18 DAVIS— THE TEST OF A [April 23, 



Upon students of this genus rests the responsibihty of obtaining 

 data on this steriHty and, if possible, of discovering its causes. The 

 assumption that a hne represents a pure species because it breeds 

 true is not a safe foundation upon which to conduct experimenta- 

 tion in the Oenotheras. This is the assumption upon which have 

 been based many of the conclusions of the Mutationists, and from it 

 we must dissent. We cannot depart from the principles underlying 

 Mendelian methods of research which have so brilliantly opened the 

 present century of biological investigation. 



Finally what are the tests that must be applied to an CEnothera 

 species to determine whether or not it is pure. 



First. — There is the breeding test and that must be applied with 

 such experimental methods of seed germination (Davis, '15) as will 

 insure a complete progeny from the sowing, a progeny wholly repre- 

 sentative of all types of viable seeds. Even then the breeding test 

 is negative rather than affirmative in its conclusions. Should the 

 form throw off numerous variants it naturally becomes a subject of 

 suspicion, but should it breed true or relatively true that does not in 

 this group of plants prove it to be homozygous in its germinal 

 constitution. 



Second. — Information must be obtained on the character and 

 degree of sterility present, both gametic and zygotic. Sterility, 

 unless shown to be strictly physiological in its character, suggests 

 genetic impurity. 



T/tzVcf.— Cross-breeding tests must be planned and followed in 

 which the form under observation is mated with material of known 

 genetic purity. If the hybrid plants of the first generation are 

 essentially uniform and the result of a normal germination of the 

 seeds the indications are strong that the form is truly pure provided 

 that the gametes are likewise normally fertile. If the hybrids of 

 the first generation fall sharply into classes the material must develop 

 gametes of different germinal constitutions and is consequently 

 heterozygous. One favorable cross with a pure species may not be 

 sufficient to establish the purity of a form ; a number of favorable 

 tests with pure types will carry increasing conviction. 



It is thus not an easy matter to determine the fact whether or 

 not a species of CEnothera is pure, and yet this is fundamental to 



