112 MR. C. F. V. MEEK ON THE CORRELATION OF 



in 50 per cent, of these cells, is again easily recognizable on account of its 

 great breadth. I hare already produced evidence to show that rods com- 

 posing ordinary chromosomes of organisms above and inclnding Nemat- 

 helminthia have a constant diameter, viz. '83 fi, and that their lengths 

 constitute members of a general series in arithmetical progression ; and the 

 chromosome measurements of S. curtlpennis therefore afford further support 

 to this assumption. 



The accompanying figure (fig. fi) shows the complexes of Stenohotlirus 

 parallelus, S. virid%dus, S. hicolor, and S. curtipennis, the four complexes 

 being respectively marked A, B, C, and D. In my recent paper I have 

 identified rod-lengths of the general series by numerals^ which are again used 

 and are placed above the corresponding chromosomes. The drawings show 

 component rods, each spermatogonial and secondary spermatocyte chromosome 

 being composed of two and each primary spermatocyte chromosome of four. 



The rod-lengths of the five short chromosomes appear to be the same in all 

 four species, whereas those of tlie three long chromosomes are not identical 

 in any two : in S. parallelus they correspond respectively with Chromo- 

 somes 14, 15, and 17 of the general series, in S. v'lriduJus with Chromo- 

 somes 13, 15, and 17 ; in S. hicolor with ( 'hromosomes 11, V^, and 17, and in 

 *S'. cvrtipennis with Chromosomes 11, 13, and 15, 



If now we assume that the chromatin is directly concerned with the 

 transmission of the hereditary characters — and we have many reasons for 

 assuming this — we must look for tlie cause of somatic differences between 

 these species in the three long chromosomes, for the respective nuclei 

 appear to differ ouly in the lengths of these. Moreover, we must try 

 to discover how these differences in rod-lengths have occurred, for the 

 problem of chromosome function must l)e intimately connected with such 

 differences. 



Let us firstly assume that chromosome I'ods throughout the animal 

 kingdom are of fixed lengths, and that morphological similarity is invariably 

 accompanied by functional similarity. This assumption carries with it the 

 further assumption that in the course of evolution certain rods have dis- 

 appeared from each complex, local conditions having determined which 

 should persist and which should be eliminated : moreover, it postulates a 

 greater number of chromosomes in primitive than in highlv organized types, 

 and we must expect to find allied organisms possessing niuny chromosome 

 lengths in common. The former of the last-named corollaries is, however, 

 not supported by actual investigations, and Avith regard to the latter I have 

 already shown that Forficida does not possess one rod-length in comilion with 

 Stenohotlirus. ^Ve are accordingly faced by a complete contradiction, for, if 

 a definite chromosome rod-length is invariably correlated with a definite set 

 of somatic characters, no such set of characters can be possessed by both 

 earwings and grasshoppers — members of sister families. 



