1618 THE VOYAGE OF H.M.S. CHALLENGER. 



lead into the six connected tangential tubes. In Aulosphcera and Auloscena, where a 

 radial spine arises at each nodal point, its axial filament arises from the centre of the star, 

 and is usually prolonged on the inside into a free centripetal beam (PI. 109, figs. 6, 7; 

 PL 110, figs. 4, 6). This centripetal beam is a slender, very thin, conical tube (often 

 with a six-sided pyramidal base) and perhaps of great morphological interest as a rudiment, 

 being homologous to the hollow radial beams which connect the two concentric spheres of 

 the Cannospheerida. Possibly the Aulosphserida have been derived from the latter by loss 

 of the inner shell and reduction of the radial beams ; the centripetal beams which are found 

 in many (but not in all) Aulosphserida, may be the last remnant of those radial beams. 

 They contain also an axial filament, as a direct internal prolongation of that of the 

 external radial spine. The axial filaments of the radial spines are usually connected with 

 their thin wall by a variable number of scattered transverse threads, or sometimes even 

 by thin transverse septa (PI. 110, figs. 4-7, &c.). This structure corresponds to the 

 similar axial filaments and their thin filiform transverse branches which we have found in 

 other PR/EODARIA, e.g., in the Tuscarorida and Circoporida. 



The Aulonida (with polygonal meshes) exhibit in general the same structure as the 

 Aularida (with triangular meshes). But whilst in the latter usually six tangential tubes 

 are united at each nodal point, in the former only three or four tubes become united. 

 Therefore also the number of the small sutural partitions or astral septa, which radiate 

 from the central openings of the nodal cavity, is six in the latter, three or four in the 

 former. Correspondingly three or four small pores lead from the nodal cavity into the 

 surrounding tangential tubes. 



The true nature of the nodal points, and the intimate structure of the different parts 

 here united, is a matter which it is very difficult to explain. It is a certain fact, well 

 established by hundreds of observations, that in the complete and well-preserved skeletons 

 which are perfectly purified by fire, or by hot mineral acids, and afterwards dried, all the 

 tubes of the skeleton, the tangential as well as the radial cylinders, become filled up by 

 air. Each tube contains usually one large cylindrical air-bubble, with two hemispherical 

 ends. But the air-bubbles of the neighbouring tubes are completely separated one from 

 another by the thin astral or sutural septa, and the air-bubble of the radial tubes is also 

 separated from the former. The central cavity of each nodal point is therefore surrounded 

 in the Aularida by six, in the Aulonida by three or four separate cylinders of air. This 

 fact seems to be explained only on the supposition that each single tube has two terminal 

 pores or fissures, which open into the two nodal cavities on its two ends. The radial 

 tubes must also possess at least one small opening, probably on their base, and probably 

 they have another on their distal apex. In no other way can it be explained, that 

 in all complete, well-preserved and purified skeletons, each single tube constantly becomes 

 easily filled by an air-bubble after drying. 



In the living Aulosphserida the cavities of all tubes are filled up by a jelly-substance, 



