2 
Proceedings of the Royal Society of Edinburgh. [Sess. 
mean. As the range of variation of the cephalic index is usually quite as 
great as the range of variation of the absolute dimensions, it is difficult to 
understand the belief in the value of indices of the Empirical school.” 
Myers (19), referring to this Empirical method, when reviewing some 
recent work of Fawcett and Lee, also says : “ It is to be hoped that never 
again will the old school be tolerated which collects a few measurements, 
dissects them, and publishes ill-founded conclusions. 
“ The Rational method aims at determining in a group of people (a) the 
frequency curve for each dimension ; (6) the variability about the mean ; (c) 
the mode or most frequent dimension ; (d) the correlation between pairs of 
dimensions by the application of rigid mathematical analysis ; (e) the 
estimation of the deviations (due to random sampling) of observed values 
from true values.” 
Quetelet (4) was the founder of this rational method of treating anthro- 
pometric statistics, and, as far back as 1846, applied this method to statistics 
of stature and chest measurements ; but his theory was practically neglected 
by anthropologists till quite recently, when it was revived and extended by 
Galton (5) and (6), and Pearson (7) and (8). 
Fawcett and Lee (1) were the first to apply statistical methods to any 
series of skull measurements, though Stieda (3) published the first scientific 
determination on the variabilities of the skull. 
That the Rational method is to replace the Empirical or old school 
methods is borne out by the published opinions of recent investigators ; e.g. 
Macdonnell (2) says : “ I venture to think that the chief aim of craniologists 
at present should he to table the means, standard deviations, and correla- 
tions of further long series of skulls.” And again the same writer says : 
“ Only when that collection is far more complete will it be possible to state 
general conclusions applying to the whole range of craniology ; and when 
such tables are formed for forty or fifty long series, we shall have far more 
light, not only on intra-racial, but on inter-racial problems.” 
Fawcett and Lee (1) state that “ the correlation of the mean values of 
the chief craniological characters in fifty or one hundred races would he a 
most valuable investigation, breaking practically untrodden ground.” 
“ The first determination of the correlation of any parts of the skull 
was,” say the same authors (1), “ made in 1895, and published by Pearson 
(8), who correlated length and breadth of skull in modern Germans, 
modern French, and the Naquada crania. In his memoir on Spurious 
Correlation (9), published in 1896, he gives the correlation values of the 
length, breadth, and height, and the two cephalic indices.” 
Further correlation results were published in 1899 by Dr Alice Lee (12) 
