704 
GENETICS: B. M. DAVIS 
A CRITICISM OF THE EVIDENCE FOR THE MUTATION THEORY 
OF DE VRIES FROM THE BEHAVIOR OF SPECIES OF 
OENOTHERA IN CROSSES AND IN 
SELFED LINES 
By Bradley Moore Davis 
DEPARTMENT OF BOTANY. UNIVERSITY OF PENNSYLVANIA 
Communicated by R. Pearl. October 31. 1917. Read before the Academy. November 20. !917 
The mutation theory of Professor De Vries rests so very largely upon 
deductions from his studies on species of Oenothera that any discus- 
sion of it naturally centers upon the interpretation of the behavior of 
these plants when selfed (in-bred) and in their crosses with one another. 
Of these species Oenothera Lamarckiana stands as the form most thor- 
oughly studied with respect to its habit of throwing off in successive 
generations variants with numerous distinguishing characters of such a 
nature that they can with certainty be separated and would rank in 
systematic botany as clearly defined new species arising suddenly and 
fully formed from the parent type. Professor De Vries calls these 
variants 'mutants' and interprets their appearance as the spontaneous 
origin by mutation of new species from a plant, Oenothera Lamarckiana, 
which he believes to be representative of a pure species. De Vries is 
not willing to allow that in 0. Lamarckiana this phenomenon may be 
the direct result of an impure or hybrid constitution. The behavior of 
O. Lamarckiana and certain other forms in this genus is, therefore, to 
De Vries direct experimental evidence of the origin of new species 
through wide and discontinuous variations, the result of spontaneous 
internal manifestations within the parent plants. De Vries further 
believes that mutations play a very important part in organic evolution 
and that they largely supply the material, i.e., the variations, upon 
which natural selection can operate. 
There is no question of the facts as brought out in the extensive and 
patient work of De Vries; they have been repeatedly confirmed. Oeno- 
thera Lamarckiana if grown in sufficiently large cultures may be expected 
to produce in each generation approximately the same set of 'mutants.' 
The proportions differ but they are apparently fairly stable for each 
variant; some make up about 1-2 per cent of the cultures, others are 
much less common. Certain of the 'mutants' breed fairly true when 
selfed while some are more unstable than the parent Lamarckiana. 
A significant feature of this performance is the clear expression of order 
and system in the appearance of precisely the same types through suc- 
cessive generations and we have no reason to suppose that 0. Lamarck- 
