114 ME. C. F. IT. MEEK ON THE CORRELATIOX OP 



The second and alternative hypothesis postulates a continuous linear 

 growth o£ chromosomes in the course of phylogeny, and is based upon 

 data that support direct correspondence between the degree o£ somatic 

 complexity o£ an organism and the total volume of its germ-cell chromatin ; 

 it moreover offers a logical explanation o£ the evolution of various rod- 

 lengths. The measurements that I have given can be only approximations, 

 and the difference between terms of the general series is probably smaller 

 than that shown, but these measurements suffice to prove differences between 

 complexes of allied species ; and it is only reasonable to suppose that such 

 diff'erences are of comparatively recent origin and have evolved by some 

 continuous process accompanying the somatic diff'erentiation of the species. 

 In certain cases the process may have been complicated by degeneration, 

 possibly resulting in the complete disappearance of a particular chromosome 

 from the complex^ but, even if this additional factor is eventually established, 

 I am aware of no reasons for discarding the assumption that the guiding- 

 principle in complex formation is and has been a purely linear growth of 

 component rods. 



I have already pointed out that if this second hypothesis is subsequently 

 proved, morphological identity of chromosomes can be no guide to functional 

 identity outside the narrowest hmits of our classification ; but in the case 

 of allied species, which must have evolved along almost parallel lines, we 

 may reasonably hope to establish correlation of individual chromosomes 

 of the respective complexes, and thus form a basis upon which to attempt 

 correlation of rod lengths and definite somatic characters, 



Let us therefore consider again the complexes of these four species of 

 Stenobotlirus. The lengths of the five short chromosomes are the same in all 

 cases, and their identities appear consequently to be established ; the three 

 long chromosomes, however, are not the same in any two complexes, and 

 correspondence is therefore not at once ap}»arent. If we accept the first 

 hypothesis, which postulates invariable correspondence between definite 

 rod-lengths and definite sets of somatic characters, w^e must realize that no 

 long chromosome is common to all four species; (Jbromosome 11 is absent in 

 S. parallelus and S. viridulus. Chromosome 13 is absent in S. j^arallelus, 

 Chromosome 1-4 is absent in S. viridulus, S. hieolor, and S. curtipennis, 

 Chromosome 15 is absent in *S'. hieolor, and Chromosome 17 is absent in 

 S. curtipennis. If, on the other hand, we accept the second hypothesis, 

 which postulates continuous linear growth of rods, we must realize that in 

 these complexes the short, medium, and long chromosomes of the three are 

 probably respectively identical. Thus Chromosome 13 of S. virididus does 

 not correspond with Chromosome 13 of S. hieolor and S. curtipemiis, but 

 corresponds with Chromosome 11 of the two latter and with Chromosome 14 

 of S. parallelus, for these chromosomes constitute the shortest member of 

 the long group in each case ; similarly, Chromosome 17 of S. parallelus. 



