 7'00, most of them being less than 5'5 ; hence the 
measurements of the pelvic bones, other than the pubic height and the height 
of the sacrum already considered, although in some few cases varying to about 
the same extent as the lengths of the long bones, have, on the whole, a tendency 
to be more variable. 
* Biometrika, Vol. i. p. 408, 1902. 
t Proc. Roy. Soc. lxi. p. 343, 1897. 
X Phil. Trans. Roy. Soc. Series B. clxxxix. p. 135, 1897. 
§ The Chances of Death, Vol. i. p. 256. London, 1897. 
