684 



SCIENCE. 



[N. S. Vol. XXIII. No. 592. 



needed research. This volume is divided 

 into six principal sections: 



I. The Introduction. — This section con- 

 tains a brief review of the leading theories 

 of evolution and a statement of his own 

 method of studying the problem. 



II. Elementary Species. — According to 

 de Vries the species of systematists are 

 not single groups of individuals inter- 

 grading among themselves and separated 

 by lacunae from neighboring groups, but 

 are aggregates of such units. He, there- 

 fore, speaks of 'systematic species' and 

 'elementary species,' an 'elementary spe- 

 cies' being one of the units that go to make 

 up the ' systematic species. ' I think it very 

 unfortunate that de Vries has introduced 

 the term 'elementary species' as opposed 

 to 'systematic species.' It clearly shows 

 that he is not in touch with refined modern 

 systematic work, for his 'elementary spe- 

 cies' is the systematic species of all modern 

 systematists with whose work I am fa- 

 miliar, while his 'systematic species' has 

 no status. However, his account of how 

 species established by the older systematists 

 are now being subdivided into smaller and 

 smaller units as our studies become more 

 detailed and more exact, is very interesting. 

 He discusses elementary species not only 

 in nature, but also among cultivated plants, 

 and gives valuable information regarding 

 the mode of their selection among the lat- 

 ter, especially wheat. 



III. Retrograde Varieties. — I think de 

 Vries again unfortunate in his attempt to 

 define a variety as an elementary species 

 that has lost a character. According to 

 the usually accepted definition of variety 

 in this country there is intergradation with 

 the typical form of the species: a variety 

 would be represented by a secondary mode 

 in the species curve. The inference from 

 de Vries 's definition would be that we can 

 not have retrogressive species. It is well 

 kno-rni in paleontology that not only single 



species, but whole groups of animals have 

 undergone retrogressive evolution— the 

 Baculites, for instance. In this section de 

 Vries brings out a very important distinc- 

 tion between two classes of phenomena that 

 have been characterized as atavism. Ileal 

 atavism is defined as the reappearance of 

 ancestral characters in a species that is 

 pure bred. The other phenomenon, and 

 the one that is usually designated atavism, 

 is considered false atavism by him, and the 

 name 'vicinism' applied to it. 'Vicinism' 

 is due to the crossing of species or varieties. 

 A plant the result of such a cross, though 

 apparently a pure strain, may produce off- 

 spring of the type of the ancestor that it 

 least resei^bles. This section contains 

 much information on crosses, including a 

 statement of Mendel's law. 



IV. Ever-sporting Varieties. — The term 

 'ever-sporting variety' is used for those 

 forms regularly propagated by seed, of 

 pure, not hybrid, origin, but which 'sport 

 in nearly every generation.' An interest- 

 ing account is given of how he tried to 

 obtain a pure striped and a pure red 

 variety from a striped variety of snap- 

 dragon, known as Antirrhinum majus lu- 

 teum ruhro-striatum, which sports into red 

 flowers, but he was not successful. The 

 seed of red individuals, or even seed from 

 self-pollinated red flowers in a raceme of 

 mostly striped flowers, would produce a 

 certain number of striped flowers. The 

 seed from striped individuals always gave 

 a percentage of red ones. It is interesting 

 to note that stocks producing double flowers 

 are grown from single-flowered individuals. 

 These experiments are so interesting, and 

 so suggestive of further research, that it is 

 a great temptation to give a detailed ac- 

 count of them, and also to describe the 

 experiments in trying to get a race of 

 five-leaved clover, and those with poly- 

 cephalic poppies and monstrosities. His 

 experiments with double adaptations are 



