370 
De Vries on the Origin of Species 
which exhibit this deviation D, and that the ancestry of these parents is supposed 
to be an average sample of the whole antecedent generations, or to have zero 
deviation. If not only the parents, but the grandparents exhibit deviation B, it is 
clear from Natural Inheritance, pp. 134 — 137 (" Separate Contribution of Each 
Ancestor ") that the mean deviation of the offspring will be ??iore than ; while 
with increase in the number of generations during which the ancestors have 
exhibited this deviation the ratio between the mean filial deviation and D 
continually approaches unity. This is stated still more clearly in the memoir 
on Basset Hounds*. Professor de Vries, however, regards the ratio between 
parental and filial deviation as fixed, so that the only ancestors whose peculiarities 
directly affect the individuals of a generation are their immediate parents. Mr 
Galton's view of the effect of regression follows inevitably from the general theory 
of chance, if we regard the character of an individual as a phenomenon due to a 
series of complex groups of causes, among which are the characters of each ancestor. 
The view which Professor de Vries implicitly adopts, that the characters of re- 
mote grandparents are of no effect except indirectly by determining the characters 
of parents, will not commend itself to naturalists as in accord with experience. 
The statements as to the character of regression among the offspring of 
" mutations " are also unsupported by anything like satisfactory proof The view 
held is clearljr set forth in many parts of the work, but especially in Die Lehre von 
der einseitigen Steigerung von Variabilitdt durch Auslese, pp. 416—422. In this 
section many cases, which are commonly adduced as evidence of the production of 
stable races by selection of variations, are treated as examples of mutation ; and 
the treatment brings out clearly the nature of the conceptions involved. An 
example cited is that of Anemone coronaria, quoted also by Darwin. The 
Rev. W. Williamson, after cultivating this plant for some years, found an individual 
with a single small additional petal. Among the offspring of this plant more 
supernumerary petals appeared; and by continued selection during some years, a 
" double " variety was established. On such cases the following comment is offered 
(pp.419 — 420): "Eine solche Verbesserung geschieht, wenn sie einmal moglich ist, 
rasch und mit zunehmender Geschwindigkeit. Daher die Vorstellung von der 
zunehmenden Variabilitat. Die Erklarung liegt aber einfach darin, dass man, wie 
im vorigen Paragraphen ei'ortert wurde, in Bezug auf das neue Material anfanglich 
Minus- Varianten findet, welche, sobald sie isolirt sind, im Folge des Regressions- 
gesetzes, sich nicht dem Merkmal der Art, sondern dem Mittelwerth der neuen 
Varietat nahern." That is to say, the original Anemone coronaria which presented 
an extra petal had undergone a change, which involved not only the obvious 
structure of one abnormal stamen, but the whole power of hereditary transmission. 
It and its offspring belonged thenceforth to a new '"' elementary species," and 
therefore exhibited regression to a new mean, involving the possession of double 
flowers. The only thing which a cultivator can do towards producing such a 
double flowered species is to watch for the first appearance of a mutation, and if 
* R. S. Proc. Vol. LXi. p. 403. 
