370 Darivinism, Biometry and Some Recent Biology. I. 
evidence we have at present, I should not expect to find a correlation between 
fertility and a somatic character ; if it does exist in any case, I should anticipate 
very rapid changes going on in that species. 
(4) One of the most recent investigations in this direction is that of Raymond 
Pearl on the inheritance of fecundity in Barred Plymouth Rocks. His paper on 
the number of eggs of this race is an extremely interesting one. It is true that 
he only gives the crude correlations and does not give the data which would be 
necessary for their correction. Thus he had 31 mothers* and 180 daughters, and 
some of his " arrays " were probably due to a single mother bird, but this cannot 
be ascertained from his Tables. In working at fertility in man I found exceedingly 
small values, until correction was made for weighting of the more fertile mother, 
and then the values appeared to be small but appreciable. If we put aside, how- 
ever, the small number of mothers used by Pearl and this question of correction, 
there cannot be the least doubt that Pearl's data show a low, if not zero, intensity 
of heredity in fertility, thus confirming in poultry what we already know of many 
mammals, i.e. the absence of inheritance in the case of fertility which seems 
a necessary foundation of the idea of Darwinian evolution. 
Pearl does not take this zero correlation as confirmation of existing results for 
fertility. He considers it in some way a confirmation of the pure line theory of 
Johannsen. We are sorry that we cannot in the least agree with him on this point. 
For if his view were correct, every high correlation would tell against such a theory, 
which it certainly does not ; and in the case of mammals, where the fertility 
correlations are zero and the body characters highly correlated between parents 
and offspring, one correlation would refute, another confirm the pure line theory 
in the same species. Nor again can data taken from measurements of a character 
highly influenced by environment be in the least conclusive, if they are handled 
in the manner recently adopted by Jennings, who selected a few individuals with 
large values of a character and compared their offspring with the general population. 
There is no security whatever that the individuals selected did not have excess of 
character owing to environmental and not to hereditary conditions. Such sweeping 
and dogmatic statements as those made by Jennings that : " The ' standard devia- 
tion ' and ' coefficient of variation ' express in a pure race mere temporary conditions 
of no consequence in heredity " merely beg the question of what a pure race may 
be, or else show a sad ignorance of what work has been done on this very point. 
The only manner in which it appears to me that an answer can be found to 
the question which Pearl and Jennings take as now settled — that is to say, the 
problem of whether or no selection within the individuals descended from a single 
unit can or cannot produce an effect — is to correlate ancestry with offspring in 
species in which self-fertilisation or parthenogenesis takes place. If this be done, 
then the relation of offspring to parent ought to be identical with the relation of 
* Nothing short of .500 mothers would really give a convincing result, I think, where the actual 
correlation is so extremely small. See Pearl, K. and Surface, F. M. : " Data on the Inheritance 
of Fecundity obtained from the Eecords of Egg Production of the Daughters of "200-Egg Hens'." 
Maine Agricultural Experiment Station, Bulletin, No. 160, 1909. 
