M. K. A. Zittel on Fossil Lithistidce. 129 



from the internal skeleton of various Lithistid genera. Amono- 

 existing Lithistid^ the genera Kaliapsis^ Discodermia, Rhaco- 

 discula^ and Theonella approach the above-mentioned fossil 

 forms. 



The union of these corpuscles is effected as follows : — The 

 dilated and ramified extremities of two or more arms of neio-h- 

 bouring quadriradiates come close together, whereby their 

 root-like processes are so closely interlocked that the skeleton 

 cannot easily break up into its individual particles. 



In the genera Spongodiscus and Plinthosella the more or 

 less regularly quadriradiate skeletal corpuscles are remarkable 

 for their knotty structure and the slight branching of their 

 arms. Almost the whole surface of these quadriradiates is 

 beset with blunt, warty excrescences ; one of the four arms is 

 sometimes abbreviated, and their ends are somewhat thickened. 

 The axial cross in the interior consists of four short, capillary 

 canals, which may be much enlarged by subsequent influ- 

 ences*. The union of these warty quadriradiates is effected 

 by the ends of neighbouring arms being brought close together, 

 so that an apparently connected, coarsely fibrous skeleton is 

 produced. In general, in the Lithistidse, the whole sponge- 

 body (leaving out of consideration the surface-spicules) con- 

 sists of similar skeletal elements, so that it is a matter of 

 indifference from what part of them a sample is taken for 

 microscopic examination. In some Tetracladina, however 

 {Siphonia^ Phymatella, Aulaxmia), the base is distinguished 

 from the upper, true sponge-body by a different microstruc- 

 ture. In these the normal quadriradiates, furnished with 

 strongly furcate arms, become gradually more irregular below, 

 and are in part converted into elongated siliceous fibres ra- 

 mosely forked at the ends and also furnished with lateral 

 processes. Between the elongated fibres there are a number 

 of shorter branched skeletal corpuscles, which on the whole 

 may be recognized as irregular quadriradiates (PL VIII. fig. 2). 

 It is a remarkable circumstance that the root-elements, 

 which are longitudinally distorted, do not possess four axial 

 canals, but only a single and generally short central canal. 



These last-mentioned root-elements serve to unite the Te- 

 tracladina with another section of Lithistidaj, which I group 

 together under the denomination Megamoeina {/xopwy, par- 

 ticle) on account of their unusually large and elongated skele- 

 tal elements. 



In these the quadriradiate structure almost entirely disap- 



* Figures of such quadriradiates are given in my monograph of the 

 genus Coeloptychimn, pi. vii. ligs. 16-20. 



Ann. & Mag. N. Hist. Ser. 5. Vol. ii. 9 



