1916] Gates, — Mendelian Interpretation of Oenothera Crosses 199 



As the matter is one of some importance, involving as it does the 

 whole question of the interpretation of mutations in certain aspects, it 

 may be worth while to point out some of the difficulties which East has 

 failed to see. He will then have the opportunity of explaining them 

 on a Mendelian basis. 



It would be quite impossible in a limited space to discuss all the 

 classes of cases which do not conform to the Mendelian conception, 

 but a few of them, belonging to one class only, may be pointed out. 



In the first place let us consider what de Vries calls mutation crosses, 

 such as Oc. Lamarckiana X rubrincrvis or its reciprocal. In such 

 crosses the F 2 splits into the two parental types x and both breed true in 

 later generations. We are assured by some that this can be explained 

 as ordinary segregation, but for such an explanation the following 

 assumptions must be made: (1) that Lamarckiana is heterozygous for 

 the rubrincrvis character, (2) that it breeds true both before and after 

 the cross because the rubrincrvis germ cells either fail to develop or 

 fail to fertilize each other, (3) that about 50% of the Lamarckiana germ 

 cells are rubrincrvis in character, since rubrincrvis usually appears with 

 this frequency in crosses with Lamarckiana. Thus far the assump- 

 tions, though improbable are not impossible, and the fact that 

 Lamarckiana may show 50% or more of sterility leaves the interpre- 

 tation a loophole through which to crawl. 



Now let us go a step further. If other flowers on the same Lamarck- 

 iana plant used to cross with rubrincrvis are pollinated by nanclla 

 the dwarf mutant, the F] will again contain the parent forms Lamarck- 

 iana and nanclla in widely fluctuating percentages, and the same 

 result occurs in the reciprocal cross Oc. nanclla X Lamarckiana. We 

 must now apply the above Mendelian hypotheses mutatis mutandis 

 to these crosses, and assume that some 50% or more of our Lamarck- 

 iana germ cells are now nanclla. The same must be done for all the 

 other forms which show a similar behavior in crosses with Lamarckiana. 

 This is of course absurd, for it assumes that 50% of the Ljamarckiana 

 germ cells are at the same time nanclla, rubrincrvis, oblonga, etc. The 

 only way out of this difficulty that I can see is by the further assump- 

 tion that when one crosses Lamarckiana with nanclla pollen all the 

 rubrincrvis germ cells present obligingly disintegrate and disappear, 



1 The form of rubrincrvis derived from this cross has since been found to differ in certain 

 particulars and has been called subrobusla. Hut this does not alter the Interpretation of the 

 facts. 



