294 ALLEN. [VOL. I. 



are derived. Meanwhile the endodermal layer has thickened 

 in four regions equidistant from each other at the distal end of 

 the medusoid. At first the cells in the thickened region are 

 irregularly arranged, but later they form themselves into two 

 rows with a space between them. In a few cases no cavity 

 was found, and in two of the gonophores examined it extended 

 halfway around the bell. The remainder of the wall consisted 

 of a single layer. PI. I, Fig. 4, represents the condition found 

 in most of the medusoids. Agassiz ('62), p. 259, states that 

 there are neither radial nor circumoral canals in this species 

 but the position and mode of development of these cavities 

 leave little doubt that they are rudimentary radial canals. No 

 circular canals were observed, and the radial canals were never 

 found connected with the body cavity of the medusoid in any 

 of the hundreds of sections studied. 



Spermatogenesis . 



The large nucleated cells lying between the manubrium and 

 the inner wall of the bell become the sperm mother-cells, 

 which finally break up to form the sperms. In the first divi- 

 sion the karyokinetic figures are distinct and show spindles 

 and prominent chromosomes. The later stages were difficult 

 to study because of the minuteness of parts, and I was unable 

 to demonstrate clearly the exact number of spermatozoa de- 

 rived from a single germ cell, but I think four are usually 

 formed. Their structure could only be made out in particu- 

 larly favorable sections, but was easily demonstrated by crush- 

 ing the gonophore and allowing the sperms to escape. They 

 consist of a pear-shaped head with a very long, slender tailpiece. 

 When fully developed the male gonophores are spherical, and 

 the walls are so thin that their structure can only be deter- 

 mined by the use of very high powers. They bear no ten- 

 tacles, although the ectoderm is sometimes thickened slightly 

 in the regions where tentacles arise in the female. I examined 

 carefully a large number of mature male gonophores to learn 

 whether or not the ectodermal layer of the manubrium was 

 formed and discovered a definite transparent layer next the 



