As the existing monotremes in many respects represent a spe cialized side-group of the earliest mammals, the marsupials, on the whole, are the most central group of the mammalian class; consequently the problem of the origin of the marsupials largely resolves itself into the problem of the origin of the mammals, which is treated more fully under MAMMALIA. Nevertheless we present here a short review of the palaeontologic data bearing on the possible relationships of the marsupials to the several Mesozoic orders with which they have been thought to be connected.
The multituberculate or allotherian mammals (see MULTITU BERCULATA), which range in time from the Upper Triassic to the Lower Eocene, appear to have been a side branch of the primi tive mammalian stock, paralleling the monotremes on the one hand and the diprotodont marsupials on the other but not closely related to either.
The fossil mammals called triconodonts from the Jurassic of England and Wyoming have usually been referred to the mar supials, but according to the recent studies by Simpson the tri conodonts formed a special group of their own, with no valid claims to relationship with the marsupials. These were small predatory mammals, the largest about the size of a cat, with jagged cheek teeth, each molar crown comprising three cusps ar ranged in a longitudinal line. A braincase of Triconodon reveals an extremely primitive stage with large olfactory lobes and unex panded forebrain. On the whole, the triconodonts stand on a lower level of evolution than any true marsupials.
Another Jurassic group, the Trituberculata, includes various small lower jaws, of which the most widely known has received the name Amphitherium prevostii. One of these jaws from Oxford
was examined by Cuvier, who noted its resemblance to the jaw of the opossums but also noted that it had ten teeth on each side behind the canine, a greater number than that of any carnivorous marsupial. In 1888 H. F. Osborn pointed out that the lower molar teeth of Amphitherium consisted of two moieties : in front an elevated triangle of sharp cusps and behind a low heel or talonid; here clearly was a distinct prophecy of the famous "tuberculo-sec torial" lower molars of Eocene mammals. Whatever its precise relationships, the lower jaw of Amphitherium could be changed into that of a primitive marsupial by the inturning of the angular process, by the reduction in the number of the cheek teeth from ten to seven ; and by the further development of the heel of the lower molars. Even less change would be required to transform it into the jaw of a generalized placental mammal, in which the angular process is not inflected.
The American Opossums.—The oldest known fossil which may be referred without any doubt to the group of marsupials is a fragmentary skull and lower jaw from the Upper Cretaceous of Montana, described by W. D. Matthew in 1916 as Eodelphis browni; this was discovered by Barnum Brown in association with the skull of a Cretaceous dinosaur. As Matthew pointed out, the pieces of the skull and the lower jaw match closely the corre sponding parts of existing opossums, which thus take their place among the most primitive known mammals. Matthew has also traced the fossil history of the opossum family through several Eocene formations of Wyoming into the genus Peratherium, which is found in the Oligocene formations of western North America and France. In South America fossil opossums of the genera Microbiotherium and Proteodidelphys, date from the Lower Mio cene and earlier formations. From these and similar facts Matthew has pointed out that the family of the opossums, like that of many other families of mammals, appears to have origi nated in North America and then to have spread on the one hand into Europe and on the other into South America. Possibly the common Virginia opossum, together with related genera in Mexico and Central America, may represent a reflux wave of immigrants coming from South America in Pliocene times.