212 



a cylindrical segment. In Iho cavity of every zooecium we may distinguish be- 

 tween a narrower proximal and a much wider distal part, curving outwards to 

 the surface under a right or an obtuse angle (Fl. VII, figs. 4 a, 4 f , PI. VIII, ligs. 

 1 b, 1 c, 2 b). While the distal broader part, wliich bears the aperture, readies 

 the axis of the colony, the case is dillerenl with Ihe narrower pari whieli only 

 reaches the frontal end of a separating wall, that separates the broad ends of 

 two zocpcia for a short distance, but these two zoo'cia are not situated in 

 two adjoining longitudinal rows, but in two longitudinal rows separated by a 

 third. The narrow proximal part of a zoacium, which has a triangular trans- 

 verse section (PI. VII, figs. 4 b, 4 c) and is closed proximally by the pari 

 of the distal wall, which is furnished with rosette plates, does not join the 

 corresponding part of another zoa^cium but the broad part of the zotrcia in the 

 two neighbouring rows, with which it is connected by a multiporous, in these 

 zocecia inwardly arched rosette-plate. Each zorecium has thus on either side 

 two multiporous rosette-[)lates, one arching inwards in the broad i)ait, and one 

 arching outwards in the narrow part. As a transverse section shows, the broad 

 part of a zooecium is nearest the axis separated by a separating wall from the 

 broad part of another zooecium, further outwards from the narrow part of a 

 neighbouring zooecium, and nearest the frontal side from the ooecium of the same 

 zooecium. A longitudinal section of a joint has a dilTerent ai)pearance according to 

 whether it is cut right through the axis or beside it, as in the latter case we 

 may see not only the cavities of the two zooecia, which have been mostly aflected 

 by the sections, but also a number of smaller cavities which have ari.sen by the 

 intersection of the stellate, adjoining separating walls and lead into a number of 

 intermediate zo(ccia. This may easily be seen on imagining a section carried 

 through fig. 4 c on PI. VII. Time has not permitted me to enter thoroughly into 

 the classification of this family. It may however be reasonable to suppose llial 

 the large genus Celliilaria may naturally be divided into several, possibly accord- 

 ing to ditTerences partly in the chitinous ridges surrounding the aperture partly 

 in the tooth-like processes of the latter. A generic division based only on the 

 difference in form of colony, on the other hand, I cannot acknowledge as natural. 

 Of the .species described in the work of d'Orbigny mentioned above the 

 following may, I lliiuk, be referred to this family: Ksclutid Biva (PI. (Jti.S), E. 

 Artemis (PI. 667), Escharinella elcijdns (PI. ()8I5), Esclutrella Anjus (PI. 666), Escluiri- 

 fora rhomhoidalis (PI. 684) and E. crassu (PI. ()(S4), of which the three last named 

 in contrast to the other members of this family have a smaller iunni)er of large 

 pores surrounded by a raised margin. While none of these figures show any 

 teeth in the aperture, the latter according to Waters is in E. Argus furnished with 



