8 
BULLETIN OF WISCONSIN NATURAL HISTORY SOCIETY. VOL. 5, NO. 1. 
would be only an improved race, i. e., merely a better kind of the 
same species — a kind that remains better only so long as cultivated 
under favored conditions, reverting to the common level as soon 
as left to itself under normal conditions. 
To remove any doubt as to these points, let me cite a few pas- 
sages from De V ries : 
(1) "I designate as mutation-theory the proposition that the 
characters of organisms are built up of units sharply differentiated 
one from another. These units may be combined in groups, and 
the same units and groups occur in allied species. Transitions, 
such as we so often see in the external forms of plants and 
animals, do not exist between these units any more than between 
the molecules of chemistry." Vol. i, p. 3. 
(2) "In the field of the doctrine of descent, this principle leads 
to the conviction that species have not arisen one from another 
by Hozving transitions, but by distinct steps. Every new unit added 
to the older units, forms a step, and separates the new form, as 
an independent species, sharply and completely from the species 
out of which it arose. The new species thus comes into being at 
a single stroke. It arises from the earlier form zvithout any visible 
preparation, without transitions." 
(3) "The mutation-theory, in my opinion, is supreme not 
only in the doctrine of the origin of species, but also in the whole 
field of hybridisation. Here it conducts us to the principle, that 
not the species, but the simple specific characters, the so-called 
elements of the species, are the units with zvhich zve deal in hybrid- 
isation" 
(4) "Every character, it is true, arises from one already 
present, not however from its normal variation, but through a 
change which, however small, is yet sudden. Provisionally, this 
change may be most simply likened to a chemical substitution." 
In his preface, p. IV., as follows : 
(5) "These jumps, or mutations, of which sport- variations 
are the best-known examples, form a special part of the subject 
of variation. They occur zvithout transitions, and are relatively 
rare, while ordinary variations are continuous and always present. 
(6) "The whole subject of variation thus divides into two 
parts, one of which deals with the ever present, individual, or 
fluctuating variation, the other, with mutation. Phenomena of the 
first class obey the well-known laws of probability and depend 
