576 Species According to the Theory of Mutation. 
the transposition of an internal character; from being 
latent it becomes active; from semi-latent, semi-active; 
and so on. If new factors are becoming active for the 
first time after having been latent through a shorter or 
longer series of ancestors, we speak of progressive muta- 
tions. If the active characters again become latent, the 
process is a retrogressive one. In all other cases it is 
degressive. 
The phenomena of hybridization find a ready ex- 
planation in the principles derived, in the first part of this 
volume, from our consideration of the origin of species 
and varieties. There are two main types of crosses, the 
bi-sexual or Mendelian and the uni-sexual. The former 
conform to the laws of segregation, they lead to various 
combinations of elementary characters, and thus can lead 
to the origin of as many new races as the number of pos- 
sible combinations indicates. These races are constant; 
the hybrids, however, always exhibit segregation in the 
formation of their sexual cells and sometimes even in the 
formation of buds. The hybrids of uni-sexual crosses 
on the other hand are constant ; so far as my experience 
goes, they do not segregate. If they are fertile they are, 
as a rule, as true from seed as their parents; but they 
may inherit the inconstancy of these (if, for instance, 
one of these belonged to an eversporting variety) and 
transmit it to their posterity. 
A strong body of facts, which have been given else- 
where lead to the conclusion that crosses follow MENDEI/S 
laws if one of the two parents stands in the relation to the 
other of having arisen from it by retrogressive or de- 
gressive mutation. This means that the two parents of 
the cross possess exactly the same internal elementary 
characters; but that one or more of these occur in dif- 
