I0 CTENOPHORA. 



ings of the canals into the genital organs and the tentacle base can only be seen in sections. Comp. 

 PI. IV. Figs. 3 4. PI. V. Figs. 6, 7, u, 12. PI. VI. Fig. 7). The transverse main vessel, of course, corre- 

 sponds to the transverse, perradial, main vessel of typical Ctenophores; the interradial vessels are not 

 represented, the genital vessels, which evidently correspond to the adradial vessels of typical Cteno- 

 phores, proceeding directly from the main vessel. Meridian vessels are not developed, and likewise 

 there are no pharyngeal vessels ("Magengefasse"). The two diagrams (Figs, i 2) may serve to bring 

 out clearly the relations between the gastrovascular system of Tjalfiella _and typical Ctenophores. 



As seen from the figures on PI. I the branching gastrovascular canals in the walls of the bod}' 

 do not form anastomoses, or at least only exceptionally. (In Fig. 8, PI. I a single anastomosis is ob- 

 served). The branches on the two sides of the "chimneys" do not unite on the outer side; likewise 

 the two branching canal systems on the same side of the body do not unite in the middle line (over 

 the sagittal plane). 



The structure of the gastrovascular canals is that typical in Ctenophorans, two lateral thicken- 

 ings forming prominences into the lumen of the canals, while the upper and lower (outer and inner) 

 wall is formed of low epithelium; this structure is found in the whole of the branching gastrovascular 

 canals in the body wall as well as in the main vessel (PI. VIII, Fig. 2; comp. also the different sections 

 represented in PI. IV and V). The two prominent ridges consist of more or less strongly vacuolated 

 cells. (Though the vacuoles may partly be due to the preservation it is beyond doubt that part of 

 them are real enough; also in other Ctenophores the cells of these ridges are generally highly 

 vacuolated). The nuclei are spread irregularly in the ridges, while in the low epithelium of the two 

 other sides of the canals the nuclei are arranged rather regularly in a single layer (PI. VIII, Fig. 2). 



Cilia do not appear to occur in the gastrovascular canals; at least I have been unable to find 

 any traces of ciliation here, while in the pharynx e. g. it was very easy to observe. 



Rosettes, which had, of course, to be sought for in the thin walls of the branching canals, 

 appear to exist as in other Ctenophorans. I have not succeded, it is true, in getting a fully satis- 

 factory view of them, but of their existence I have become convinced. In one case I have seen the 

 whorl of cilia of a rosette most distinctly projecting into the lumen of the canal. 



Here may be included an observation regarding the food of Tjalfiella. In one specimen was 

 found in the suboral cavity a shrimp of ca. i cm length, filling out the whole cavity, the tail lying 

 partly within the one chimney. It was lying with the back downwards. As it was already half decom- 

 posed it would scarcely be possible to identify the species, and it was therefore not dissected out. 

 This observation proves that Crustaceans, and not especially the smaller forms, make at least part of 

 the food of this Ctenophore; whether it also catches other animals as prey must remain an unsolved 

 question for the present. That it must catch them by means of its tentacles seems beyond doubt. 

 - It is worth noticing that the food of Tjalfiella consists of Crustacea (partly at least), thus being in 

 accordance with what is the case also in other Ctenophores (except the Beroids, which prey on other 

 Ctenophores); these latter, however, take only small pelagic Crustaceans (and occasionally other 

 pelagic animals), according to the observations of Chun (Monogr. p. 240), with which my own obser- 

 vations on Bolina and Pleurobrachia agree, while Tjalfiella takes larger, non pelagic, Crustaceans. For 

 the rest it is scarcely probable that it keeps to the Crustaceans alone; it would appear more probable 



