8 Introductory. 



genes obtained in both species can be used for comparison. Further 

 details concerning this case are given in a later section. 



Owing to the fact that the material for the present work, and for 

 other investigations of a similar kind, has been selected largely on 

 the basis of apparent chromosome relationships, these may be con- 

 sidered briefly here. 



CHROMOSOME RELATIONSHIPS. 



The known types of chromosome groups found in the genus 

 Drosophila, with the addition of one type from the related genus 

 Cladochceta, are represented in figure 1 (from Metz, 19166). A 

 detailed comparison of the various types is given in the paper just 

 cited, and we may hmit ourselves to features concerning the species 

 under genetic observation. The types represented by the latter 

 are shown in figure 2.^ In this figure the sex chromosomes, where 

 identified are represented in solid black, and where not definitely 

 known, the ones presumed to be the sex chromosomes are cross-ruled. 

 Considering first the sex chromosomes, it may be noted that two 

 main kinds are involved — the short, rod-like form found in D. 

 melanogaster, simulans, and virilis, and the long V-shaped form 

 found in willistoni and obscura. In funebris the sex chromosomes 

 have not been identified, although the behavior of the longest pair 

 suggests that it represents the sex chromosomes. 



Comparing the autosomes, it is seen that the same two sorts are 

 represented — short rods and long V's — in addition to the small, 

 dot-like pair common to all save possibly D. willistoni. In both the 

 autosomes and the sex chromosomes, it is to be noted that the 

 V-shaped members are approximately twice the size of the rod-like 

 ones (with the exception of one pair in D. funebris) and that each 

 has a constriction in the middle. This suggests a possible relation- 

 ship between the types which may best be appreciated, perhaps, by 

 imagining all of the V-shaped chromosomes broken in the middle, 

 at the point of constriction. With this alteration all of the groups 

 conform essentially to one type, that represented by D. virilis. 



For the present purposes, this series of chromosome groups repre- 

 sents an almost ideal condition. There are clear-cut differences 

 between the types, yet the differences do not appear to be purely 

 at random, and it is possible to compare the individual tjT)es, chro- 

 mosome for chromosome. 



The relations between the types suggest the hypothesis that the 

 latter have arisen by a process involving the breaking up of large 

 V-shaped chromosomes into rods or, vice versa, by the fusion of rods 

 to form V's. If this hypothesis is correct, genetic analysis should 

 reveal the fact through the resemblance between groups of linked 



* Exclusive of two or three species which have only been studied slightly. 



