118 MR. C. r, 17. MEEK OX THE COREELATION OF 



chromosomes o£ a complex are qnalitatively different, as we have reasons for 

 believing, each must be concerned with a definite set of characters : difference- 

 in length of two corresponding chromosomes may therefore be connected 

 with differences in several characters, and, even if the character under 

 consideration is included in these, other factors are equally bound up in 

 the chromatin rod, and may be responsible for apparently irreconcilable 

 lengths. 



In the circumstances I am inclined to think that the fourth explanation, 

 possibly coupled with the third, will eventually be found to account for our 

 present failure. It is difficult to believe that the obvious somatic differences 

 mentioned in our list are not in some way connected with the lenoths of the 

 three long chromosomes, but until a thorough analysis has been made of both 

 internal and external characteristics of these species we cannot hope to 

 correlate somatic characters and chromosome rod-leu oths in the oenus. 

 In 1908 McClung indicated a course of investigation upon Acridiida? to 

 be carried out upon these lines, but I have seen no paper by him on the 

 subject : possibly he and his followers have been able to throw some light 

 upon this difficult problem. 



Resume. 



Each ordinary spermatogonia] and secondary spermatocyte chromosome 

 of S. curtipennis is composed of two equal rods, and each primary spermo- 

 tocyte chromosome of four. The diameter of these rods is invariably 

 "83 fi and consequently lends further support to the assumption that the 

 chromatin thread-width is constant in all organisms above and includinsr 

 Nemathelminthia. 



The lengths of the ordinary rods constitute members of a general series 

 in arithmetical progression ; the five short chromosomes ai'e respectively 

 identical with those of other members of the genus, but the lengths of the 

 three long chromosomes once more enable the species to be identified. 



A comparison between *S'. parallelus, S. virididvs, S. hicolor, and aS'. curti- 

 pennis fails to establish correlation of somatic characters and chromosome 

 rod-lengths, but we have reason for believing that the obvious characteristics 

 upon which identification is based are in some way connected with the three 

 long chromosomes : our present failure is probably due to ignorance of the 

 less obvious somatic characteristics and t'> the lack of trustworthy methods 

 of identifying corresponding chromosomes in the respective complexes. 



