Progressive, Retrogressive, Degressive Mutations. 571 
so we can imagine that a new character remains latent 
for some time after its first origin, its phylogenetic birth, 
if I may so express it, and does not become active until 
after the lapse of a lesser or greater period of time. 
According to this view, every progressive mutation 
is fundamentally a double process, and consists in the 
production of a new internal factor and in its activation. 
Both may sometimes happen at the same time, but this- 
is not necessary. It is therefore desirable to apply sepa- 
rate names to the two processes ; the internal one I have 
called premutation, the externally visible one a mutation 
sensu strict o. 
The premutation is therefore of a hypothetical, the 
mutation, however, of an empirical nature. 
It further follows from this that an internal factor 
does not of itself lead to the origin of an external change. 
As in ontogeny so also in phylogeny an elementary char- 
acter can be sometimes active, but at other times latent 
or inactive. If a new character emerges from its original 
latent conditions and becomes active, we call the process 
a progressive mutation ; conversely we can denote its 
return from the active to the latent condition as a retro- 
gressive mutation. 
The experience of gardeners and of the systematists 
with the smaller species and varieties teaches that retro- 
gressive mutations of this kind are common phenomena. 
Almost any character may disappear. This applies not 
only to the superficial characters, such as color, hairs, or 
thorns, but also to those deeper ones which affect the 
inner organization of the plant, such as the decussate ar- 
rangement of leaves, and even the symmetry of the organ- 
ism. Spiral torsion and peloria show how profound an 
alteration in the appearance of a plant or in the structure 
