344 ROBERT H. BOWEN. 



been rather tacitly accepted as an explanation. My observations 

 on Euschistus do not bear out this view, and I should like there- 

 fore to make a brief survey of the problem as a whole in an effort 

 to find, if possible, a consistent basis for the whole series of ob- 

 servations. 



It has often been pointed out that the chondriosome-material 

 in dividing cells (spermatocytes, and others) tends to appear in 

 the form either of granules or of rods and threads of various 

 dimensions. In fact one can arrange these various morphological 

 types in a rather instructive series. There is first the case in 

 which the chondriosomes are scattered rather evenly throughout 

 the cytoplasm in the form of small granules, as, for example, in 

 the guinea pig (Duesberg, '10). An orderly arrangement with 

 respect to the achromatic division figure here seems entirely lack- 

 ing. In many pulmonates (Gatenby, '18) conditions are very 

 similar except that the granules are much larger, and in many 

 Lepidoptera (Gatenby, '17) enlarge to form vesicular bodies, in 

 some cases few in number and of considerable size. From such 

 a condition it is a short step to that in the scorpion Opistha- 

 canthns, where a definite(?) number of " chondriosome spheres" 

 occur, involving a further element of regularity which seems 

 lacking in the other cases mentioned. All the chondriosome for- 

 mations so far considered agree in one particular ; they do not 

 appear to be subject to any influence other than that of chance. 

 Only in a single case, so far as I know, has a definite orientation 

 of small scattered chondriosomes been described. This is the 

 condition in Ascaris described by Hirschler ('13) where the 

 slightly elongate chondriosomes are all distinctly oriented toward 

 the centrioles. This observation should be confirmed, for if cor- 

 rect it throws much light on the whole mechanism by which 

 granular chondriosomes are distributed in cell division. 



The chondriosome-ring of Ccntrurus offers an interesting in- 

 troduction to the rod type of mitochondria, for after becoming 

 drawn out along the spindle it breaks apart into two rod-like 

 halves which are subsequently divided across and the parts dis- 

 tributed to the daughter cells. It is only a step from this case to 

 that of Pahidina (Gatenby, '190) where there is a limited num- 

 ber of heavy rods which become arranged along the spindle in 



