68 Miscellaneous. 



I have found an interesting form which so far unites this family 

 ■with the Chlathridae, that we find in this sponge projecting, obtuse, 

 perfectly smooth spicules in great quantity, hut no tubercular 

 epicules. 



Among the other three families which include 0. Schmidt's 

 groups of the same names, we should then have the Desmacidonidiu* 

 to distribute, 



I have proposed the foregoing classification without reference to 

 the flcsh-spiciilcs, the multifarious anchors, hooks, &c. of the Dcs- 

 macidouidse, and without taking into account the siliceous stars in 

 the cortex of many Monactinellida?. 



I regard it as preferable not to adduce the flesh -spicules in 

 the formation of tho principal groups, but rather to employ them as 

 generic characters. Although I have long been inchned to this 

 view, I have hitherto refrained from bringing it forward, as autho- 

 rities like Yosmaer and 0. Schmidt lay great stress upon the sepa- 

 ration especially of the anchor- and hook-bearing sponges and their 

 collocation to form a whole. In the same way, however, that 0. 

 Schmidt united sponges with and without siliceous stars in tho 

 family Gumminea?, I think we may bring together Monactincllid 

 families with and without anchors &c., provided they agree in tho 

 structure of the fibrous skeleton. 



I believe that these flesh-spicules are of coenogenctic origin. 0. 

 Schmidt has demonstrated that in part they contain much more 

 organic material than those siliceous structures which occur in the 

 corals. Their coenogenctic character is especially proved by the 

 extraordinary multiplicity and variability of their forms. In sponges 

 which show themselves to be nearly allied by the character of their 

 fibrous skeleton we often find anchors in one and nothing siliceous 

 in the parenchyma of another ; while, on the other baud, hooks 

 occur in very different sponges. Wo usually meet with them in 

 sponges the fibres of which contain Monactincllid spicules. But I 

 possess a well-preserved siurit-specimen of a sponge from I'ort 

 Phillip, which is a true Hircinia, and the fibres of which contain no 

 siliceous spicules, while in its parenchyma, besides the filaments, 

 great quantities of S-shaped double hooks are to bo found. 



The same applies to siliceous stars. These may sometimes occur 

 and sometimes bo deficient in nearly allied forms, while, on tho 

 other hand, they arc found in very different sponges. 



If we represent the system of the sponges in the form of a genea- 

 logical tree, taking the above conceptions as the foundation, we 

 arrive at the following conclusions : — 



1. From the Myxospongiae originates a series of forms, the cen- 

 tral members of which resemble the SpongidiB. From the sides of 

 this series branches are given off, at the extremities of which stand 

 the Aplysinaj aud Hircinise. The Chalarchidae and Chalca'nida? are 

 placed in the upper part of the series, from the end of which the 

 llenieridaD, Suberitidae, and Echispidse radiate like an umbel. The 



* com-sc I cau here refer ouly to the most important groups. 



