82 MR. p. H. CAEPENTER ON THE GENUS ACTINO^IETRA. 



geucrally more or less concealed by the calcareous network occupying the opening of the 

 central funnel. They are the ventral continuations of the live canals enclosed by the 

 radial spout-like processes of the rosette, and they pass downwards and outwards in the 

 reentering angles of the star into the intermuscular furrows on the distal faces of the 

 component radials (PI. IV. figs. 12 a, 14, 17, /,). These reentering angles, which are 

 boiuided by the superior margins of the two muscular fossa; of each radial, are more open 

 -in Aiit. celtica (fig. 1) than in Ant rosacea (fig. 17), so that the rim of tlic central funnel 

 becomes more nearly pentagonal, having somewhat the shape of a Gotiiasler. This is 

 still more marked in ^«^. Eschrichtii, while m. Actinometra \ihecova.QS ^ regular pentagon 

 (PI. V. fig. 4, PI. VI. figs. 5, 12, 23). 



In correspondence with the nearly vertical position of the distal faces of the radials in 

 Actinometra, thcu' ventral faces, which in Ant. rosacea and Ant. celtica have a very 

 steep inward slope, occupy a nearly horizontal position, sloping but very gently inwards 

 towards the central space, so that the opening of the funnel becomes widely expanded. 

 Its inner walls, formed by the adjacent ventral faces of the contiguous radials, which are 

 relatively much larger than in Antedon, are generally more or less scul^jtured out into 

 a series of radiating ridges and furrows, the number and distribution of which vary in 

 different species. 



(§ 55) In Act. i)ectinata the ventral surface of each first radial (PI. V. fig. 9 a) is 

 nearly as even and regular as that of Ant. celtica (PI. IV. fig. 4) or Ant. rosacea (PL IV. 

 tigs. 12 a, 17), and in some cases it may be even more so. It is, however, both absolutely 

 and relatively larger, as it is not encroached upon by the distal face, which stands nearly 

 at right angles to it, and the furrow [v.r.f) occuping its median line is far more distinct 

 than in either of the two species of Antedon. These points are clearly seen in a com- 

 parison of figs. 5 on Plates IV. and V., Avhich represent the ventral aspect of the calyx, 

 as seen after removal of the visceral mass, in Ant. celtica and Act. 2}ectinata respectively. 

 In the former (PI. IV. fig. 5) the second and third radials and the bases of the arms are 

 at a higher level than the pentagon of the first radials, owing to the inclination of the 

 distal articular faces of the latter ; but in yictlnometra (PL V. fig. 5) the whole ventral 

 surface of the calyx is in one horizontal plane, as the opposed articular faces of the first 

 and second radials are parallel to the vertical axis of the calyx, and not more or less 

 inclined to it, as in Antedon. 



We have seen that in Ant. rosacea the lateral margins of the ventral faces of the first 

 radials (PL IV. figs. 12 a, 17) are somewhat depressed, so that when two pieces are in 

 contact a shallow interradial groove marks their line of union on the ventral side. It 

 is deepened into a furrow {v.i.f) by the elevation at its sides of the ridges forming the 

 superior margins of the muscular fossa; of the inclined distal face. This interradial 

 depression also occurs in Act. 2)cctinuta (PL V. figs. 5, 9 a &c.) ; but as the A-ertical 

 lamellae (fig. 9 c, g) in which the muscular fossae are excavated are very small, and do 

 not extend inwards so as to encroach upon the ventral face, there are no ridges at the 

 sides of this interradial depression (fig. 5, v.i.f) converting it into a deep furrow as in 

 Ant. rosacea (PL IV. fig. 17), so that it is far less conspicuous than the corresponding 

 radial furrow (PL Y. figs. 5, 9 a, v.r.f). 



