118 CHAS. B. WILSON. 



the ability to restore lost parts so quickly, so easily^ and with 

 so little apparent inconvenience during the process, that any- 

 thing short of being absolutely eaten whole could scarcely 

 prove fatal. 



"We are left with the two most interesting questions to 

 answer. First, how is fission actually accomplished ? The 

 only description of the process that can be found is the one 

 given by Benham (10). From a study of sections in different 

 stages of fission he concludes that a double row of nuclei 

 appear in the connective tissue along the line of rupture, and 

 that the longitudinal muscle-fibrils are severed between these 

 rows of nuclei, either through the cell substance of the con- 

 nective tissue becoming actively contractile and nipping off 

 the fibrils, or in consequence of the fibrils being eaten through 

 by some solvent action. 



Neither of these conclusions would explain fission in Cere- 

 bratulus, for the whole process, under irritation, occupies only 

 a fraction of a minute. Hence the connective tissue would 

 have no time to become contractile, but must exist in that 

 condition all the while, ready to act at an instant's notice, 

 i.e. it must itself be muscle. On the other hand, any sol- 

 vent powerful enough to act in so limited a time would not 

 be likely to stop with the mere severing of the longitudinal 

 muscle; it would dissolve everything within reach, and prevent 

 any regeneration. How, then, is fission accomplished ? 



The rupture takes place through one of the intestinal cceca, 

 where the body-walls can be most easily divided. 



It may take place at any cascum, and sometimes under ex- 

 treme irritation the worm will divide itself into fragments so 

 small that they contain but a single pair of pouches. 



If there were any special apparatus for producing fission it 

 ought to reveal itself opposite each of the caeca, at least in 

 specimens killed after irritation. But careful examination 

 fails to reveal anything of the sort. 



What we do find is that Cerebratulus has a thick layer of 

 circular muscles, which are almost entirely lacking in Cari- 

 nella. 



