654 



SCIENCE 



[N. S. Vol. XXVI. No. 672 



when we come to estimate the minimum 

 number of transmitting elements, it is su- 

 perfluous to postulate additional elements 

 as instruments in eifecting these alterations 

 in pattern, seeing that the change may very 

 readily be imagined as due to a series of 

 quantitative subtractions from the quali- 

 tative elements. If then we can thus re- 

 gard the distribution of color as dependent 

 on subtraction-stages of some one element, 

 say the chromogen, we are naturally led to 

 refer the various intensities to another sim- 

 ilar but also definite series of subtraction- 

 stages in which the subtraction is spread 

 over the whole field, and so on for the other 

 qualities. 



Two fairly distinct classes of difference 

 may thus be presumed to exist, those de- 

 pending on the qualitative elements and 

 those due to quantitative subtractions from 

 them. The latter may be again subdivided. 



It is scarcely necessary at this time to 

 repeat that almost all the subtraction- 

 stages fully studied are fairly definite, and 

 their existence implies no suggestion of 

 general failure of segregation. Interesting 

 experiments have recently been made by 

 Castle and McCurdy, exhibiting positive 

 results of selection inside the limits of one 

 of these stages, viz., the so-called hooded 

 type in the rat. Nevertheless, the maxi- 

 mum result attributable to selection in such 

 eases is a modification within the limits of 

 one particular varietal type. 



Siich evidence provides no escape from 

 the conclusion that each genetic variety 

 comes into existence by a special addition 

 to or s.ubtraction from the genetic equip- 

 ment. 



Of all the results to which experimental 

 work has led us, that which to me is the 

 most astonishing is the fact that the same 

 systems of transmission should be followed 

 by characters which, by whatever test they 

 are judged, must be supposed to be most 

 diverse in physiological causation. Natu- 



rally when we are dealing with changes in 

 color, for instance, or in the reserve ma- 

 terials of a seed, we surmise that the crit- 

 ical factor is a certain ferment, or rather, 

 the power to produce that certain ferment. 

 It is perhaps not too wide a stretch of im- 

 agination to regard susceptibility to fun- 

 goid disease as caused by some similar 

 body. The diversity of these ferments 

 must anyhow be very great, and it seems 

 very strange that all these multifarious po- 

 tentialities should exhibit gametic allelo- 

 morphism. Let us take an illustration. 

 Color, as we can prove in regard to several 

 plants, and in regard to the plumage of 

 fowls, is due to the meeting of two com- 

 plementary factors. One is presumably a 

 ferment. Recent research strongly sug- 

 gests that it is a tyrosinase. The other is 

 referred to as a chromogen. But whatever 

 they are, the two bodies, or rather the fac- 

 tors which produce them, must be of ut- 

 terly different nature, and yet, genetically, 

 the two potentialities are treated similarly. 

 Each is allelomorphic to the absence of 

 such a power. 



How much more astounding is it, that 

 when we pass to qualities such as length 

 of stalk and shape of flower, or of a cock's 

 comb, the quality of the hair in rabbit, we 

 still find the same rules in strict and un- 

 deviating operation. Any scheme of hered- 

 ity on a scale comprehensive enough to 

 deserve the title of theory must deal with 

 this surprising fact. 



There is another extraordinary feature 

 in the behavior of allelomorphs which, 

 though Imown clearly in a few cases only, 

 must certainly play a great part in the 

 fuller elucidation of heredity. This is 

 partial gametic coupling. 



Mr. Punnett and I have for some time 

 been engaged in studying this phenomenon 

 in the sweet pea {Lathy rus odoratus) and 

 we have recognized indications of the same 

 thing elsewhere. The section will perhaps 



