M. L. TiLDESLEY 
247 
Here again we find our Burmans in close proximity to the Malayan Groups of 
the West Indian Archipelago. For each index of the nasal bridge the Burmese 
Group C gives a higher value than any other race for which it has been obtained, 
English and Egyptian included. It is regrettable that at present we have not these 
indices for the bridge of the nose in the case of more Caucasian races, nor for 
Chinese or Dravidians. It is clear that a complete table — especially with rather 
longer series for some of the races here provided — would be of the greatest sug- 
gestiveness. We need also to discover, if possible, the values for palaeolithic and 
neohthic man. The nasal bridge index and the differential cephalic index [{B — H)/L] 
seem to me perhaps the more capable than any other cranial indices of providing 
adequate racial characterisations. 
12. Coefficient of Racial Likeness. 
We turn now to the problem already stated (p. 244) of obtaining a measure of 
the relationship between different races. For the general solution of this problem, 
an extensive knowledge of the degree of correlation between different cranial 
characters is required. We have two types of correlation in craniometry. 
First, the intra-racial, that is, the correlation of characters in the individual, 
the measure of this relation being obtainable from a group of individuals to which 
he belongs. A good deal is known about such correlations already; more will be 
known when the long series of Egyptian skulls — 1800 of one period — now in hand 
in the Biometric Laboratory, is completely reduced. But we do know quite enough 
to assert that the correlation is never very high between cranial characters which 
do not have any portion in common, and which are not right and left measurements 
of homologous characters. It is indeed often wholly neghgible. 
The second form of correlation is that between the means of characters in different 
races, and to obtain a measure of this we must have a large number of races whose 
means have been established. This may be conveniently termed inter-racial corre- 
lation to distinguish it from intra-racial. 
There are at present, however, so few races for which we have a knowledge of 
the mean values of forty or more characters, based upon a series of adequate 
dimensions, that the principal inter-racial correlations remain largely unknown. We 
do know that inter-racial and intra-racial correlations differ considerably in a 
number of cases: thus, for example, the value of the relation between orbital and 
nasal index in the individuals of one race is considerably different from that between 
their mean values in a group of races. There is, however, on the basis of our present 
knowledge, no reason to believe that inter-racial correlations are likely to be at all 
intense. Accordingly, when we consider the difference in cephalic index of two 
races, and then proceed to consider the difference in zygomatic breadth, any infer- 
ence of significant difference in the first will not be weakened by the fact that we 
find a significant difference in the second: in other words, the two judgments may 
be considered independent in the present state of our knowledge of inter-racial 
correlation. That is to say, we shall obtain at least a first approximation to an 
