896 



SCIENCE 



[N. S. Vol. XXVII. No. 701 



Administration des Arcliives Italiennes de 

 Biologie, Via Acquarone, Genova, Italy. 



Charles S. Minot 

 Habvaed Medical School, 

 May 29, 1908 



SPECIAL ARTICLES 



AN INTERPRETATION OF ELEMENTARY SPECIES 



The original idea which led to the develop- 

 ment of the theory of so-called elementary 

 species is found in Darwin's gemmules. Ex- 

 istence of these gemmules was proposed to 

 explain the supposed transmission of acquired 

 characters. Weismann, acting on Darwin's 

 idea as a suggestion, developed a very elah- 

 orate theory of heredity. To consider the 

 relation of Weismann's philosophy to the 

 subject in hand would take us too far from 

 our present object, though this relation is 

 important. De Vries, going directly back to 

 Darwin and doing away with that part of 

 Darwin's theory which postulated the migra- 

 tion of gemmules of the various cells of the 

 body to the germ cells and assuming that the 

 germ plasm is composed of these gemmules — 

 or as de Vries calls them, pangens, has devel- 

 oped a very elaborate theory, not only of 

 heredity, but also of evolution, based on the 

 assumption that the individual is merely an 

 assemblage of parts, each of which constitutes 

 an hereditary character and each of which 

 develops from a particular pangens in the 

 original germ plasm of the fertilized egg. 

 He conceives a definite species to be made up 

 of a definite number of these hereditary char- 

 acters. The addition of a new kind of pan- 

 gens to the germ plasm causes the developed 

 organisin to differ more or less from other 

 individuals which preceded it. If this differ- 

 ence relates to a single pangens, then the new 

 and modified form of the organism is looked 

 upon as an elementary species. It differs 

 from its congeners by an elementary differ- 

 ence. The ordinary species may contain with- 

 in it a large number of elementary species, 

 each differing from those nearest related to it 

 by the possession of a single pangens not pos- 

 sessed by its nearest relatives. 



The work of Nilsson in Europe and of 



Shull in this country have been considered as 

 strengthening the idea of elementary species. 

 Nilsson has been able to obtain varieties of 

 wheat and other plants that may be assumed 

 to be absolutely uniform except for such dif- 

 ferences as are caused by environment. Some 

 of the distinct strains differ very little, but 

 this difference is absolutely constant, and the 

 different individuals within one of the ele- 

 mentary species are as like each other as so- 

 called identical twins. They offer no further 

 chance of improvement by selection. Shull 

 has, in like manner, obtained supposedly ele- 

 mentary species of corn which breed true, the 

 various individuals of a given strain being as 

 much alike as identical twins. He was led 

 to look upon a corn field as simply a hetero- 

 geneous collection of these elementary species 

 and hybrids between them. 



These so-called elementary species can easily 

 be accounted for on the old Darwinian idea 

 of gradual evolution, as will be shown below. 

 They are, therefore, in no wise a confirmation 

 of the pangens theory of de Vries. The dem- 

 onstration is as follows: Let A, Table I., rep- 

 resent a Mendelian character which is more 

 or less variable in the different individuals in 

 which it appears, these differences being hered- 

 itary. Let B and C represent other Men- 

 delian characters similarly variable. The 

 variations in these characters may have come 

 about gradually, as Darwin supposed variation 

 to occur, or they may have come about in any 

 other manner. Suppose A^ represents the first 

 character as it appears in a particular homo- 

 zygous individual. A^ may represent this 

 same character in another homozygous indi- 

 vidual, the difference between A^ and A^ being 

 so slight as not to be certainly discernible. 

 In like manner A' differs from A' so slightly 

 that the two can not be certainly distin- 

 guished, but A^ differs from A^ suiEeiently to 

 be distinguished. So with the other A's. 

 Any one of them in the series from 4' to A^" 

 differs so slightly from adjacent A's as not to 

 be certainly distinguishable from them, but 

 may be distinguished with more and more 

 certainty as we recede from the selected A 

 in the series. The exponents of jB and C have 



