368 M. F. E. Schulze on the Relationship 



fig. 29 of pi. xv. of his memoir, and which I have also 

 detected in many swarm-larvae, is nothing- but the perfectly 

 solid non-granular terminal part of the cells in question, from 

 the slightly convex free surface of which the delicate flagellum 

 originates. 



In my investigations of the swarm-larva? of Sycandra 

 raphamiS) which can hardly differ essentially in the structure 

 of its larva? from Sycandra compressa, and of many other 

 sponge-larva?, I have endeavoured, always in vain, to dis- 

 cover anything like the collar at the free extremity of the 

 cylindrical flagellate cells. But when I compare the figures 

 that Saville Kent gives of his " svvarm-geinmules " with 

 the images that one obtains by tearing up living Sycandrce 

 under the microscope, I cannot avoid the supposition that 

 what Saville Kent has described as a perfectly mature " swarm- 

 gemmule," and finally figured in his ' Manual of the Infusoria ' 

 (pi. ix. fig. 25), is nothing more than a separated portion of 

 the layer of collared cells, which has rolled itself up, so that the 

 basal extremities [of thecells] are turned inwards and the collars 

 outwards. Such deceptive images often come under observa- 

 tion when living Sycandra' are torn up in sea-water. Not unfre- 

 quently a separated sheet of cells becomes so completely rolled 

 up, that it even appears like a closed vesicle when rotating, 

 although, as a rule, it only forms more irregular structures, 

 such as are figured by Saville Kent in pi. vi. fig. 17 of his 

 memoir "On the Embryology of Sponges," in the 'Annals 

 and Magazine of Natural History,' ser. 5, vol. ii. 



Moreover, a hemispherically rolled layer of cells may occa- 

 sionally attach itself to the posterior extremity of one of the 

 oviform larva?, which are usually present in numbers, in such a 

 manner as to produce the form figured by Saville Kent {he. 

 cit. pi. vi. fig. 16, and ' Manual,' pi. ix. fig. 26), the posterior 

 extremity of which appears covered with collared cells. 



While, on the one hand, from my own observations, I 

 cannot accept the composition of the bodies designated "swarm- 

 gemmules " by Saville Kent out of collared cells equivalent 

 to Choanoflagellata, on the other hand I must assert that, 

 even if their structure were of the kind stated by Saville 

 Kent, their nature, as colonies of Flagellata, would be by 

 no means proved. They would then as now have to be re- 

 garded and described as true sponge-larva?, because they have 

 been produced by segmentation from a fertilized sponge- ovum, 

 and become afterwards transformed in the same way as the 

 well-known larva? of other animals, such as the Hydroida, 

 into mature Metazoa. Saville Kent will admit neither of 

 these arguments as conclusive. He rather endeavours to 



