scHUfHERT.] PRINCIPLES OF CLASSIFICATION. 117 



meauing of the deltidium and deltidial plates, and subsequently, from 

 the works of otliers, chiefly Kovalevsk}', was able to demonstrate the 

 great morphologic signifieance of the deltidium. Without any injus- 

 tice to the monumental work of Waagen — and there is no more careful 

 work oil tlie r>raehiopo<la — it can safely be asked, Were Waagen's 

 suborders based on a fundamental morphologic character of general 

 importance throughout or on ontogeny? Mesokaulia and Daikaulia 

 are the only two of the seven suborders having, as now understood, 

 the re([uired ordinal characters, and these divisions were established 

 by Waagen on the form, general expression, and the liosition of the 

 pedicle, and not on the morphologic development of the [>edicle open- 

 ing. Four of the other five suborders aie based on sui)erfamily and 

 the liftli on family characters. Five of Waagen's seven suborders, 

 therefore, are here retained as superfaniilies, and practically in the 

 sense of their author. 



Since orders are established on the nature of the pedicle opening, 

 persistent internal characters of the shell are, as a rule, used for 

 superfamily purposes. Such are the absence or presence of a spon- 

 dylium (Strophomenacea and Pentameracea, respectively); the absence 

 or presence of calcareous brachial supports, and their nature (crura 

 only in the Rhyuchonellacea, loop in the Terebratulacea, and spirals in 

 the Spiriferacea). 



Families within the superfaniilies are based upon a combination of 

 external and internal generic characters common to many genera, or 

 even to one genus. Such characters are: Outer form; nature and posi- 

 tion of muscles (Obolida^, Lingulidte, etc.); internal plates (Trimerell- 

 ida', Lingulasmatida^, Pentameridffi); peculiarities of the cardinal 

 process (Orthida^, Strophomenidiie) ; imperfection or perfection or per- 

 sistent peculiarities of ordinal and superfamily characters (Orthid;e, 

 Trematida', Discinidsie, Siphonotretida?, etc.); simplicity or complexity 

 of the jugum (Hiudellina', Diplospiriuiie, etc.); and occasionally the 

 nature of the shell structure (Rhynchospirina?). When families are 

 large it is not rare to find groups of genera having a common origin 

 which have characters in common but not differentiated sufficiently 

 to introduce new characters of family importance. In such cases it 

 is advisable to divide the family into subfamilies, which facilitates 

 systematic review and discussion. Such is the case in the large fam- 

 ilies Strophomenidaj, Terebratulidse, Terebratellidai, Spiriferidae, and 

 Athyrida'. 



No division, however, has any value unless the group contains forms 

 of but one j^hylum. A phylum, or line of descent, can not originate 

 twice. It happens, however, that the same or nearly the same combi- 

 nation of mature characters is developed along different i)hyla. When 

 this occurs the ontogeny will show it. It is therefore not correct to 

 group these different stocks as belonging to one family. For instance, 



