THE HYDEOSPIEES AND 8PIEACLB8. 105 



as in Pentrmites and Pentremitidea (PL I. figs. 4, 5,10; PI. IV. fig. 17; PI. V. 



fig. 15). 



The spiracles of Granatocrinus differ from those of all other Blastoida in the fact 

 that they actually perforate the centre of the deltoid piece, and are altogether inde- 

 pendent of the lancet-plate and side plates of the ambulacra, neither of which take- 

 any part in their formation (PI. VII. figs. 3-6; PI. VIII. figs. 12, 17; PI. IX. 

 figs. 5, 8, 14, 15). This causes a considerahle modification in the relations of these 

 openings to the hydrospire-sacs. In most other Blastoids the spiracle is merely a 

 portion of the proximal end of the radial sinus, which is not completely filled by the 

 lancet-plate and side plates (Pentremites, PI. I. figs. 4-6 ; Schizoblastus, PI. III. 

 figs. 1-3; Metablastus, PI. 111. figs. 14, 15; Oryptoschima, PI. V. figs. 23, 24; 

 Mesob last us, PI. VI. figs. 7, 8, 13; and Orophocrinus, PI XI. fig. S). 



The opening which is left in this way is hounded, partly by the proximal end of 

 the deltoid and partly by the lancet-plate, either alone, as in Cryptdblastus melo 

 (PI. VII. fig. 15), or in combination with the side plates as in the Pentremitidae 

 (PI. I. figs. 5, 6 ; PI. V. fig. 19; PL VI. fig. 13; PL XVI. fig. 21). In all these 

 cases the hydrospire-canal which leads backwards from the spiracle towards the 

 hydrospire-sac passes over the concealed lateral portions of the deltoids, which meet 

 one another beneath the central end of the lancet-piece (PL I. figs. 5-7 ; PI. V. 

 fig. 24; PL XII. fig. 13 ; PL XV. figs. 4, 10). But in Granatocrinus this is not the 

 case at all. The lancet-plate together with the thin hydrospire-plate at its sides 

 completely fills the sinus, just as in Cryptoblastus melo (PL VII. fig. 15). But there 

 is no notch at the proximal end of the lancet or deltoid plate as in this species, by 

 which the hydrospire-canal continued onwards beneath the contiguous edges of the 

 exposed parts of the deltoids and the lancet-piece could eventually reach the exterior. 



In Granatocrinus Norwood! (PL VII. figs. 7-9) as in Acent rot re mites and in 

 Cryptoblastus melo (PL VII. fig. 15) the hydrospire-plate does not extend further 

 upwards than the suture between radials and deltoids, which last plates are com- 

 paratively small in all three types. In Acent rotremites this is the position of the 

 spiracles, while in Cryptoblastus they are nearer the peristome, as just described. 

 But in Granatocrinus the hydrospire-canal has no such means of opening externally, 

 as the lancet-plate comes into close contact with the deltoids (PL VIII. fig. 20), 

 and it finds its way to the exterior by being continued onwards through the substance 

 of the deltoids till it reaches the surface (PI. VII. figs. 3-G, 10, 11). Its course is 

 well shown in the internal casts represented in PL VI. fig. 20, and PL VII. figs. 7-9. 

 As a general rule the two tubes piercing each deltoid, which communicate with the 

 hydrospires of adjacent ambulacra, converge rather rapidly and meet enc another 

 some little distance before they reach the external opening ; but in one specimen 

 that we have seen they open separately (PL VII. figs. 12, 13), and one of the two 

 openings is not so much a perforation in the substance of the deltoid as a notch in 



r 



