176 r. KEEBLE AND F. W. GAMBLE. 



As Table III sliows^ we liave, Ly adopting this procedure, 

 succeeded in obtainiug batches of Convoluta whicli remained 

 absolutely colourless and uninfected. 



The several columns of this table give the results obtained 

 in the cases of : 



(i) Animals left in association with their capsule-remnants. 



(2) Animals removed from their capsule-remnants. 



(3) Animals which, having been so removed, were subse- 

 quently submitted to the risk of infection by placing them in 

 fresh unfiltered sea-water. 



(4) Animals hatched without any precautions in ordinary 

 sea-water. 



The animals hatched in unfiltered sea -water became 

 uniformly green after two or three days. Among those left 

 with capsule-remnants infection occurred less rapidly and 

 more sparingly. Thus, as Columns 3, 4, 5 of Table III show, 

 eleven individuals out of fifty-nine became infected during 

 the eight days which followed after general infection had 

 declared itself among the animals reared in unfiltered sea- 

 water. After seventeen days, infection had become general 

 in all these cases (Columns 3, 4, 5). 



The contrast between this result and that set forth in 

 Columns 1 and 2 of the same table is emphatic and conclusive. 

 In those cases (Columns 1, 2, Table III) in which the animals 

 had been separated at the time of hatching from their capsule- 

 remnants the total numbers of animals showiuG: infection 

 Avere, in the one experiment (Column l),fivc out of forty-four 

 animals examined, and in the other none out of forty-seven 

 examined. We conclude, therefore, that the green cells of 

 Convoluta are of intrusive origin, or, to use the terms 

 employed already, that they arise as the result of an infection 

 from the water of the sea. From a single infecting cell, or 

 at most from two or three, are produced the vast numbers of 

 the green cells of the adult Convoluta: infection taking place 

 normally during the first three days after hatching. 



Experiments made in the course of this inquiry enable us 

 to answer two other questions — viz. May the green cells of an 



