PEOF. W. A. HEKDMAN 02f BRITISH TUNICATA. 451 



Test thin, but tough and leathery, whitish on section ; inner 

 surface smooth and shining, adhering rather firmly to the mantle. 



Mantle very muscular, thick, and opaque. The external 

 muscle-bands run circularly and the internal ones longitu- 

 dinally, forming a complete coating. Numerous endocarps 

 project from the inner surface of the mantle. 



Branchial sac with four folds on each side, the two dorsalmost 

 distinctly larger than the rest. On each fold there are from 

 12 to 14 bars, and from 10 to 12 in the interspaces. The 

 meshes are small and nearly square, containing from 3 to 10 

 narrow stigmata, and most frequently divided by a narrow hori- 

 zontal membrane. The transverse vessels are of three sizes. 



The dorsal lamina is a plain membrane with transverse ribs. 



The tentacles simple, about 16, some very large, but varying 

 much in size and length. 



Dorsal tubercle prominent, almost circular in outline, the 

 horns slightly curled inwards, opening occasionally anteriorly, 

 but more often to the left side. 



Endostyle very broad, and considerably convoluted for a por- 

 tion of its length, sometimes for the whole distance. 



The gonads consist of one or two convoluted tubes on each 

 side, with two or three very short branches, if any, and with the 

 duct at the anterior extremity of the main tube and directed 

 anteriorly (see Pi. XXXVI. fig. 2). 



Subfamily CrNTHiiisr^. 



FOEBESELLA TESSELLATA {Forles). (PI. XXX YI. figS. 3-10.) 



A number of specimens dredged lately off the west of the Isle 

 of Man, about 9 miles off Contrary Head, depth 46 fathoms, have 

 enabled me to make a careful re-examination of this species. 

 The specimens are mostly attached to dead shells of Pecten max- 

 imus, and they present a very great range of variation in shape, 

 colour, texture, and general appearance — so much so that at 

 first I was under the impression that I bad before me two or 

 three species ; and now I can see that forms corresponding to 

 Porbes's two species Cynthia tessellata and C. liviacina are repre- 

 sented in the series, and that it would be possible to pick out 

 and describe even more divergent specimens (see PI. XXXVI. 

 figs. 3-7). 



In regard to shape, the typical form is like half a small walnut, 

 but some are hemispherical while others are nearly quite flat, the 



