1 4 
A. D. Darbishire. 
think in terms of masses, the statement that ancestry does not 
make a difference is demonstrably false. 
I ought to have said that my experiment demonstrated that, in 
the particular case dealt with, the characters of a given generation 
were determined by the potentialities existent in the germ-cells of 
their parents (as according to the Weismanno-Mendelian view); and 
not by the somatic characters of their parents and remoter 
ancestors (as according to the contributional view). 
In a word, my “ theory of ancestral contributions” is a term 
which I have coined in order to be able to refer briefly to the generally 
accepted view of the relation between successive generations of 
organisms, which the Weismanno-Mendelian view is gradually 
superseding. And a Anal demonstration of its falsity or correctness 
will not affect the Law of Ancestral Heredity as formulated by 
Professor Pearson one way or the other ; though a demonstration 
of its truth would profoundly modify our attitude to the commonly 
accepted interpretation of Mendelian hereditary phenomena. 
Probable Error. 
Attention to the exact numerical proportions, in which the 
various types dealt with in breeding experiments occur, has played 
a very important part in the progress of our knowledge of heredity. 
It is perhaps going too far to suggest that it was neglect to pay any 
attention to numerical proportions which prevented John Goss (who, 
as we have seen, observed the essential Mendelian phenomena) from 
enunciating the interpretation of them which we now associate 
with Mendel’s name ; but it is certain that an accurate record of 
these proportions was one of the conditions essential to the discovery 
which Mendel made. 
In my fifth lecture 1 pointed out that many cases which, when 
first observed, were regarded as instances of the 1:2:1 ratio have 
been subsequently regarded as instances of the 3:9:4 ratio. Here 
again close attention to the exact numerical proportions themselves 
is necessary; for on it depends the answer to the question to which 
of these two types of ratio the particular case of segregation under 
observation belongs. Again, those who maintain that the theory of 
ancestral contributions holds good within the limits of a single unit- 
character would probably not contend that, for instance, the 
diminution of the proportion of the recessives in successive 
generations, as we proceed from the first cross, would be any more 
than a very small one in each successive generation ; but they 
