The Role of the Chromosomes 215 



another in maturation and then separate in the reduction 

 division, so that different germ cells receive different contri- 

 butions. We have not however considered how the chromo- 

 somes pair, whether end to end, or side by side. Much dis- 

 pute has arisen over this question, which is probably due to 

 the occurrence of different methods in different species. In 

 some cases at least there is definite evidence of a side by side 

 conjugation or ' * parasynapsis, " and in some of these it has 

 been shown that the chromosomes do not lie parallel but wind 

 about one another, forming a more or less twisted braid. 

 When this occurs it is probable, although not definitely proven, 

 that in separating the chromosomes do not unwind but rather 

 pull apart irrespective of the twist, so that a part of one 

 chromosome may now be switched over into another chromo- 

 some and vice versa. This phenomenon of "crossing over" 

 as it is called, would readily explain the 8.5% of gray vestigial 

 and black long flies obtained in the last experiment; for if 

 the chromosomes carrying the black vestigial and gray long 

 determiners were occasionally to wind about each other and 

 then separate without untwisting the determiners would be 

 apt to get mixed up and find themselves in the wrong pews, 

 so to speak. Why this should occur only in the female and 

 not in the male is a problem. Possibly more extensive experi- 

 ments would show it to occur in the male also. However, the 

 lack of cross-overs in the male of another species of Drcsn- 

 phila (viriles) suggest that it is a constant feature of this 

 genus. 



Now if this explanation of crossing over be correct, we 

 should expect those characters to cross over most frequently, 

 the loci of whose determiners in the chromosomes are most 

 widely separated, for in the twisting, points at either end of 

 a pair of chromosomes would be more apt to interchange 

 places from one chromosome to the other, than closely adjoin- 

 ing points. And as a matter of fact, great differences do 

 exist in the amount of crossing over between different charac- 

 ters. On the basis of these differences Morgan and his 

 students have made a plan of the chromosomes, locating with 

 a high degree of probability in tiny threads of protoplasm, 

 possibly the diameter of the finest filame'it of a spider's web, 

 the determiners of more than two hundred characters already 

 studied in the fruit fly. 



The fact that all these characters fall into four groups in 

 respect to their linkages, thus corresponding to the four pairs 

 of chromosomes in the fruit fly; and further, that the cal- 

 culated separation between the extremes in each series, based 

 upon the frequency of the cross-overs, corresponds to the 

 relative lengths of the chromosomes, is very strong addi- 



