7()() DR. It. lutoo.M OX TniT\ r.oi OX lMiiy.'{. 



and ill the skull only the inflexion of the jiiw and the palatal 

 AiU'iiities. All the other supjiosed n:ai>upial chniacters are 

 primitive features jtrohaljly present in all early niannnals. 



On the other hand, there are one or two characters strongly 

 opposed to any association of tlie Multituberculata witli Mar- 

 supials. (1) While there is no evidence of any reduction in the 

 iiuuiher of molars, tliere are never more than three (or according 

 to Gidley two). In Tritylodoii we seem to have the full set of 

 four piemolax's and three molars. In Ctenacodon, Flagiaidax, 

 and Ptilodus we get a progressive reduction of the premolars but 

 the molars remain constant. Further, the mnltituberculate con- 

 dition is quite iinlike anything known in Marstipiaks. It may 

 have been an evolution from oi-dinary tritubercular molars, but 

 if TrUjhjpluis is really Ti-iassic then the ISIultituberculates aj)])eared 

 much before the eafliest-knowu trituberculate types and may 

 ]iossibly have been independently evolved from a prototherian 

 ancestor. (2) The presence of a well-developed septomaxillaiy in 

 Triij/lodon suggests affinity with the ]\Tonoti-emes and is opposed 

 to a close I'elationship with the Maisupials. Both Echidna and 

 Ornithorhynchus, as shown by Gaupp, have large septomaxillaries. 

 No marsupial is known to have a septomaxillary, though a small 

 one appears to occur in Dasypus, as I have shown. (3) The far 

 backward position of the glenoid cavity is also against a marsupial 

 affinity and in favour of the IMultituberculates being a more 

 primitive t}pe. 



In Upper Triassic time we have numerous Cynodont reptiles, 

 some so like mammals that it is often difficult to be sure whether 

 tliey are mammals or not. We also have the earliest-known 

 supposed mammals, repr-esented by jaws with very primitive teeth. 

 These eai-liest mammals can liardly have been other than Proto- 

 therians, and as the Multitiiberculates also appear to arise in the 

 Triassic and Triiylodon is certainly not later than Lowei- Jurassic, 

 it seems most likely that they are an ofishoot from the early 

 Prototherian group. The Multituberculates can be traced i-ight 

 on to Eocene times and form a well-marked group with no close 

 affinities to any other later mammals. The Marsupials of to-day 

 ai-e ap[)arently a lately evolved group which sprang from a di- 

 ])liy()dont ])lacental ancestor ])robably in Cretaceous times. A 

 /)id('I/Jiys-]\]<.e form is perlia]is the most jirimitive tyjie, and from 

 it liaxe l)een derived the various Polyprotodont types, including 

 ('(I'DoIfstes, which for many reasons I have long looked upon as a 

 slightly moditied ]\)ly|)rotodont, and later on in Australia have 

 been evolved the Diprotodonts. 



If we are right in conelmling that the Diprotodont uiai'supials 

 originatx^d in Australia in Tertiary times from a I'olyprotoihmt 

 ancestor which was itself derive<l in later Me.sozoic times from a 

 diphyodont placental, it is difficult to believe that the Multi- 

 tuberculates which originated in Ti-ia.ssic times can be in any 

 way nearly related to them. In- the present state of our know- 

 ledsie it seems wisest to leave the Midtitul erculata as a distinct 



