01^ THE ORIGIN- OF GENERA. 103 



cnt geological Idstory, passed into another suborder or order hii 

 the assumption or loss of the character or characters of that to 

 or from luhich they ivere transferred, and that without necessary 

 loss of their generic characters. 



I will cite a probable case of this kind, the facts of which I 

 have already adduced. 



It has been shown above that the genera of six of the fami- 

 lies of the Batrachia anura form series characterized by the suc- 

 cessive stages of ossification of the skull, terminating in a der- 

 moossified condition, with over-roofed temporal fossae. That in 

 nearly all the other families similar relations between genera exist, 

 but are nowhere carried so far. The character attained by all 

 the first series is now only generic ; but should all the genera of 

 each of the six families assume this character in time, as is neces- 

 sary in accordance with a development hypothesis, it would at 

 once possess a new and higher importance, and would become or- 

 dinal or otherwise superior. It would define a series homologous 

 with all those types which had not attained it. This character 

 of the over-roofing of the temporal fossae has actually attained a 

 family significance among the Testudinata — e. g., as defining the 

 marine turtles ; and similar characters are found by Owen to 

 characterize the Labyrinthodontian order of Batrachia.* 



Agassiz has pointed out a similar and more extended case, in 

 the lieterocercal and Homocercal ganoids. Had we not so many 

 of the closest approximations between members of these groups, 

 they would stand in the systems as two great homologous series, 

 with their contained heterologous genera. As it is, these heter- 

 ologous terms or genera are evidently so nearly allied that Agassiz, 

 in the " Poissons Fossiles," has thought it best to arrange the latter 

 together, thus instituting a system transverse, as it were, to the 

 other. This may be necessary, since Kolliker points out transi- 

 tional forms, and perhaps certain types may have begun to aban- 

 don the heterocercal form near the period of reproduction, pro- 

 ducing offspring somewhat protean in character, preparatory to 

 an earlier appearance and consequent permanence of the homocer- 

 cal type. This is to be derived from the history of the metamor- 

 phosis of Amblystoma. 



In the same manner the development of the convolutions of the 

 brain does not define groups of the highest rank, since it pro- 



* The roof here alluded to by Owen includes some two distinct bones not known 

 in the arch of the Anura, and therefore different. 



