ON THE OEIGIN OF GENERA. 59 



the assumption of the adult dentition, since among the higher 

 Mammalia at least we know of no genus which, however similar to 

 undeveloped stages of the higher, never loses the milk dentition. 

 It is nevertheless an important fact that among smooth-brained 

 mammals, or many of them, but one tooth of the second series 

 appears ; and inasmuch as smooth-brained forms of the higher 

 orders have become extinct, it is not too much to anticipate that 

 a type of permanent milk dentition will be found among the ex- 

 tinct forms of the same high orders. 



As an example of exact parallelisms in series of series, I select 

 the following : 



1. In the batrachian family Cystignathidae there are six groups 

 or sets of genera. In the highest of these we have an ossified cra- 

 nium and xiphisternum — i. e., in the Cystignathi ; in the Pleuro- 

 demee the cranium is not ossified, thus representing the Cystigna- 

 thi while incomplete ; in the Criniae the xiphisternum is cartilagi- 

 nous, as well as the fronto-parietal region, being an equivalent of 

 a still lower stage of the Cystignathi. From this simplest type we 

 can find a rising series by a different combination of characters ; 

 thus the Ceratophydes add an osseous cranium to the incomplete 

 xiphisternum, while two succeeding groups diverge from each 

 other at the start, the Pseudes loosening the outer metatarsus in 

 their development to maturity, while the Hylodes add by degrees 

 a cross-limb to the last phalange. The Ceratophrydes and Criniae 

 are stages in the develoj^ment of these ; but neither one of them 

 is a step in the development of the other. They are measured by 

 adaptive characters purely. 



2. The whole suborder of the Anurous Batrachia, to which 

 the above family belongs, the Arcifera, differs from the suborder 

 Eaniformia by a character which distinguishes a primary stage of 

 growth of the latter from its fully developed form. That is, the 

 Eaniformia present, at one period of their development, a pair of 

 parallel or over-lapping curved cartilages, connecting the procora- 

 coid and coracoid bones, which subsequently unite and become 

 a single, slender median, scarcely visible rod, while the bones 

 named expand and meet. The first condition is the permanent 

 and sole systematic character of the Arcifera.* 



Objection. — It may be objected, by those who have observed 



* This may be rcarlily understood by comparing my monograph of the Arcifera, 

 "Jour. Ac. Nat. Sci.," Pliil., 1866, with Dugcs's work, or Gegcnbaur and Tarker's 

 memoir on the Shoulder-Girdle. 



