HYDROMEDUSAE. 725 



The Absence of a Circular Canal. Brandt (1838) in his description of Proboscidactyla 

 Jiavicirrata states that a circular canal was not observed, but a very definite inner marginal 

 edge was visible. When I was examining sjiecimens of Willia luutabilis I noticed that what 

 then was regarded by me as the circular canal was very slender and inconspicuous compared 

 with the radial canals. As I could not make out a definite canal with the microscope a 

 series of sections was cut. The sections showed a solid chord of cells without any opening in 

 the centre. Sections were also cut of Willia stellata and these failed to show a circular canal. 

 I have also examined Willia stellata alive when an active circulation was going on in the 

 radial canals and inside the basal bulbs of the tentacles, but I did not see a circulation round 

 the margin of the umbrella. Externally the solid chord of cells looks like a circular canal 

 and expecting one to be present it is ea.sy to mistake it for a canal. As every tentacle is 

 in direct communication with the radial canals, the circular canal became functionless and it 

 has now ceased to exist. 



The Clusters of Nematocysts on the Ex-umbrella. In several species the nemato- 

 cysts on the ex-umbrella have been carefully described, and it has been shown that their 

 arrangement is a definite one. Radial clusters of nematocysts, situated midway between every 

 two tentacles, extend from near the margin of the umbrella to about halfway or nearly to 

 the top (if the umbrella. The clusters are connected with one another by a kind of canal 

 (just under the surface of the ex-umbi'ella), along which the nematocysts travel. An investigation 

 of the nematocysts in Willia stellata when alive showed that isolated nematocysts travelled 

 along the margin of the umbrella, then entered a canal, and after proceeding a short distance 

 stopped. A cluster is formed by the accumulation of nematocysts. I could not find a cluster 

 of nematocysts on the very margin of the umbrella, but always isolated nematocysts ; the first 

 cluster being at a little distance from the margin. In the earliest free-swimming stage there 

 is only a single cluster between every two tentacles and as the umbrella increases in size 

 other clusters are formed. The number of nematocysts in a cluster, the shape of the cluster, 

 and the number of clusters in each row, I found to be very variable in the adult of Willia 

 stellata. 



Genus Pr-oboscidactyla, Brandt, 183.5. 

 Generic Character. Williadae with four radial canals leaving the stomach. 



Proboscidactyla flavicirvata Brandt, 1835. 



Proboscidacti/la jiavicirrata, Brandt (1835); Brandt (1838, p. 390, Taf XIX.); Lesson (1843); 

 L. Agassiz (1862); A. Agassiz (1865, p. 173, figs. 280—282); Haeckel (1879); Proboscidactyla 

 brevicirrata, Haeckel (1879, p. 160); Murbach and Shearer (1902), (1903, p. 178). 



The description of this species by Brandt and in addition the figures drawn from life by 

 Mertens show that the main radial canals have two principal branches (a wide bifurcation), 

 which run to the margin of the umbrella. From the inner side of each of the principal 

 branches a number of branchlets run to the margin and the branchlets again branch near 

 the margin. Mertens' figures show that a main canal has 18 branches which are in connection 

 with 18 tentacles. The medusa should have altogether about 72 tentacles, a number far in 

 excess of those of any other described species of the family. 



93—2 



