POLYPIDES OF CEPHALODISCUS NIGRESCENS. 25 



The pigmentatiou * of the skiu is deep over the visceral mass (figs. 7 and 8), 

 less deep over the gonads and along the ventral surface of the stolon, which 

 touches the interior of the tulje, and over the part of the surface of the body 

 between the buccal shield and the stolon. The axes of the plumes have a double 

 black band running along them, the two bands converging towards the tip, which 

 is entirely black. There is a black edge to the buccal shield. Near the posterior f 

 edge of the buccal shield is a curved line of brilliant red, and a similar red 

 pigment is seen in the oviducts. 



The blackish colour of the body is due to a superficial layer of large cells, 

 which are brown in colour when mounted iu glycerine and examined under the 

 microscope (see fig. 9, plate 3) ; they do not stain with haematoxylin or borax-carmine 

 solution, they seem to have no nuclei, and they contain each one small spherical 

 pigment granule, sometimes two. In very thin sections each cell appears to be 

 composed of thirty or forty closely packed polyhedral grains of a dark straw or a 

 raw sienna colour. This is the appearance presented in sections or teased preparations 

 made from the material fixed in formalin, or iu Perenyi's fluid. In that fixed in 

 picric acid solution the pigment cells are considerably swollen and are not brown 

 in colour, but each cell shows its one or two small black spheres, as in the 

 other material. Pigmented cells similar to those of the epidermis occur in 

 the wall of the pharynx and in the pleurochords, either as solitary cells or in 

 small groups. 



At intervals in the sections of the pigmented epithelium are seen masses which 

 stain deeply, each about as large as one of the pigmented cells. In carmine- 

 stained sections these are uniform in tint, but in those stained with haematoxylin 

 they appear finely granular. They are presumably the secreting cells of the 

 epidermis. They show no nuclei, and there is no distinction between the cytoplasm 

 and the deeply-staining mass, which latter I take to be the secreted material not 

 yet emitted. These cells occur in all parts of the superficial layer, over the body, 



* The species now under consideration was termed C. nigrescens (15) before the publication of the ' Siboga ' 

 Eeport on the Pterobranchia. In that Report Harmer notes that the epidermis of ail parts of the body of 

 C. sihogae is pigmented (10, pp. 8 and 49). He also states that a median line of pigment occm-s on the anterior 

 side of the stolon of C. gracilis (p. 52), and that deserted stolons of individuals of this species are deeply 

 pigmented (p. 93). 



t The usual orientation of the body is accepted in this communication ; the plumed end of the body is 

 regarded as the anterior end, and the surface of the body upon which the mouth opens, and from which the 

 stolon arises, as the ventral siyface. Harmer, in his most i-ecent work on the subject {10, p. 23), adopts a new 

 orientation, taking the main axis of the body as jsassing from the middle of tlie buccal shield through the central 

 nerve mass, and ending on the rectal side of the body a little below the anus, so that the visceral mass is regarded 

 as a ventral downgi-owth. He denies that in the bud the elongating caecal end of the intestine leaves its position 

 in the part of the " body " near the stalk and migrates along the aboral surface of the pharynx, subsequently 

 opening to the exterior in the position in which the anus is found in the adult, and suggests (10, p. 100, par. 3) a 

 lateral pinching-in which separates a dorsal intestine from a ventral stomach, with formation of a median 

 mesentery between. One objection to the acceptance of Hai-mer"s orientation is that in the adults of C. nigrescens 

 the notochord is a long structure lying parallel with the thick face of the shield, and more or less in a Ime with 

 the pliaryngeal axis (tex-t-fig. 10, p. 34), and it is not unreasonable to regard this as determining the median axis 

 of the polj'pide. For practical purposes the earlier terminology is preferable. 



