Loop-hearing JBrachiopoda. 389 



which in other genera are very delicate. Bouchardia, 3Iagasella 

 (M. Cicmingi), and Agulhasia, have excessively thickened hinge 

 plates and muscular fulcra; the deltidial plates are obsolete, but the 

 pedicle is enclosed by the growth and overlapping of the .edges of 

 the delthyrium. Megathyris and Cistella deposit lime on the inte- 

 rior of the valves, often in the form of nodes and ridges, while the 

 pedicle in its growth encroaches upon the beaks of both valves to 

 the final elimination of the deltidial plates. Platidia represents the 

 extreme of this encroachment. Thus, there are incomplete ty^jes of 

 brachial development in these genera, accompanied by positive evi- 

 dences of senility and retrogression in other shell characters. 



Classification. 



In the light of the previous discussion, the classification of the 

 families and genera of Ancylobrachia, with the exception of the 

 fossil forms, is comparatively simple. Among the extinct genera of 

 the Terebratulidge, or the family in which the loop is a development 

 of the descending branches, Centronella and Rensselceria belong to 

 one sub-family ; IStringocephalus to another ; and Megalanteris, 

 Cryptonella, JDielasma, Dictyothyrls, and others, will come in the 

 Terebratulinse. 



As for the long-looped mesozoic forms, it is evident that many 

 of them underwent metamorphoses in their development while 

 attached to a dorsal septum. The presence with them of such 

 genera as Kingena, Ismenia, and Muhlfeldtia, points to their inti- 

 mate relations with the boreal stock of the Terebratellidae, and 

 therefore indicates that the mesozoic species with long recurved 

 loops will in all probability be found to agree with Macandrevia, 

 JEudesia, Dallina, Trigonosemus^ JLyra, Terehratalia, Laqueus, 

 Kingena, Ismenia, and Muhlfeldtia. Magas is exceptional, as it 

 belongs to the austral group of genera, and suggests the likelihood 

 that some of the higher species and genera associated with it geo- 

 logically may prove to belong to the same branch. 



The septum in the TerebratellidiB appears before the branches of 

 the loop, and is an important character in the series of metamor- 

 phoses until the last stage. At first it is merely a support to the 

 introverted growing cirrated portion of the lophophore. This elon- 

 gation of the cirrated margin next results in an arch on the septum 

 in the platidiform and ismeniform stages. Furthe^ elongation pro- 

 duces the median coiled arm, and the structures impeding its growth 

 become resorbed, so that the transverse space between the lamellae 



