W. B. Scott — Variations and Mutations. 367 



in a series of repeated parts consisting of numerous and uni- 

 form members. 



In considering the structure of the extremities, we shall 

 find the same discrepancy between the facts of variation and 

 those of phylogenetic progress. The history of the changes 

 in foot-structure has been worked out in several mammalian 

 phyla with great minuteness of detail, palseontological material 

 being particularly rich in this respect. Just as in the case of 

 the teeth, there is a marked individuality in the elements 

 which compose the manus and pes, an individuality which 

 may be traced unchanged throughout vast periods of time. 

 Indeed, we may speak with perfect propriety of the history of 

 the astragalus or of the third digit and such history might be 

 written, for many groups, with much accuracy and fullness. 

 The limits of phylogenetic change in the extremities are more 

 narrowly fixed than in the case of the teeth. Here it is dem- 

 onstrably true that " the series consisted ancestrally of some 

 maximum number, from which the formula characteristic of 

 each descendant has been derived by successive diminution." 

 In the Cetacea, however, the number of phalanges has, in all 

 probability, been greatly increased beyond that found in the 

 ancestors of the order and in the Sirenia the same process 

 seems to be in an incipient stage, but in the terrestrial mammals 

 there is no such increase of parts. In every such phylum the 

 Eocene ancestors have the largest number of carpal, tarsal and 

 digital members and from these the successive changes are 

 always by means of the gradual diminution and loss of ele- 

 ments, partly through the coossification of members originally 

 distinct, partly through the entire suppression and loss of mem- 

 bers. These changes are, as has been said, extremely gradual 

 and rudiments often linger a very long time after they have 

 lost all obvious functional importance. 



It may be objected that the phyletic series have been 

 arranged on this presupposition and that a different arrange- 

 ment would bring forth a different result. But this is not the 

 case. As we have already seen, the geological order of suc- 

 cession is a bar to any wholesale reconstruction of the phylo- 

 genies, and I know of no case in which there is any reason to 

 derive a four-toed genus from a three-toed genus. Didactyl 

 forms, it is true, do occur in the Oligocene and upper Eocene, 

 perhaps even earlier, as, for example, Anoplotherium^ Xjlpho- 

 don, I)iplojpus, Elotherium, etc., but these are all types, which 

 for reasons independent of their foot-structure, we must regard 

 as terminating their several series and as having died out with- 

 out leaving any descendants behind them. That exceptional 

 cases of the addition to parts to the extremities, or of the reac- 

 quisition of lost elements, may hereafter be discovered, is quite 



