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



collected and cemented together into definite fibres as in Desmacidon reptans, nobis, 

 or they may be glued on to the surface of the sponge without any definite arrangement, 

 as in Polymastia agglutinans, nobis (PI. XLI. fig. 6 ; PL XLII. fig. 1). (The main 

 skeleton also may be replaced to a greater or less extent by foreign bodies, but this 

 subject will be treated of more fully in our discussion of the so-called " Phoriosjwngm.") 



We ought not to leave this part of our subject without a few words as to the arrange- 

 ment of the mierosclera. These, as we have already pointed out, do not, as a rule, take 

 part in the formation of the skeleton proper, and their function is in most cases extremely 

 diflicult to determine. Occasionally, however, it is sufiiciently obvious, and in such cases 

 it nearly always appears to be protective.^ Thus we know of several instances in which 

 hooked mierosclera are arranged around the walls of the canals, with one end embedded in 

 the wall and the other projecting freely into the lumen of the canal. In Esj)erella 

 murrayi, nobis, the inhalent, and apparently some of the exhalent canals also, are plenti- 

 fully armed with sigmata, each with one hook projecting into the canal (PI. XLVIIL 

 figs. 2, 2c), and in lojylion chelifer, nobis, we have observed the very peculiar bipocUli of 

 that species arranged in a similar manner. In these cases there can be little doubt that 

 the mierosclera in question serve to protect the sponge from the ingress of noxious 

 parasites, such as minute Crustacea, to whose attacks sponges seem to be peculiarly 

 subject, by way of the canal system ; just as the dermal skeleton serves to hinder 

 these creatures from boring their way into the sponge at any part of the surface, which 

 is a common occurrence in cases where the dermal skeleton is not sufficiently strongly 

 developed. The arrangement of special mierosclera {e.g., amphiastra in Axoniderma, 

 discastra in Latrunculia, spirulse in Spirastrella) to form a dermal armour, has already 

 been referred to and needs no further comment. 



More difficult to understand are certain cases in which the mierosclera are attached to 

 the fibre of the proper skeleton. This arrangement has been described by Bowerbank * 

 in the case of the diancistra oiHamacantha johnsoni, and the diancistra of our Vomerula 

 esperioides are arranged in a similar manner, each being cemented on to the spiculo-fibre 

 by the back of the shaft, while the sharp, cutting teeth project outwards. Moreover we 

 have observed a similar arrangement of the large anisochelse of E.spereUa simonis, nobis, 

 the spicules being attached to the skeleton fibre in groups by their small ends. Probably 

 in these cases also the spicules in question are really defensive. 



It is much more common, however, for the mierosclera to occur simply scattered 

 irregularly through the mesoderm or in the limiting membranes without any pretence of 

 arrangement, and they are often so minute and insignificant that we cannot believe 

 that they now fulfil any function whatever. 



1 We have, however, suggested elsewhere (p. xxxi) a different function for the triehodragmata, viz., that they serve, 

 like straw in mortar, to bind together the soft gelatinous tissues in which they lie. 

 - Men. Brit. Spong., vol. i. p. 35. 



