198 



SCIENCE. 



[N. S. Vol. VII. No. 163. 



ural selection to eliminate either the orig- 

 inal type or the new variety. It would, 

 moreover, be illogical to presume that the 

 same agent that has been instrumental in 

 causing the disappearance of the pigment 

 in the natural fish is the same agent, in the 

 same locality, and under apparently the 

 same conditions, that is instrumental in 

 again producing pigment on the lower side. 



A Precise Criterion of Species. C. B. Daven- 

 port. 



In order to decide whether two allied 

 groups are species or varieties it is neces- 

 sary first to give an exact quantitative ex- 

 pression to the two best criteria of species 

 — divergence and segregation —by the use 

 of the modei-n mathematical method of 

 studying individual variation. Divergence 

 is the distance between the modes of the 

 two groups in question expressed in units 

 of the average deviation from the mean of 

 the individuals of one of the groups. Segre- 

 gation is inversely proportional to the num- 

 ber of intergrades, or it is the height of the 

 lowest ordinate between the two modes ex- 

 pressed in units of the height of one of the 

 modes. 



An examination of the usage of systema- 

 tists will tell us what least degree of diver- 

 gence or segregation is usually expected in 

 distinct species. Leaving the precise de- 

 termination of this least degree still unde- 

 cided, we may conclude : A group of allied 

 individuals giving a dimorphic curve of 

 a (differential) character consists of two 

 species either when the minimum between 

 the two groups is m [20] % or less, of the 

 shorter mode, or when their modes diverge 

 by n [10?] or more times the smaller aver- 

 age deviation of the two subgroups. Other- 

 wise the dimorphic curve indicates two 

 races. 



Certain Results from a Study of the Variation 

 of Liitorina. H. C. Bumpus. 

 The critical study of variation in 100 



specimens of N"ecturus revealed the fact 

 that those individuals which were abnormal 

 so far as location of the pelvis was con- 

 cerned were also abnormal in respect to 

 many others, and only remotely related 

 characters ; that those individuals which 

 were unstable in respect to one character 

 tended towards instability in respect to all 

 characters ; that variation of one organ was 

 an indication of probable variation of other, 

 if not of all, organs. The examination of 

 over 1,700 sparrows' eggs encouraged the 

 belief that this principle of the general in- 

 stability of variants might be of further ap- 

 plication, and in a recent article by Have- 

 lock Ellis, on Genius and Stature, it is 

 concluded, on anthropological data alone, 

 that those variations of mind which have 

 been instrumental in producing men of emi- 

 nence are accompanied by striking and re- 

 mote physical variations ; that those who 

 have attained distinction as warriors, states- 

 men, scientists or writers have generally 

 been above or below the mean of stature. 



The speaker then exhibited 1,000 shells 

 of Littorina littorea which had been ar- 

 ranged in order, according to their shape, 

 from extreme elongation to extreme ventri- 

 cosity, and showed that those at the ex- 

 tremes tended toward excessive variation in 

 weight. Both ventricose and elongated 

 shells were far heavier and lighter than the 

 species of more ordinary form. 



Grafting Experiments on Tadpoles, with Special 

 Reference to the Study of the Gh'owth and Re- 

 generation of the Tail. R. G. Harrison. 

 The method of grafting amphibian larvis, 

 as described by Born in his exceedingly 

 original and suggestive paper [Archiv f. 

 Entwicklungsmechanik, Bd. 4], may be ap- 

 plied to the study of the normal growth 

 of the embryo. Thus, when portions of 

 larvae of Rana virescens and R. palustris 

 are combined in various ways to form a 

 complete normal organism, the sharp con- 



