6o 



shewiiiij: the continuity which Masterman describes between each one of them and two of the 

 arms. On the contrary, the arm-j^rooves, in man)' favourably orientated specimens, die away in 

 approaching the mouth, and tht- inner surface of the arm-base becomes smoothly convex in 

 places where grooves ought to occur on Mastermax's hypothesis. This is seen, for instance, 

 in figs. 143—147 a'l- XII), figs. 66, 65 (PI. VI) and figs. 47—45 (PI. V). 



In order to obtain as much certainty as possible with regard to the relations of the 

 oral grooves, I made plasticine reconstructions of two individuals of C. gracilis, namely those 

 from which fig. 25 and figs. 62 — 70 were respectively drawn. 1 failed in these to find any 

 evidence of continuity between the oral grooves and the food-grooves of the arms. 



It may further be remarked that it is improbable, on a priori grounds, that a system of 

 small grooves could be effective in all the parts where they have been described by Master.man, 

 who has himself (98, 2, p. 509) seen the difficulty of supposing that they could direct the flow of water. 



I come next to the pair of "ventral grooves" described by Masterman (98, 2), and 

 marked v.g. in his fig. 93. These grooves, which are supposed to start as the "ventral" or 

 anterior grooves of the pharyngeal diverticulum, jjass down the sides of the mouth, just inside 

 the oral aperture, and can then be recognised "by the fact lliat they are situated exactly internal 

 "to the line of junction of the mesentery between collar- and trunk-cavities, with the gut wall". 

 They are thus shewn in M.\sterman's fig. 93, though not on the left side of fig. 16 of his 

 earlier paper (97, 2), to which a reference is expressly given. In this latter figure, the left 

 groove lies entirely in the region of the collar-cavity. 



The series of sections from which figs. 152 — 157 are taken shews the "ventral" grooves 

 in an unmistakeable form in the pharyngeal diverticulum (figs. 152, 153); but on pa.ssing to 

 the mouth they communicate with the oral grooves (figs. 154, 155), and no trace can be seen 

 of them in fig. 156, which passes through the mouth. I am satisfied that the grooves are really 

 absent in this section, but the discrepancy between Masterman's results and my own is ncjt 

 hard to explain. The parts of the "ventral grooves" which Masterman shews at the sides of 

 the moutli appear to me to be outside the mouth, and to belong to the recess (PI. I\', 

 figs. 34, 42, op.rec), external to the oral a]jerture, which is formed by the base of the operculum. 

 The recess may altogether disappear when the operculum is straightened out and its free edge 

 is directed ventrally, as in figs. 156, 157, but it becomes conspicuous in certain other positions 

 of the operculum. 



Figs. 119 — 123 (PI. X) are frontal sections of C. levinseni which shew what I think 

 there can be no doubt are the parts of Masterman's "ventral grooves" that are supposed to 

 e.xtend along the sides of the mouth. In fig. 123, the groove on the right side is continuous 

 with a part of the basement-membrane extending, from the anterior horn of the third body- 

 cavity (d.c.^a.), between the collar-cavity {/>.c.') and what may be an oral vascular sinus {or.s.). 

 On the left side, the wall of the groove projects into the collar-cavity. These are so much like 

 the relations shewn by Masterman in fig. 16 of his earlier memoir (97, 2) that there can be no 

 reasonable doubt that the grooves are the ones allu^i^d to by him. lint a further consideration 

 of figs. 119 — 123 and of the other sections belonging to the same series shews that the two 

 "ventral grooves" represent the junction of the base of the operculum with the mouth, and that 



