A STUHY OF RESTITUTION MASSES. 457 



contraction of the plasmodium away from them, and secrete a perisarc 

 within 12 or 18 hours. This resumption of function in so short a time 

 and the locaUzation of the secreting power so that secretion of perisarc 

 takes place only on the outside of the mass, is somewhat remarkable, 

 considering that all the normal relationships between the cells must have 

 been completely upset. Contraction of the plasmodium of endoderm cells 

 still continues, eventually causing the mass to contract away from the 

 perisarc, and it is noteworthy that when this occurs no fresh perisarc is 

 secreted by the peripheral cells. If, however, a small piece of the perisarc 

 is removed, it is rapidly re-secreted by the cells in the neighbourhood, and 

 the gap is healed. From this it would seem possible that the secretion 

 of perisarc is a direct reaction of the ectoderm cells when in contact with 

 sea-water, and that it is not produced when they are in contact with the 

 fluid filling the space between the shrunken mass and the perisarc, and 

 that accordingly this fluid has a different constitution to sea- water. 



During the retraction of the plasmodium, the endodermal cells gather 

 together their ramifying processes and again become differentiated, and 

 so verv slowly resume their normal form. Of the manner in which many of 

 them become collected so as to form definite tubules, we are unable to 

 offer any explanation. The appearance of sections at a period when 

 the earliest stages of tubule formation are apparent suggests that the 

 plasmodial masses, which will later develop into endodermal tubules, 

 become arranged and segregated before they have differentiated into 

 recognizable endoderm cells, but this is the merest surmise. In cases like 

 this, where individual cells cannot be watched through their modifica- 

 tions and development, the value of the evidence of sections of different 

 masses is always doubtful. Considering that no sign of cell division was 

 ever seen in our experiments it would seem strongly probable that a 

 certain amount of actual migration and rearrangement of the cells 

 within the masses must occur, but no clue is given as to the form or 

 mode of action of the forces causing this rearrangement. As in the 

 case of Wilson's experiments, however, it is difficult to imagine any 

 " form regulation " force coming into play after isolation and subsequent 

 agglomeration of the individual cells forming the original organism. 



We consider that the granules found in the cytoplasm of the endodermal 

 cells in the older restitution masses are probably in the nature of the pro- 

 enzyme of the digestive ferment. Since the tubules are closed, and there is 

 no food stimulus, the granules are not discharged, but accumulate in the 

 cell up to a certain limit ; when this is reached the cell ruptures and 

 autolvsis ensues. From the fact that these granules are formed, it 



