THE BIONOMICS OF CONVOLUTA EOSCOFFENSIS. 867 



adult Convolutas as well as on the eggs, we show that 

 the evidence is against Haberlandt's theory ; and that 

 Georgevitch's view of direct infection, though probably 

 correct, is insufficient, inasmuch as there is no infection of 

 Convoluta by a green cell. We show that the problem of 

 infection is at once more complex and more difficult than has 

 been supposed. More complex, inasmuch as the green cells 

 of Convoluta develop from colourless cells within its body. 

 More difficult, because it is impossible, by ordinar}^ means, 

 to cultivate Convoluta in a sterile medium; and this not 

 because Convoluta does not hatch out and grow well in 

 filtered or heat-sterilised sea water — this it does, — but because 

 the egg-capsule is far from sterile, being, in fact, infested by 

 all manner of gi-een and colourless cells, spores, and even 

 diatoms. 



Though we have failed to isolate an infecting organism, 

 we bring forward evidence in favour of the theory that the 

 green cells of Convoluta are the result of an infection. 

 This evidence is derived from experiment, and from observa- 

 tions on the position in which the colourless progenitors of 

 the green cells are first found in the body of the recently 

 hatched animal. We trace the stages in the development of 

 these colourless cells into the green cells, and the migrations 

 of the latter through the body to their final disposition as a 

 dense green layer just beneath the skin. 



Finally, we supplement and extend the observations of 

 Haberlandt on the tropisms of Convoluta, and endeavour 

 to show how these tropisms co-operate to form the habits and 

 determine the distribution of the animal. 



The research was commenced at Roscoff in the summer of 

 1901, and since then has been carried on at Tregastel (Cotes 

 du Nord), and in the laboratories of Owens College, Man- 

 chester, and University College, Eeading. 



