THE BIONOMICS OF CONVOLUTA ROSCOFFENSIS. 387 



ensis; but we liave not been able to undertake pure cultures 

 of its green cells or of its persistent colourless cells with a 

 view of proving the origin of this capsular flora from the 

 disjecta of the parent. The ground for the belief lies in the 

 fact that every important kind of capsular cell (except the 

 filamentous ones) can be matched by cells in Oonvoluta 

 itself. Experimental proof, however, is needed, since a great 

 exotic fauna and flora of alga=^, diatoms, and Protozoa also occur 

 on the capsule. The amount of contamination adhering to 

 the capsule varies considerably, but we have not seen a single 

 case of a capsule free from some one or other of these 

 remarkable cells. 



It will now be clear that the constant presence of these 

 contaminating tissues is a source of error which precludes 

 the use of egg-capsnles in any experiment with sterile water 

 on the nature of the green cells and their colourless pre- 

 decessors in Convoluta. That precautions of no ordinary 

 kind have to be taken in isolating larv» and transferring 

 them into jars of sterilised water, also follows ; for these 

 larvae, even before escaping from the capsule, may ingest 

 some of the colourless capsular cells, or some of the latter 

 may be transferred with the larvse by the action of the 

 pipette. 



To obviate these sources of error is a matter of the greatest 

 difficulty, which, indeed, we cannot claim to have com- 

 pletely overcome. But experiments in this direction — such, 

 for example, as seizing Convoluta at the moment of hatching, 

 and washing them several times in sterile water before trans- 

 ferring them to other sterile water — all point to the conclusion 

 that the more precautions of this kind are taken, the fewer 

 colourless antecedents of green cells are registered in larvee 

 reared in sterilised or filtered water. 



These precautions were taken in our last experiments. 

 Advanced embryos as well as larvte were taken out of the 

 capsule, and carefully and repeatedl}^ and appropriately washed 

 before being placed in the normal, filtered, or sterilised sea 

 water. Control cultures were arranged so that access to the 



