4.60 FREDERICK KEEBLE. 
under experimental conditions is, as has been shown (Keeble 
and Gamble, 1907), in part due to the regular and close 
relation which exists between the infecting organism in its 
free state and the egg-capsules of the animal and in part to 
the ubiquity of the infecting organism. In that species the 
free, four-flagellated, Chlamydomonadine alga which consti- 
tutes the infecting organism settles habitually on the gelatinous 
egg-capsules, and, undergoing division, gives rise in the 
empty capsules to spherical colonies. It was the observation 
of this fact which led to the isolation of the infecting organism 
of C. roscoffensis. 
In C. paradoxa this habit on the part of the free infect- 
ing organism of frequenting the egg-capsules does not appear 
to obtain, or, if it obtains, is much more difficult of demon- 
stration. 
Repeated experiments were made in which the capsule- 
remnants, left after the larvee had escaped, were put in 
filtered sea-water and kept under observation ; but, though 
these capsules developed a fairly luxuriant flora of diatoms, 
brown flagellates and yellow-brown cells resembling in 
number of chloroplasts and in refractive inclusions the 
yellow-brown cells of the adult C. paradoxa, it was not 
found possible to induce the infection of larval animals by 
putting them in contact with these possible sources of infec- 
tion. Too much weight must not, however, be attached to 
these negative results ; for the experimental difficulties are 
greater in the case of C. paradoxa than in the case of 
C. roscoffensis, partly because the former animals are 
more difficult to obtain in any considerable numbers, and 
partly because of the difficulty of imitating in the laboratory 
the conditions under which C. paradoxa lives in the open. 
For the same reasons all attempts to isolate the infecting 
organisms or to obtain it in its free stage have failed. The 
yellow-brown cells of the adult or young infected animal 
cannot be induced to grow when removed from the body of 
the animal. The same fact has been established with respect 
to the green algal cells of C. roscoffensis. They also fail 
