26 THE VOYAGE OF H.M.S. CHALLENGER. 



ends. The vertical position of the articular faces is well seen in some forms of Actino- 

 metra lineata, which has an extremely "wall-sided" calyx (PI. V. fig. 2e) ; while in 

 Actinometra paucicirra their lower portions actually slope inwards as seen in PI. V. 

 fig. 3c. The ventral faces of the radials, which in Antedon have a steep inward slope 

 (PI. I. fig. 86), are almost horizontal in Actinometra, sloping very gently inwards towards 

 the central space. Hence the opening of the funnel becomes widely expanded, and 

 when the radial pentagon is viewed from above little or nothing is seen besides the proper 

 ventral faces of its component radials. All the species of Actinometra which I have 

 examined have smaller muscle-plates than those of any Antedon except Antedon 

 macronema (PI. IV. figs. 3a, b), so that the distal faces of the radials are very low and 

 the muscular fossae often quite inconspicuous (PL IV. figs. 4a, 5c ; PI. V. figs. 1-5, b, 5c). 

 They are separated from the lower pair of fossae by fairly prominent ridges which are 

 either horizontal or curved slightly upwards. These start from the sides of the radial, 

 run inwards towards the middle line, and then turn downwards so as to leave between 

 them a wide furrow, which gradually dies away below with the disappearance of its 

 bounding ridges. No recent Actinometra has the distinct rim on the ventral side of the 

 opening of the central canal that exists in every Antedon, even in Antedon carinata 

 (PI. III. fig. la) and in Antedon macronema (PL IV. fig. 3a), perhaps the nearest 

 approach to it being in Actinometra meridionalis and Actinometra pulchella (PL IV. 

 figs. 4a, 5c), where the lower edges of the ridges bounding the intermuscular furrow are 

 somewhat thicker than usual. 



These differences in the structure of the calyx in the two chief genera of Comatulse 

 are of considerable importance. For it is only by means of an acquaintance with them 

 that the generic determination of the fossil Comatulae becomes at all possible. Every one 

 hitherto found in the Tertiary strata and in the Chalk, of which the entire calyx is known, 

 is an unmistakeable Antedon, both in the characters of the centro-dorsal and in those of the 

 radials. Antedon sequimarginata from the Gault is as clearly an Antedon as Actinometra 

 loveni from the same formation is an Actinometra. But some of the Neocomian and 

 many of the Jurassic Comatulae are less easily identified. The wide and low radials with 

 marked intermuscular furrows of Actinometra cheltonensis from the Inferior Oolite, and of 

 Actinometra ivurtembergica from the Corallian of Nattheim, indicate the generic position 

 of these types pretty clearly ; while Antedon scrobiculata with its high articular faces, 

 much narrower above than below, is an undoubted Antedon. But on the other hand, 

 the low and wide radials and thin centro-dorsal of Antedon picteti and Antedon 

 infracretacea are very suggestive of Actinometra; though in both types the articular 

 faces of the radials have a considerable slope and are altogether much like the corre- 

 sponding parts of other species which are unquestionably referred to Antedon. For the 

 present, therefore, the systematic position of these and of other somewhat generalised 

 types of early Comatulas must remain in doubt. 



