PROCEEDINGS OF THE ANATOMICAL AND ANTHROPOLOGICAL 



earliest prehistoric races are usually lower than those of living civilised 

 races, but some of the living lower races (''.//., the Andaman Islanders) 

 have higher. indices than the higher or more civilised races. The belief 

 is gaining ground among anthropologists at the present day that the 

 value of indices as criteria of race has been greatly exaggerated. 



2. THE RATIONAL METHOD. 



The rational method of treating anthropometric statistics aims 

 at determining in a group of people the frequency curve for each 

 dimension, the variability about the mean, the mode or most frequent 

 dimension and the correlations between pairs of dimensions by the 

 application of rigid mathematical analysis. Another very important 

 question which is dealt with by the rational method is the estimation 

 of the deviations (due to random sampling) of observed values from 

 true values. It is seldom that an anthropometrician is able to measure 

 the whole population of any district which he is surveying. He has 

 to be content, usually, with a very small fraction of the whole popula- 

 tion. This sample is selected at random, and the measurements ob- 

 tained from one sample are not likely to be exactly identical with the 

 measurements of the next sample which might be selected. Allow- 

 ance must therefore be made for possible deviations due to random 

 sampling before we can regard any conclusions indicated by the 

 observations as significant. This testing of the significance of de- 

 viations of mean values, frequencies, standard deviations, correlations, 

 etc., obtained from random samples is one of the most important 

 parts of the rational method of treating statistics. 



The founder of the rational method of treating anthropometric 

 statistics was Quetelet, who, in his Lettres sur In Theorie des Prob- 

 dbilites published in 1846, has shown that in certain anthropometric 

 statistics, as those of stature and chest measurements, the curve of 

 frequency is approximately of the same form as the " binomial curve ". 

 Quetelet's theory was practically neglected by anthropologists till 

 quite recently, when it has been taken up and greatly extended by 

 Francis Galton and Karl Pearson. 



