S2)ongc-fauna of Norway. 445 



immediately on its under surface very definite minute round 

 nucleolated nuclei 0'000125 inch in diameter, each of which 

 is situated in the midst of a cluster of fine granules ; it is 

 clearly a layer of pavement-cells from which the cell-outlines 

 have disappeared. Very fine fibrils are usually apparent 

 Avandering over its lower surface ; they are probably the 

 tenuous ends of branching processes extended fi-om the cor- 

 puscles of the underlying connective tissue. The ectoderm is 

 continued inwards as an epithelial Hning (PL XVII. fig. 47) 

 to the incurrent canals or vesicular system, from no part of 

 which is it absent. In describing the ectoderm of Tetilla we 

 stated that the characteristic minute spicules of the sponge 

 (hamates) appear to contribute to its composition ; similar 

 components appear also in the ectoderm of Thenea. The 

 spinispirules which in this sponge represent the hamates of 

 Tetilla are associated, wherever they occur, with a small round 

 nucleus, which lies close to their shaft between two of its 

 spines j Avhen the shaft is curved the nucleus lies in its con- 

 cavity (PI. XV [I. fig. 24). The nucleus of the spinispirules 

 is undistinguishable in character from that of the ectodermic 

 cells; and in many cases one can see in the epithelium lining 

 a vesicle a nucleus otherwise precisely similar to its fellows, 

 but here embraced by the concave shaft of a spinispirule, and 

 so closely as to show that it belongs to the spicule, which on 

 its part lies so near to the epithelium that its minute spines 

 project through it (PI. XVII. fig. 47). The nucleus is clearly 

 a part of the epithelium ; but likewise it belongs to the spicule; 

 and thus it would appear that the spicule is a genuine com- 

 ponent of the epithelium. But spinispirules in association 

 with epithelial and epidermic nuclei are far from uncommon, 

 indeed remarkably frequent ; so that we are led to conceive of 

 these membranes as to a considerable extent composed of 

 spicule-bearing cells. Further, as in Tetilla, we are brought 

 to the alternative of regarding the ectoderm as a skeletogenous 

 tissue, or of admitting that mesodermic cells may find their 

 way into it and contribute to its formation. 



Endoderm. — This lines the excurrent system of vesicles as 

 an epithelium which does not differ from the ectoderm except 

 when it forms the walls of the tlagellated chambers. The 

 flagellated cells, in their present state, are rounded or oval 

 bodies 000125 iiich in jliamcter, with a well-marked round 

 nucleus containing a nucleolus. They are seated on the walls 

 of the chamber, about O'OOOll inch remote from each other 

 on the average, and number about forty to a chamber. Some- 

 times one is to be observed markedly larger than the others, 

 0*00028 inch in diameter ; and sometimes a little heap of four 

 Ann. & Mag. N. Hist. Scr. 5. Vol. ix. 31 



