Early mammalian radiations

Journal of Paleontology, Nov 2001 by Cifelli, Richard L

Haramiyida, Theroteinida, and Eleutherodontida.-The firstknown of the earliest mammals are peculiar taxa discovered in the 19th century and now referred to the Haramiyidae (see Simpson, 1947). Then-known specimens, isolated teeth only, were reviewed by Simpson (1928), who considered haramiyids to be probable mammals and very tentatively suggested a possible relationship to Multituberculata.

Haramiyids vaguely resemble multituberculates in having rows of cusps on their cheek teeth, but even orienting these baffling specimens was problematic until a large sample was described from a single, Triassic locality in France (Sigogneau-Russell, 1989; Butler and MacIntyre, 1994). Resemblances of haramiyids (placed in their own order by Hahn et al., 1989) to multituberculates appeared closer with the description of primitive multituberculates from the Jurassic of Portgual (Hahn, 1973) and the recognition that haramiyids, like multituberculates, apparently chewed with a distinctive, longitudinal movement of the jaw (Butler and MacIntyre, 1994). Matters are further complicated by consideration of two other early mammal groups with vaguely similar multicusped teeth: the Late Triassic Theroteinida (Sigogneau-Russell, 1983; Hahn et al., 1989) and the Middle Jurassic Eleutherodontida (K. A. Kermack et al., 1998). These, like haramiyids, are known principally or only by isolated teeth. The most important recent discovery relevant to one or all three of these groups comes from the Triassic of Greenland. Haramiyavia, known by jaws, cranial pieces, and a partial skeleton, was initially interpreted as a haramiyid (Jenkins et al., 1997) and later compared favorably with Theroteinidae (K. A. Kermack et al., 1998). It presents a few multituberculate-like characters, though not, apparently, longitudinal jaw movement. Some recent studies tentatively recognize all of these groups as a monophyletic group, Allotheria, that also includes multituberculates (see Butler, 2000, and references therein). The rationale for considering these enigmatic and poorly known mammals in the present context is that this hypothesis is difficult to reconcile with the stratigraphic record and most of the recent, comprehensive hypotheses of mammalian relationships. These studies commonly place multituberculates high on the mammalian tree, usually within the crown group formed by living taxa (e.g., Rowe, 1988; Rougier et al., 1996a; Luo et al., 2001). Earliest evidence of crown mammals is Middle Jurassic (Rowe, 1993). But haran-tiyids are known from beds as old as Norian: hence, if they are related to multituberculates, an astonishingly early divergence of crown mammals-not to mention a series of putative sister taxa to crown mammals-is implied.