212 Bulletin American Museum of Natural Hidory. [Vol. XXVII, 



There can scarcely be much doubt that the lower jaw designated as 

 Parabderites bicrispatus by Ameghino (/. c, Fig. 74), from the Colpodon 

 Beds, pertains to a Caenolestoid. Now the jaw and teeth of Parabderites 

 are extremely like those of ArcJiaodoIops from the Notostylops Beds (Ame- 

 ghino, /. c, Fig. 75) and these offer a close comparison with the Polydolops 

 teeth, which in turn lead readily and unmistakably into all the other "-dolops" 

 genera of the Notostylops Beds. None of these genera show any clear 

 evidence of derivation from tritubercular and tuberculo-sectorial types, but 

 on the contrary certain of them {Propolymastodon, Pliodohps) approximate 

 the Multituberculate type, as represented in Ptilodus and Polymastodoii , 

 in the following characters: 



(1) The angle in PropoJipnastodon is strongly inflected. This, how- 

 ever, would be true of almost any INIarsupial. (2) The second incisors 

 (Ameghino) are much enlarged and subscalpriform, with .the enamel con- 

 fined to the anterior edge; they are followed by a diastema; the remaining 

 incisors and canines having apparently disappeared. (3) The last pre- 

 molar is a high pointed tooth, grooved anteriorly, larger than in Polymastodon, 

 but smaller than in Meniscoessus. (4) The molars are elongate-oval in 

 crow^n view, presenting two rows of tubercles on the inner and outer edges 

 respectively, and separated by a deep median longitudinal valley. (5) 

 Pg in Poli/dolops is reduced to a small peg at the base of the enlarged pj, 

 very much as in Meniscoessus. 



To offset these resemblances we have chiefly the following differences, 

 which are all such as might separate more advanced from more primitive 

 forms. 



The North American Polymastodonts have advanced beyond Propoly- 

 mastodon : (a) in the loss of nig; (b) in the strictly rectangular outline of the 

 base of the cusps in Polymastodon and in the crescentic shape of the cusps 

 in Meniscoessus (c) in the presence of a third external row of cusps in the 

 upper molars, whereas in Polydolops there are only two rows. In Plio- 

 dolops, it is true, there are three rows but the cusps are very irregularly 

 arranged and of polygonal outline at the base, whereas in the Multituber- 

 culates they are arranged in very straight antero-posterior rows and are 

 sharply rectangular at the base. 



There are, however, certain objections against taking these resemblances 

 and differences at their face value, as Dr. Ameghino does, and concluding 

 from them that the Polymastodontidae are derived from the Promysopidse 

 (Promysops and Propolymastodon) : 



(1) Conclusions as to the genetic relations of orders, when based on 

 dental remains alone, should always be accepted with caution (pp. 107, 108). 



(2) Marsupials in general and especially South American families of 



