May 30, 1913] 



SCIENCE 



821 



any color. These two things may for the 

 moment be called "A" and "B." Both 

 A and B must obviously have been present 

 in the original purple race from which the 

 two white races are descended. One race 

 has at some time in its past history lost A, 

 the other B; and each loss has produced a 

 specific type of white race which breeds 

 true. By crossing the two A and B are 

 again brought together, thus restoring the 

 original combination AB; hence the "re- 

 version" to the purple wild type. Now, 

 the things which we have called "A" and 

 "B" may very well be different chemical 

 substances. If we assume them to be borne 

 by different chromosomes, brought together 

 in the hybrid, the whole matter becomes at 

 once clear and simple. It seems probable 

 that all kinds of reversion may be ex- 

 plained by the same principle. 



The second case is that of "criss-cross" 

 heredity in the short- and long-winged 

 flies, where the sons are like their mother, 

 the daughters like their father. The ex- 

 planation of this case is less easy to follow 

 than that of reversion, but is more specific. 

 This case, and many others of similar type, 

 may be completely explained through our 

 knowledge of the relation of the chromo- 

 somes to sex. These flies agree with the 

 general rule already referred to, that the 

 males contain a single X-chromosome, or 

 sex-chromosome, the females two. AU the 

 facts revealed by experiment are very 

 simply and completely accounted for by 

 the single assumption that the X-chromo- 

 some is responsible not only for sex, but also 

 for the short-winged character. Specific- 

 ally, the assumption is that the short wings 

 are due to the lack or defect of something 

 (let us again say some definite chemical 

 substance) that is contained in the X- 

 chromosome. Let us see just how this 

 works out. We may write the formula for 

 the short-winged female as xx (the small 



letters indicating the defect in X that 

 is responsible for defective wing develop- 

 ment), while that of the normal (long- 

 winged) male is XO. Such a female pro- 

 duces eggs of only one type, x, while the 

 normal male produces sperms of the two 

 types X and no X or 0. Fertilization thus 

 can only give rise to the two combinations 

 xX. and xO, the former being females, the 

 latter males. The males are short-winged 

 because they contain only the defective x. 

 The females likewise contain such an x, 

 but they are nevertheless long-winged be- 

 cause they also contain a normal X, which 

 is sufficient to ensure normal wing develop- 

 ment. It follows that the daughters are 

 long-winged like their father, the sons 

 short-winged like their mother. I need not 

 trace this explanation further into its de- 

 tails. It is enough to say that upon this 

 one assumption the results of many other 

 kinds of crosses among these flies work out 

 perfectly; and a similar explanation will 

 completely account for other cases of criss- 

 cross fertilization, for such curious phe- 

 nomena as the heredity of color-blindness 

 in man, and many other cases in which 

 particular somatic characters are linked 

 with sex in a definite way. These cases 

 offer, indeed, a brilliant example of the 

 clearness and simplicity of the causal ex- 

 planations that microscopical research has 

 helped to give of complicated special phe- 

 nomena of heredity. 



As a third and last example I select an 

 even more interesting and instructive case, 

 the complete analysis of which carries us to 

 the firing line of research in this field. It 

 illustrates the influence of the grand- 

 parents upon the combinations of unit- 

 characters seen in the grandchildren. It 

 is a very curious fact, only recently discov- 

 ered, that in certain cases hybrids of iden- 

 tical composition exhibit marked differ- 

 ences in their output of offspring that can 



