252 
A First Study of the Burmese Skull 
I have made an attempt to examine in some degree the frequency with which 
certain kinds of asymmetry occur, and the extent to which they tend to be 
associated with one another. 
Dr D. E. Derry, whom I have to thank for giving me the benefit of his anatomical 
experience in examining my series for anomahes, pointed out the frequent asym- 
metry in the occipital region, and suggested that this was usually accompanied 
by a slightly greater prominence of the frontal bone on the opposite side. I have, 
accordingly, made observations as to frontal and occipital asymmetry in this 
series of skulls, and submitted the data to mathematical processes, to see what 
association, if any, existed between the two. 
Most of the observations made were merely appreciative : if some quantitative 
measure of the degree of asymmetry could be taken, the association could of course 
be determined with much greater accuracy. 
The skull was held in such a manner that the frontal bosses and that portion 
of the squama occipitalis which lies above the region of the inion were silhouetted 
against the light, care being taken that what seemed to be asymmetry was actually 
so, and was not due to any fault of position. Three categories were made, classify- 
ing the material according to whether there was greater prominence on the left or 
right, or equally great on both. The following are the numbers which fell into the 
various classes. 
Greater on 
right 
Equal 
Greater on 
left 
Total 
Occipital prominence 
Frontal ,, 
38 
41 
34 
63 
69 
37 
141 
141 
These figures confirm the suggestion that the occipital region is usually asym- 
metrical, and that in this case it is more often the left side which manifests the 
greater development. 
To test the degree of association between the two asymmetries two methods 
were open to me, neither, however, an ideal treatment for the data. One could 
throw the "Greater on Right" and "Equal" together, thus reducing the number 
of categories to two, and use the method of "tetrachoric r " to obtain a coefficient of 
correlation. This method, however, implies linearity of regression between the two 
variates. Or one could apply the method of mean squared contingency, which 
makes no such assumption, but which would give a much better result had there 
been five categories or more instead of three. I therefore used both methods. 
The results were as follows: 
Measure of association between greater occipital development on one side and 
greater frontal development on the opposite side: 
Coefficient of correlation (Tetrachoric) r = + -6169 ± -0734. 
Coefficient of mean squared contingency corrected with class index corrections 
6*2 = -6374*. 
* Before the class index correction is applied, the value obtained for uncorrected is '4965, 
