Wilder.] 322 i June 21, 



bra; it appears, also, that even individuals of the same species may 

 vary in this manner, (Plioca gi-cenlandica) ; and this recalls a sug- 

 gestion already made by me (45, 15), which ought to be considered, 

 although, at present, its importance may seem rather ideal than real; 

 "it does not seem possible that the head and pelvis can be as strictly 

 homologous in animals having a different number of vertebras as in 

 those with the same number; in other words, the heads or the pelves 

 of tAvo animals may be cephalic or pelvic modifications of vertebras, 

 without being such modifications of the same identical vertebras." 

 Even if we exclude the skull from the category of vertebrae, the diffi- 

 culty is riot removed ; for if the atlas of Hyrax is homologous with 

 that of Elephas, then the sacrum of Elephas is the homologue of the 

 twenty-fourth vertebra and its successors, with Hyrax; or if we also 

 assume that the sacra of the two are homologous, we must homologize 

 29 vertebras in the one with 22 in the other; and, practically at least, 

 this seems to be our only course. 



I trust that the foregoing considerations will aid in removing the 

 stumbling block of numbers, from the path of those who would other- 

 wise accept the meketropy of pollex and primus. To my own mind 

 they were hardly needed, so decided was the conviction formed in 

 1866, and expressed in 51,-52 and 57, that no difference in the num- 

 bers of phalanges ought to affect our recognition of a profound mor- 

 phological law affecting the membra. 



Note. Dr. Coues has kindly placed at my disposal the ms. of some un- 

 published investigations bearing upon this subject, which so nearly accord 

 with my own views, that I add them here. April, 1872. 



Susceptibility of variation in numerical composition he believes to be, o, in di- 

 rect ratio of number of parts composing an organ, and &, in inverse ratio of mor- 

 phical differentiation and telical specialization of the parts of an organ ; and 

 that, consequent^, the value of numerical composition as a morphological or 

 taxonomic datum can be estimated with reasonable confidence of at least ap- 

 proximate accuracy. Value is inversely as variability. 



" It is notorious," he continues, " that an organ (whether central or periph- 

 eral — whether indispensable to the integrity of an animal, or merely a useful 

 adjunct to its economy) composed of a few parts, does riot exhibit the same per- 

 centage of variation in the number of these parts, as the same or a similar organ 

 does when it is composed of many parts. For instance, the normal variation in 

 the bones of the coccyx of Primates is at a minimum, if it be not, indeed, nil ; 

 whilst the ordinary individual variation in the coccyx of a longicaudate mam- 

 mal, such as the Jaculus hudsonius, for example, amounts to four or five coccy- 

 geal vertebra?. The few dermal scutes of armadillos are sufficiently constant in 

 number to afford specific characters, while the essentially similar but numerous 



