May 30, 1913] 



SCIENCE 



815 



explanation can be offered of the combina- 

 tions and recombinations of parental or 

 grandparental characters that appear in 

 definite numerical proportions in hybrids? 

 We may glance at a few particular cases 

 that will serve to place some of these ques- 

 tions before you in more concrete form. In 

 the first view we see the results of crossing 

 two distinct races of sweet peas, each of 

 which is pure white and produces only 

 white offspring so long as strictly inbred. 

 On crossing the two races the hybrids are 

 always deep purple, like the wild Sicilian 

 species, and in this respect they no doubt 

 revert to an early common purple ancestor 

 of the two white races. The offspring of 

 the hybrids include a variety of purple, 

 red and white forms, among which are 

 whites that are identical with the two orig- 

 inal grandparents. Here is a somewhat 

 similar case in domestic fowls. You see 

 two different white races, each of whi.h 

 breeds true; but when crossed together 

 they produce deeply colored hybrids, show- 

 ing a pattern of plumage that is closely 

 similar to that of the wild jungle fowl from 

 which both white races are probably de- 

 scended. How is the reversion shown in 

 this and the preceding case to be explained ? 

 Let us look at some more complicated 

 phenomena. We have here the result of 

 crossing two differently colored races of 

 fowls, the barred Plymouth Rock and the 

 black Langshan. If the barred cock be 

 paired with the black hen, all the offspring 

 are barred, like the father. If the barred 

 hybrids be paired together the progeny in- 

 cludes, on the average, three barred to one 

 black, and the black bird is always a fe- 

 male. Quite different, and even more sin- 

 gular, is the reverse cross shown in the next 

 view, where the black cock is paired with 

 the barred hen. Half the offspring are now 

 barred and half black ; and the remarkable 

 fact is that the barred birds are all males, 



the black ones all females. In color pat- 

 tern the sons are like their mother, the 

 daughters like their father — an example 

 of the so-called "criss-cross" heredity. 

 Upon pairing these hybrids together, the 

 following generation (grandchildren of 

 the original forms) includes males and fe- 

 males of both types, barred and black. In 

 the following view we see a quite analogous 

 form of heredity, observed by Morgan in 

 crossing a long-winged and a short-winged 

 race of fruit-flies {Drosophila) . When the 

 male of a long-winged race is paired with 

 the female of a short-winged race, aU the 

 sons are short-winged like their mother, all 

 the daughters long-winged like their 

 father. On pairing these two, the offspring 

 are of all four types, long wings and short 

 wings occurring in both males and fe- 

 males. 



Such results seem at first sight ca- 

 pricious, almost fantastic, but this first 

 impression is erroneous. The results are 

 not capricious, but constant. The experi- 

 ments may be performed over and over 

 again, always with the same result, so that 

 the outcome may be unfailingly predicted 

 in advance; and this demonstrates that 

 such forms of heredity, and heredity in 

 general, must be due to some definite appa- 

 ratus in the germ-cells. I shall try to show 

 that microscopical r&search has revealed to 

 us at least something of the nature of this 

 mechanism, and that it has practically 

 solved some of the very puzzles that have 

 just been propounded. In order to indi- 

 cate the nature of this solution, I must first 

 ask attention for a moment to the so-called 

 "unit-characters" and their behavior in 

 heredity, on which the attention of both 

 cytologists and experimenters on heredity 

 has been largely concentrated in recent 

 years. 



Unit-characters have become too fa- 

 miliar to require more than brief illustra- 



