A BIOMETRIC STUDY OF THE RED BLOOD CORPUSCLES OF 
THE COMMON TADPOLE {RANA TEMPORARIA), FROM 
THE MEASUREMENTS OF ERNEST WARREN, D.Sc. 
By KARL PEARSON, F.R.S. 
(1) On leaving for Natal in 1902 to take up his duties at the Government 
Museum, Pietermaritzburg, Dr Warren placed in my hands a quantity of material, 
the reductions of which will be published with many other data in a forthcoming 
memoir on homotyposis. Among this material were measurements on the red 
blood corpuscles of 71 tadpoles from a pond at Hendon. These measurements 
were, for several reasons, of peculiar interest to me. I had been working at the 
blood corpuscles of frogs, and had met with several difficulties in the investi- 
gation. One group of these difficulties turned on the problem of the " small 
sample," and the second set on the doubt which arose in my own mind as to 
whether the size of the cell in the individual of any species can be considered 
as independent of the bulk or age of the individual. 
An interdependence of the two appeared to me not improbable on the basis 
of statistical material, a part only of which has been published. It is clear that 
any correlation of the kind indicated will influence to a greater or less extent the 
determination of homotyposis in the case of such cells. From a careful examination 
of the data themselves I was unable to detect signs of differentiation * in the sizes 
of the red blood corpuscles of any individual ; nor were my biological friends able 
to give me evidence that such size differentiation actually existed for the tadpole, 
or indeed, if it existed at all, that it affected any large percentage of the red 
blood corpuscles of the frog or any other species. Differentiation was based in 
most cases on other than size characters. 
In the absence of Dr Warren from England I do not feel justified, however, 
in making him responsible for any conclusions I may draw from his measurements. 
* A large amount of algebraical work was undertaken in the endeavour to break up the frequency 
distributions of the blood corpuscles into components, but no definite evidence of size heterogeneity 
could be obtained. The apparent bimodal nature of the characters in Tables A, D and F (pp. 412, 415, 
417) emphasised in the homotyposis tables failed to allow of algebraic resolution, and is, I think, 
solely a result of the paucity of the sample. 
