MEDUSA. I. 



53 



sometimes happens, the opening of the stomach or the mouth becomes fairly large and circular. Be- 

 sides the margin is entire ..."'. The base of the stomach is broad, octangular, and the whole dorsal side 

 is entirely attached to the subumbrella ; there are, accordingly, no "grooves" or "centripetal continua- 

 'tions of the radial canals". The 8 radial canals open in the upper part of the sides of the stomach 

 through perpendicular slit-shaped openings (Plate I, fig. 10). On each side of these openings there is 

 a perpendicular fold. Probably the opening may be closed by means of these two folds, the lumen of 

 the radial canal thus being separated from that of the stomach. 



The radial canals are attached to the subumbrella along straight lines, but the lateral walls 

 of the canal are, for the greater part of their length, folded and sinuous and contain the gonads. The 

 folded gonads commence at some distance (about one-fourth of the length of the radial canal) from 

 the stomach; they are most highly developed towards the distal end; distally about i mm of the radial 

 canal is free of gonads. It is interesting that even fully developed gonads do not cover the whole of 

 the lateral wall of the radial canal in the dorso-ventral direction; they commence at a fairly consider- 

 able distance from the subumbrella, so that on each side of the radial canal nearest to the subum- 

 brella there is a stripe free of gonads; ventrally, on the other hand, there is only a very narrow streak 

 which separates the gonads of the two sides. This structure is demonstrated in Plate III, fig. 8, which 

 represents a transversal section through a radial canal with female gonads. 



The subumbrellular lines of nematocysts are clearly visible in several of the specimens exam- 

 ined. There are usually 5 7 of these lines in each octant. They issue from the circular vessel and 

 run in a centripetal direction towards the base of the manubrium; a few of the lines may reach 

 almost to the base of the stomach. It is, probably, those stripes, which were described by Wright 

 (1867, p. 42. PI. I, fig. i) as "a supplementary canal system"; Wright states, however, that they issue 

 from the sides of the stomach, running to the circular vessel, forming anastomoses with one another. 



The tentacles are hollow ; their basal part is laterally compressed. With regard to their number 

 and development I shall make the following remarks: In full-grown specimens there are about 64 

 large tentacles and about as many small ones; the latter are directed somewhat inwards (adaxially). 

 The large and small tentacles do not, however, alternate in an absolutely regular manner; now and 

 again between two successive fully developed tentacles we may find one quite small tentacle and one 

 of intermediate size, or two fully developed tentacles are placed immediately beside each other, no small 

 tentacle or tentacular bud being found between them. In younger specimens we find tentacles in all 

 stages of development, but as a rule they may be divided into three groups according to size. 

 Specimens 6 7 mm wide have, as a rule, got all of their 128 tentacles, viz. 32 fully developed, a 

 similar number of somewhat smaller, and about 64 quite small tentacles. In an individual, about 7 mm 

 wide, from the north-east coast of Iceland ("Thor" stat. 203 (04)) the mode of development of the tent- 

 acles is clearly visible from their size. Beside the 8 perradial tentacles there are, in each octant, 3 

 long tentacles, which however seem to be a little smaller than the perradial ones; further 4 tentacles, 

 somewhat smaller and still somewhat inwardly directed; and finally 8 quite small tentacles. It is sel- 

 dom, however, that the grouping in three groups of size is so distinct and regular; as a matter of 



i "Naar disse 8 Folder udbredes, soin undertiden skeer, bliver Mavens Aabning eller Munden tenimelig stor og cir- 

 kelrund. lovrigt er den heelrandet ...". 



