206 
Variation and Correlation of the Human STiull 
and Leake were working with mustard seed and using Flower's method. These 
results confirm the view that different observers using the same method may 
have an average difference of 10 cm.', and if they use different methods an 
average difference of even 20 cm.l On the other hand, with careful treatment, 
different measurers working with different methods will hardly differ in the 
averages of fairly long series by more than 10 to 15 cm.^. I think we must 
be content with this degree of agreement at the present stage. My method of 
weighing after single packing is sensibly shorter than that of double packing, and 
it appears to give results in substantial agreement with the latter. 
(6) On the degree of Homogeneity possessed by the present Series of 
English Skulls. 
Considering the history of the ground where the skulls were discovered, as set 
forth in section (3) above, we may conclude that our material is fairly homogeneous 
without further inquiry, but it may be well to compare their variability in length 
and breadth with that of other long and probably homogeneous series. 
Variability of other Series. 
(Standard Deviations and Coefficients of Variation.) 
L 
cT 
L 
¥ 
B 
<? 
B 
? 
Series 
S. D. 
C. of V. 
S. D. 
C. of V. 
S. D. 
C. of V. 
S. D. 
C.of V. 
Naqada* 
5-72 
3-09 
5-25 
2-96 
4-61 
3-42 
4-09 
3-42 
Bavarian t 
6-09 
3-37 
6-20 
3-57 
5-85 
4-89 
3-39 
Ainot 
5-94 
3-19 
5-45 
3-08 
3-90 
2-76 
3-66 
2-67 
Parisians | 
Wbitechapel English . . . 
5-94 
5-21 
6-27§ 
3-31 
6-22 
3-45 
5-28§ 
3-75 
4-77 
3-54 
Heterogeneous Scottish 
7-42 
3-97 
7-15 
4-00 
5-96 
4-13 
5-11 
3-71 
Seeing how closely our series agrees in variability for both sexes, with the 
homogeneous Altbayei'isch crania of Professor Ranke, the conclusion that we are 
dealing with reasonably homogeneous material seems justified. The individual 
crania have also what may be described as a strong family likeness, the only widely 
divergent skull in the series excluded from my data was apparently that of a 
negro. Their mean characters agree quite closely with those of another fairly long 
series of English skulls, which I shall deal with in a later memoir. Their great 
length and dolichocephaly effectively differentiate them from any German, French 
* C. D. Fawcett, Biometrika, Vol. i. p. 424. 
t Alice Lee, Fhil. Trans. Vol. 196, A, p. 230. 
X Biometrika, Vol. ii. p. 346. 
§ For an explanation of the difierence between these results and those published in Biometrika, 
Vol. [I. p. 346, see p. 191 above. 
