Variation and Heredity 289 
Many other Linnzean species are in this respect like Drada 
verna, and most varieties, De Vries thinks, are really element- 
ary species. 
Second, the polymorphism due to intercrossing is the 
outcome of different combinations of hereditary qualities. 
There are here, De Vries says, two important classes of facts 
to be kept strictly apart, —scientific experiment, and the 
results of the gardener and of the cultivator. The experi- 
menter chooses for crossing, species as little variable as pos- 
sible; the gardener and cultivator on the other hand prefer 
to cross forms of which one at least is variable, because the 
variations may be transmitted to the hybrid, and in this way 
a new form be produced. 
New elementary characters arise in experiments in crossing 
only through variability, not through crossing itself. 
Third, variability in the ordinary sense, that is, individual 
variability, includes those differences between the individual 
organs that follow Quetelet’s theory of chance. This kind 
of variability is ‘characterized by its presence at all times, in 
all groups of individuals. 
De Vries recalls Galton’s apt comparison between variability 
and a polyhedron which can roll from one face to another. 
When it comes to rest on any particular face, it is in stable 
equilibrium. Small vibrations or disturbances may make it 
oscillate, but it returns always to the same face. These 
oscillations are like the fluctuating variations. A greater 
disturbance may cause the polyhedron to roll over on to a 
new face, where it comes to rest again, only showing the 
ever present fluctuations around its new centre. The new 
position corresponds to a mutation. It may appear from our 
familiarity with the great changes that we associate with the 
idea of discontinuous variability, that a mutation must also 
involve a considerable change. Such, however, De Vries says, 
is not the case. In fact, numerous mutations are smaller 
than the extremes of fluctuating variation. For example, the 
U 
