AUGUST, 1915.} THE ORCHID REVIEW. 229 
variation and reversion is associated with details introduced by the sexual 
process or with the phenomena that precede it. 
When gametes produced by distinct species fuse together the result is a 
hybrid, in which the characters of the two parents are combined, often 
forming a more or less intermediate blend. It matters not—as a general 
rule—which parent was the seed bearer, for in the majority of cases 
reciprocal crosses are substantially identical, as also are successive crosses 
obtained between normal individuals of the same two species. Such 
hybrids are termed primary hybrids, and their behaviour indicates clearly 
that the mere act of crossing is not the cause of variation. 
But when primary hybrids are self-fertilised or recrossed with their 
parents, or with something else, the result is usually a batch of hybrids 
showing great diversity, and this variety indicates a want of uniformity in 
the characters of the gametes from which they were derived, and shows 
that something has happened in the interval between the origination of the 
hybrid and the formation of its reproductive cells. This condition of things 
either indicates an imperfect blending of the original conjugating nuclei or 
a separation of characters when the hybrid comes to form its own repro- 
ductive cells, which are known collectively as the germ plasm. The former 
is almost certainly the case, though either would indicate incompatability 
between characters derived from distinct ancestries. Incompatability 
would manifest itself at once, and would naturally be reflected in the 
teduction division. Indeed, incompatability can often be witnessed, when 
characters and tissues from diverse ancestries are developed side by side, 
thus demonstrating the well-known phrase that a hybrid is a mosaic. 
A complete blending of character would result in a batch of uniform 
secondary hybrids, and Mendel himself appreciated the fact when he pointed 
out that hybrids in which the diverse elements were permanently accom- 
modated together reproduced themselves true from seed, and had all the 
attributes of species. It is often assumed that it is in the reduction 
division that the redistribution of characters takes place, and that it is 
governed by nothing stronger than the law of chance. If the former 
condition has no stronger element of probability than the second we should, 
in the absence of positive evidence, unhesitatingly reject it. Nature under- 
stands her own business better than that. Hybridisation is not a new 
Process, it merely unites gametes of diverse instead of the same ancestry, 
and all its processes are otherwise identical. ; 
Another point that tells in favour of this contention is the substantial 
identity of the phenomena indicated with the separation of mixed character 
seen in sports and in graft hybrids, in which the aid of a reduction division 
Cannot be invoked. Sports are almost invariably seen in plants of mixed 
ancestry—hybrids or hybrid derivatives—while graft hybrids combine 
