510 



SCIENCE 



[N. S. Vol. XXVII. No. 691 



the same parents is due to any of the causes 

 mentioned above. 



But variation as between individuals of the 

 same generation may be, to some extent, due 

 to what has been aptly termed " place efiect." 

 This is especially true when a species is trans- 

 ported to new environment, a good deal of 

 variation occurring which seems to be due to 

 change in environment, such variation may 

 partake partially of the nature of fluctuating 

 variation, but this is a special case which we 

 do not need to consider here. 



This leaves two distinct types of variation 

 due to wholly different causes which have 

 hitherto been more or less confused. By some 

 writers they have both been included under 

 " fluctuating variations." The first of them 

 is well exemplified in a field of corn, where 

 ordinarily hardly any two individuals are 

 alike. It is fully demonstrated by the work 

 of Nilsson, in Europe, and Shull, in this 

 country, which will be referred to below, that 

 by far the greater part of the variation of our 

 corn field is due to the fact that we have in 

 the various individuals almost countless com- 

 binations of Mendelian characters and that 

 these combinations change with each new gen- 

 eration. The same thing is true on a smaller 

 scale with all species that cross-fertilize under 

 any conditions, and even with close fertiliza- 

 tion this condition exists to a greater or less 

 degree. If we require a corn plant to close 

 fertilize, by this process we permit the forma- 

 tion of a few individuals which are perfectly 

 homozygote; i. e., the inheritance of the indi- 

 vidual from the two parents becomes exactly 

 alike. Then if we take the pains to seek out 

 these homozygotes and propagate them, allow- 

 ing no 'cross-fertilization, we completely elim- 

 inate the type of variation here referred to 

 and get forms that vary only in response to 

 the immediate environment. Variation due 

 to this re-combination of characters with each 

 generation would naturally show correlation 

 between parent and offspring, but when we 

 have eliminated this type of variation such 

 correlation would no longer exist. 



Species that never cross-fertilize or that do 

 so very rarely, and plants that are not allowed 



to do so, have been shown by careful study by 

 Nilsson, Shull, Hopkins and others to consist 

 of mixtures of strains which when separated 

 show no variation except that due to environ- 

 ment, or rather the larger part of the species 

 exhibits this condition, for even in such spe- 

 cies there may be a small admixture of hetero- 

 zygotes. 



Nilsson in Europe and Shull in this coun- 

 try have obtained these perfectly homozygote 

 individuals, Nilsson working with wheat and 

 many other species, and Shull with corn. The 

 individuals of a generation are as much alike 

 as identical twins. These are the so-called 

 elementary species, this term being applied 

 through a misconception of their nature. 

 (This subject will be fully discussed in 

 another place.) 



Finally we have variation due to the imme- 

 diate environment of the individual. For in- 

 stance, a variation in food supply may cause 

 two individuals having identically the same 

 inheritance to differ in size and in other char- 

 acteristics. 



Dr. Shull confines the term " fluctuating 

 variations " wholly to variations of the last- 

 mentioned type. He rigidly excludes varia- 

 tions due to the heterozygote nature of the 

 parents. I fully agree with him in this use 

 of the term. When we so limit it, we may 

 then say that fluctuating variations are not 

 transmitted. W. J. Spillman 



U. S. Depaetment of Ageicultuke 



SPECIAL ARTICLES 



THE DERIVATION OF FECHNEe's LAW 



About eighty years ago E. H. Weber ob- 

 served that the least perceptible increment to 

 a stimulus affecting several of the sense 

 organs, under fixed subjective conditions of 

 attention, expectation and fatigue, bore a 

 definite relation to the amount of that 

 stimulus. 



G. T. Fechner, thirty years later, extended 

 Weber's observations and formulated his re- 

 sults in mathematical terms. Calling the 

 stimulus L and the least perceptible increment 

 hL, then Fechner's statement of what he 

 termed Weber's Law is that 8L/L = constant. 



