424 BOTANICAL GAZETTE [DECEMBER 
My evidence is appropriate, however, to the question at issue. Irre- 
spective of the individual parentage of the plants, it shows that the 
distinction between wide and narrow, or continuous and discontinuous, 
variations is artificial. That these aberrant forms should be the result 
of several generations tending in the same direction would be incom- 
prehensible in view of the sterility of some of the forms and partial 
sterility of others; and would itself be contradictory to DE VRIES’S 
idea that new forms of plants arise suddenly, without preparation or 
intermediate steps. 
The a priori objection to really discontinuous variation is the 
impossibility of really discontinuous development. Every organism 
that varies grows, and varies only as it grows. All organisms of any 
kind are indistinguishable during a considerable part of their develop- 
ment, but sooner or later their individual differences appear and 
become fixed. The tendency of heredity, as the conservative factor 
in both evolution and development, we believe is to postpone the 
appearance of deviations from the parent types. If they appear very 
late, the variations will be very small; if they appear earlier, they will 
obviously be more notable. If variations in growth appear much 
earlier than usual, the variation will be unusually profound. But it 
must be evident to anybody that it is not possible to select any point 
within the range of known deviation in the development of any organ- 
ism whatever, and to say that the differences which occur before this 
time are different in kind from those which appear at and subsequent 
to it. 
Variation, when it is just appearing, is a phenomenon involving 
small and homogeneous groups of cells; or, regarded in finer detail, 
single cells. When variation occurs it is by the unit of the varying 
structure. If it occurs early, the subsequent development of the eit 
can make it become very conspicuous; but the variation is when It #5, 
irrespective of later growth based on it. Stomata and trichomes are 4S 
a rule formed late in development, and the presence of two where one 
is normal is likely to escape our attention, as is the presence of an 
extra leaf on a tree; cotyledons are formed earlier, and an € st 
cotyledon, perhaps involving an unusual form of the whole plant seh 
grows from the seedling, is an object of interest and remark. : 
when the first step toward the formation of the extra cotyledon W 
