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



different disposition, that one would take them for an entirely new formation. And if 

 in Sycandra compreasa the individuality of the radial tubes still finds its expression 

 in the disposition of the acerate spicules, the radial tubes in Sycortis laevigata or Sycctta, 

 cupula, enclosed between two distinctly parallel layers, that of the cortex and that of 

 the gastric surface, show as yet not the slightest trace of any independence. The tubar 

 skeleton of these latter sponges is still articulated ; but, owing to the presence of an 

 independent cortex, its transformation into the non-articulated, under certain conditions, 

 may be very easily imagined. 



Let us admit that some of the dermal triradiate spicules, for instance, in Sycortis 

 Icevigata, develop a fourth apical ray — this ray will have a centripetal direction ; and 

 again let us imagine that the subgastric triradiate spicules grow larger, so that their eentri- 

 fugally directed basal ray approaches more or less nearly the dermal extremity of the tube ; 

 it is evident that the tubes thereby receive a new system of supporting spicules, which 

 render the former one — the articulated skeleton — superfluous. There are not always 

 present, however, quadriradiate spicules, which, lying with their facial rays in the plane 

 of the dermal surface, support with their apical rays the radial tubes. Such spicules may 

 be triradiate also, as is the case with Sycaltis glacialis (Kalkschwiimme, Bd. iii., pi. 

 xlv. fig. 5), Amphoriscus poculum, n. sp. (PI. IV. fig. 4), &c. 



Now I have found — and this is a very interesting fact — that in Amphoriscus poculum 

 and in Amphoriscus flamma (PI. V. fig. 36), these subdermal triradiate spicules have pre- 

 cisely the same form as those of the cortex, the former differing from the latter only in 

 the fact that one of their rays exceeds the others in length. According to the position of 

 these subdermal triradiate spicules, their longer ray being directed centripetally, and their 

 two remaining rays diverging towards the dermal surface, one would say that their longer 

 ray is the basal. This, however, is not admissible. Its length being variable, there are in 

 the species just named many subdermal triradiate spicules, which, if removed from the 

 soft parts of the sponge, would certainly be confounded with the spicules of the cortex, 

 and, as their position also is not quite constant, their rays lying sometimes in a plane 

 forming an acute or an obtuse angle with the longitudinal axis of the sponge, the only 

 deduction possible is that these subdermal triradiate spicules are nothing but spicules 

 of the cortex, modified with respect both to their form and to their position, and that, 

 consequently, their longer centripetally directed ray is one of the lateral rays. This con- 

 clusion is of great moment, for, in connection with other anatomical and emhryological 

 facts, before communicated, it presents a conclusive proof that the Sycones with non- 

 articulated tubar skeleton owe their origin to Sycones whose tubar skeleton was articu- 

 lated. The transformation of the first type into that characteristic of the family Leucones 

 can now be followed step by step. 



There are amongst the Sycones which are characterised by a non-articulated tubar 

 skeleton, forms in which the radial tubes, instead of ending each with its own gastric 



