284 THE VOYAGE OF H.M.S. CHALLENGER 



(PI. LXXI. figs. 5-9), it is not unfrequently markedly arched, and frequently bears 

 a more or less sharply defined central papilla. The toothed outer margin extends for a 

 varying width beyond the centre of the head, or may on the other hand be reduced to a 

 simple smooth fringe (PI. LXXII. fig. 8), or even wholly disappear. The number and 

 size of the marginal teeth are very variable. The normal number, twenty, is rarely 

 greatly exceeded, and frequently not attained. I may, however, call attention to the 

 fact that amoDo- the dermal clavulse I have never found the anchor-like forms with eisfht 

 to four long curved rays, which occur so frequently on the gastral side. This marked 

 diff"erence may sometimes be utilised to distinguish the two sides on broken fragments. 

 The portion of the head below the umbel is either a simple conical or trumpet-shaped 

 expansion of the stalk, or is a thickening sometimes cylindrical or even narrowed in 

 its superior portion (PI. LXXI. fig. 9). In the latter case this swollen lower portion of 

 the head may bear a special second row of teeth, which lie parallel to the toothed margin 

 of the umbel. Not only do the length and strength of the stalk vary greatly, but the 

 shape of the lower portion, which is sometimes quite gradually pointed, sometimes 

 conically truncated or even rounded ofl". The degree of roughness is also as variable on 

 the lower as on the upper portion. 



The gastral skeleton agrees essentially with the dermal, so that I may simply refer to 

 the above description of the latter, and content myself with noting the principal 

 deviations. These difi'erences are indeed notable enough to keep us from ignoring them, 

 as Carter has done in his otherwise excellent description of the skeletal elements in 

 Farrea occa. The large gastral pentacts agree almost perfectly in shape, size, and 

 disposition with those of the dermal skeleton. It seemed to me, however, that there was 

 a more frequent occurrence of the remnant of the lost sixth ray in the form of a conical 

 or rounded boss. Often enough, however, in the gastral membrane, pentacts may be 

 seen without any trace of this remnant. 



The difference between the gastral and dermal skeletons consists chiefly in the form of 

 the clavulse, which occur here also beside the pentacts, corresponding in number, position, 

 and arrangement to those above described, but differing somewhat in form. In the 

 majority of the gastral clavulge, four or eight long, slender, more or less markedly 

 recurved hooks or teeth are formed on the very diflerently shaped head portions, and 

 produce a certain resemblance to an anchor. The head at least, and the greater part of 

 the stalk, are destitute of those roughnesses which are so frequent in the dermal clavulse 

 (PL LXXI. figs. 8, 10; PI. LXXIL fig. 9). 



This difference in shape between gastral and dermal clavulse, is not of course constant 

 nor everywhere marked to the same extent. While in some specimens, and in certain 

 regions of a single specimen, almost all the clavulse of the gastral skeleton exhibit this 

 anchor-like structure, and only a few jjrovided with the flat terminal umbels occur in the 

 dermal membrane, in other regions and in other specimens the umbel-bearing forms may 



