384 bulletin: museum of comparative zoology. 



narrow limits drawn for them by that author. In most of the speci- 

 mens they reach aborally to about the level of the funnel (Plate 1, 

 fig. 1), as in Chun's specimen. But in some they are much shorter, 

 the precise length varying with contraction. On the other hand the 

 tentacle-bases in flaccid specimens may surpass the funnel aborally, 

 and extend orally almost to the ends of their sheaths : i. e. in examples 

 in which the body as a whole is shrivelled or contracted and the 

 tentacle-bases expanded, we may have precisely the characters of 

 amboinae. The bases of the tentacles are usually bowed outward, 

 i. e. away from the gastric cavity, at their mid-points (Plate 1, fig. 1), 

 as figured by Moser ( : 08a) for japonica. But they vary from this 

 condition to one in which they are practically straight (Plate 1, fig. 2) : 

 in a rather flaccid specimen 12 mm. long, only one of the bases was 

 bowed outward, and in two of the large Japanese specimens both 

 bases (Plate 1, fig. 6) showed the double curve figured by Chun ('98). 

 These facts demonstrate that there is no ground for making the curve 

 of these organs a distinguishing feature between japonica and pal- 

 mata. The tentacles are contracted, but many of them still bear 

 considerable numbers of tentilla, all of the filiform type though vari- 

 ously shrivelled by preservation (Plate 1, fig. 2). 



An important specific character in Hormiphora is afforded by the 

 level at which the adradial canals join the meridionals. In all the 

 present series the junction takes place at about the level of the funnel 

 or very slightly above it (Plate 1, figs. 1, 3), and the same is true of 

 the specimens figured by Chun and by Moser. 



Meridional canals and paddle-ribs. The ribs are as a rule rather 

 shorter than the meridional canals; but the difference in length varies 

 from practically nil (Plate 1, fig. 5) through a continuous series of 

 intermediates, to a very considerable one (Plate 1, fig. 6) ; Chun's and 

 Moser's figures all fall within the limits of variability of the present 

 series ; therefore the slight difference in this respect between japonica 

 and amboinae on the one hand, and Chun's specimen of palmata 

 on the other, is not specific but merely individual. It does not seem 

 to be influenced to any great extent by contraction. The ribs, as 

 noted by Moser ( : 08a) commence close to the apex, and end slightly 

 below the oral end of the tentacular bases. As pointed out above, 

 the ends of the meridional canals bear a similar relation to the ends of 

 the tentacular sheaths. It is impossible to make any exact statement 

 as to the proportional length of ribs to body as a whole, because 

 the latter varies so much with the degree to which the mouth or 

 pharynx is protruded or contracted. In several of the largest speci- 



