120 PL ANT- ANIMALS [OH. 



increases till it forms a multitude. Then the capsule 

 bursts and swarms of infecting organisms are liberated 

 and, ingested eagerly by C. roscoffensis, give rise to 

 the chlorophyllous cells of its body. The idea sug- 

 gests the simple method : isolate the empty capsules, 

 as well as the just-hatched young, and in a week or 

 two some of the capsules will be found teeming with 

 the infecting organism. 



On returning to Brittany in the following summer 

 the first thing done was to test the hypothesis. 

 Animals were washed and put to lay in filtered 

 sea-water ; the egg-capsules were washed likewise and, 

 when the larvae were hatching out, the latter were 

 put in one vessel and the remains of their capsules 

 in another. The animals remained colourless, though, 

 when samples of them were put into ordinary sea- 

 water, green cells made their appearance in their 

 bodies with uniform regularity. After seventeen 

 days, several small, green, globular bodies, each as 

 large as a big pin's head, made their appearance in the 

 water of the vessel containing the capsule-remnants 

 (Fig. 19). Their hue was the dark spinach-green of 

 C. roscoffensis. On microscopic examination, under 

 the slight pressure of a cover-glass, the dark green 

 mass dissolved and formed a cloud of active, green 

 flagellated cells, emerging from an egg-capsule (Fig. 20). 

 Though these free cells differed in various details from 

 the green cells which occur in the body of C. roscoffen- 



