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



the sixth ray may also appear, occur tjrpically in the two families of the Asconematidse 

 and Hyalonematidse. Further differences between the unpaired fifth ray and the other 

 four occasionally occur. 



As in hexacts, so here curved rays often occur, and the ray may be curved through- 

 out its whole length, or only in a particular portion. The curvature is frequently 

 exhibited only by the four rays which form the cross, and these are usually curved 

 towards the unpaired straight ray (PI. LXXIV. figs. 1, 2). In many of the pentacts, 

 which project from the sponge-body, this curvature of the four cruciate rays has 

 been so effected that they have assumed an anchor form, and have, in fact, the function 

 of an anchor (PI. III. fig. 23 ; PI. XXXIII. fig. 10). 



Peculiar curvatures of a different kind are exhibited by the four cruciate and 

 tangential rays of many pentact pinuli, which adhere closely to large siliceous beams 

 of hypodermal spicules (PI. LI I. fig. 6). 



The four rays, which lie at right angles to the two radials, form an acute angle with 

 each other, and are sharply curved just at their origin in those large pentacts of Rossella 

 antarctica (PI. LV. fig. 9, 13), which have been protruded from the outer skin of the 

 lateral waU. 



In the fork- or broom-like spicules {scopulie), which frequently possess four outwardly 

 directed teeth in addition to the long stalk, I was, like 0. Schmidt, unable to trace 

 into the teeth the fine axial canals, though their cross of intersection is often very plainly 

 visible just below the forking. It seems to me improbable, therefore, that the teeth 

 can be regarded as. principal rays, and all the more since their number is by no means 

 always four or five, but sometimes six or more (PI. XCII. figs. 4, 6 ; PI. XCIV. fig. 5). 

 I should be more inclined to compare them with the terminal rays of the rosettes. The 

 invariably simple and straight stalk of the scopulse either ends in a point, or is truncated, 

 or exhibits a knob-like thickening. It is usually smooth, being but rarely provided, 

 terminally or throughout, with tubercles or transversely directed prongs. The teeth, on 

 the other hand, exhibit manifold variations in number, length, form, and position. They 

 usually arise in whorls from, or just above the knot-like swelling on the stalk which 

 contains the axial cross. They usually vary in number from four to six, but in 

 individual cases more may be present. Sometimes they diverge but slightly at their 

 point of origin (PI. LXXXIV. fig. 5), and, in other cases, somewhat widely (PI. XCVIII. 

 fig. 8). After then- original divergence they may also become parallel to one anothei", 

 and to the chief axis (PL XCVIII. fig. 9). They are frequently simple in the gastral 

 part, but exhibit in the dorsal part an S-like curvature (PI. XCII. fig. 6), or are sharply 

 bent somewhat above their origin, so that the outer portion is considerably divergent 

 (PL LXXVII. fig. 10 ; PL XCIV. fig. 5). Some are pointed (PL LXXVIII. figs. 3, 5), 

 and others truncated (PL XCII. figs. 5, 7), but most of them bear a knob-like or even 

 spherical terminal swelling, which is usually provided with numerous fine backwardly 



