360 AFFINITIES OF EXTINCT SPECIES. 



for instance, the genera a}, a^, a}'^, /', m', m", m^, were dis- 

 interred, these three families would be so closely linked 

 together that they probably would have to be united into 

 one great family, in nearly the same manner as has occurred 

 with ruminants and certain pachyderms. Yet he who 

 objected to consider as intermediate the extinct genera, 

 which thus link together the living genera of three fami- 

 lies, would be partly justified, for they are intermediate, 

 not directly, but only by a long and circuitous course 

 through many widely different forms. If many extinct 

 forms were to be discovered above one of the middle hori- 

 zontal lines or geological formations — for instance, above 

 No. VI. — but none from beneath this line, then only two 

 of the families (those on the left hand, a}^, etc., and Z>^*, 

 etc.) would have to be united into one; and there would 

 remain two families, which would be less distinct from 

 each other than they were before the discovery of the 

 fossils. So again, if the three families formed of eight 

 genera (a^* to wi^*), on the uppermost line, be supposed to 

 differ from each other by half-a-dozen iuiportaut char- 

 acters, then the families which existed at a period marked 

 VI would certainly have differed from each other by a less 

 number of characters; for they would at this early stage of 

 descent have diverged in a less degree from their common 

 progenitor. Thus it comes that ancient and extinct 

 genera are often in a greater or less degree intermediate in 

 charactei' between their modified descendants, or between 

 their collateral relations. 



Under nature the process will be far more complicated 

 than is represented in the diagram; for the groups will 

 have been more numerous; they will have endured for 

 extremely unequal lengths of time, and will have been 

 modified in various degrees. As we possess only the last 

 volume of the geological record, and that in a very broken 

 condition, we have no right to expect, except in rare cases, 

 to fill up the wide intervals in the natural system, and 

 thus to unite distinct families or orders. All that we have 

 a right to expect is, that those groups which have, within 

 known geological periods, undergone much modification, 

 should in the older formations make some slight approach 

 to each other; so that the older members should differ less 

 from each other in some of their characters than do thQ 



