280 REPORT OF THE CONFERENCE ON GENETICS 
result from impure segregation in the second generation onwards—depart- 
ing from the true Mendelian system. I have, however, observed similar 
graduations among Primula, Verbena, and Beet hybrids of the first 
generation ; yet these cases do not yet appear to me to be sufficiently clear 
to justify the publication of further details as yet. The production of 
many types (Pleiotypie) even in the first generation, with constancy of 
each single form, constitutes the characteristic of the Hieraciwm hybrids, 
according to Mendel, as well as of the so-called Macfarlane hybrids 
generally, and, moreover, among others, of the G’nothera-mutants of de 
Vries.* He regards, as is well known, such a condition as a proof that the 
differences among these forms or characters are specific, considering a 
Mendelian behaviour as an expression of merely racial difference. 
At this point those cases of crossing should be mentioned (such as 
Triticum vulgare x Tr. polonicum) in which the splitting (Spaltwng) at 
the second generation intensifies one of the parental characters such as 
the length of ear of one parent, and permits the appearance of a long 
series of gradations in that respect; but according to my investigations 
to date it produces no, or almost no, absolutely pure representative of the 
other parental character, e.g. the ear-length of Triticum vulgare.t The 
same is true of the close packing of the ears in crosses between square- 
headed forms of wheat and narrow-eared forms. Even in the following 
generations the one form of parent never again comes out pure, that is, 
free from a more or less distinct trace of the other.t 
Among the most significant results of hybridisation considered as a 
source of progress in evolution is the mcdern production of really new 
forms which cannot be regarded as mere combinations of characters 
already visible in the two parents. These cases should be spoken of as 
hybrid-mutations. 
Some years ago I was able to state that in not a few cases such 
novelties produced by crossing have a regular behaviour and exhibit 
Mendelian ratios. For example, as my experiments have shown—with 
which those of Bateson and Saunders are in complete accord—there are 
certain races of peas (Piswm arvense), beans, stocks, and barley which 
when in-bred remain perfectly pure, but on crossing with another strain, 
chosen almost at random, give rise to new characters. For such forms 
I have proposed the term cryptomeric. In the cases I studied the two 
parents do not seem to take an equal share in the production of the new 
form ; for the possession of the hidden character by one of the parents 
reveals itself in spontaneous variation. The other parent thus plays the 
part of “activator” or complement. The Mendelian ratios conform to 
the conclusion first promulgated by Correns, that in these cases two pairs 
of characters are concerned, namely, the possession or want of the 
character and of the activator. These regular hybrid-mutations are thus to 
be described as degressive or retrogressive in de Vries’ sense ; since there 
is a loss or a gain on one side only, and not a variation in several respects, 
* Extreme cases (monotypic constancy, monolepsis) of this kind are presented by 
the false hybrids of Millardet. 
+ Similarly, according to Biffen’s communication to this Conference, from the 
cross Tr. vulgare x Tr. dicoccwm no pure dicoccum reappears. 
t¢ This case must be carefully distinguished from that of general dominance or 
partial dominance, and from the phenomena seen in the false hybrids of Millardet, 
