This is, in one way, a poor time to review the iViesozoic mammals and the origin of mammals. Discoveries and studies announced but not yet published fully, if at all, are extraordinarily rich and important. They include: late Cretaceous mammals from Wyoming (McKenna and associates at the University of California) ; Triassic mammals or very advanced mammal-like reptiles from England (Kermack and Mussett; preliminary notes have been published), China (Olson and Patterson) and South Africa (Crompton; the first of a series of studies has been published) ; and tritylodonts from Arizona (Lewis and Colbert). When available these prospective works will throw a flood of new light on the whole subject and will certainly greatly modify any conclusions that can now be reached. On the other hand, this pleasing prospect (complete fulfillment of which may be long delayed) makes it all the more interesting to survey what we do know now and what it may mean, to look at the base for the next advances. There are also several recent publications of major importance that merit review and that suggest reconsideration of various points, especially: Patterson (1956), Peyer (1956), Kiihne (1958), Crompton (1958), and Kermack and Mussett (1958 a, b). This review, centered on the affinities of the known forms, will consider those formis in sequence from later to earlier: Late Cretaceous. Although this situation is due to change rapidly, at present the published data on late Cretaceous mammals have not been really significantly increased in the last thirty years. The known specimens are from Mongolia (Simpson, 1928d, and references there) and North America (especially Silmlpson, 1929b; later studies by me and others so far add little of real importance). All fall into two very clear-cut groups, multituberculates and Theria with tribosphenic molars. The latter include unquestionable marsupials (e.g., Eodelphis) and placentals (e.g., Zalambdalestes). Some of the marsupials are closely similar to living opossuLms in the dentition and lower jaw. In spite of considerable diversity they represent or include very primitive forms from which any or all later marsupials could be structurally derived. The placentals are generally classed as Insectivora. They are, however, such not in the same sense as the later, specialized insectivores, but in the sense that they lack known clear-cut ordinal characters within the Theria and that "Insectivora" has been used as a convenient catch-all for such virtually nonordinal groups. Their degree of diversity and any special relationships to known later groups are yet to be established. Early Ciretaceotus. Endotlerium Shikama, 1947, was described as a placental mammal from the Jurassic of South Manchuria. Kiihne (1958) accepted Shikama's views as to age and affinities. He concluded that placentals arose in the Jurassic from pantotheres, and that marsupials arose in the early Cretaceous from placentals. Patterson (1956),' however, pointed out that Endotherium is probably