Paradigm Shifts In Mesozoic Mammal Evolutionary Studies

A significant conceptual advance from recent studies of early mammals is the understanding that the greatest tax-onomic diversification and morphological divergence tends to appear in the earlier periods of each major clade (Rowe, 1993; Cifelli, 2001; Luo et al., 2002). The most significant diversification of Mesozoic mammals occurred long before the splits of modern lineages of monotremes, marsupials, and placentals. It is near the root of the mammal tree that many basal branches split off, so as to generate a very bushy branching pattern of the stem clades in the Jurassic (Rowe, 1999). Based on the almost fully resolved tree of all phylogenetically relevant taxa, Mesozoic mammals have some 26 or more clades, which appeared in the fossil record by several successive bursts of diversification of stem lineages. Most of these stem lineages are dead-end evolutionary experiments that did not survive the Cretaceous-Tertiary extinction. Of all 26 or more clades (figure 1.1), only four major lineages—the mono-tremes, cimolodontan multituberculates, marsupials, and placentals—have a significant presence after the Cretaceous-Tertiary extinction.2

Most of the orders and families of the Late Triassic and Early Jurassic, such as sinoconodontids, morganucodon-tans, and kuehneotheriids, represent a hierarchical series of stems at the very base of the mammalian tree. They are distantly related to modern mammals and did not directly give rise to the major orders and families of the mammalian crown group of the Middle and Late Jurassic. (It remains unclear whether haramiyidans, which are among the earliest mammals, are related to multituberculates, which appeared and diversified considerably later; see Rowe, 1993; Butler, 2000; and chapters 8 and 15.) Similarly, the major mammalian groups of the Late Jurassic— dryolestoids, docodontans, Multituberculata (sensu stricto), and triconodontids—represent evolutionary dead ends in

2 Monotremes, cimolodontans, marsupials, and placentals are the four main groups that survived the Cretaceous-Tertiary extinction. It has been reported that one dryolestoid, otherwise known only from Jurassic-Cretaceous, is found in the Paleocene of South America (Gelfo and Pascual, 2001). There is also another report of a therapsid taxon from the Paleocene of North America (Fox et al., 1992), but this has been debated (Sues 1992).

table 1.1. A Cladistic Classification ofMesozoic Mammals at the Suborder to Family Levels

Class Mammalia Clade of (Sinoconodon + Crown Mammalia) Family Sinoconodontidae

Family incertae sedis Adelobasileus (sedis mutabilis) Clade (Morganucodon + Crown Mammalia) Order Morganucodonta

Order incertae sedis, family Kuehneotheriidae (sedis mutabilis) Clade (Docodonta + Crown Mammalia) Order Docodonta

Order and family incertae sedis Reigitherium (sedis mutabilis), Woutersia (sedis mutabilis) Clade (Hadrocodium + Crown Mammalia)

Order and family incertae sedis, Hadrocodium Clade of mammalian crown group

Clade Yinotheria (Shuotherium + Crown Monotremata) Order Shuotheridia Subclass Clade Australosphenida

Order and family incertae sedis, Asfaltomylos, Ambondro (sedis mutabilis) Order Ausktribosphenida

Stem ausktribosphenidans Ausktribosphenos, Bishops (sedis mutabilis) Order Monotremata

Family Kollikodontidae (sedis mutabilis) Clade (Steropodon + Crown Monotremata) Family Ornithorhynchidae Family Steropodontidae Family Tachyglossidae Clade (Eutriconodonta + Allotheria + Trechnotheria)

Order incertae sedis, family Tinodontidae (sedis mutabilis) Order Eutriconodonta Clade (Allotheria + Trechnotheria)1 Subclass Allotheria

Order Haramiyida (sensu Butler, 2000) Order Multituberculata

Suborder "Plagiaulacidans" Suborder Cimolodonta Clade Trechnotheria (Spalacotheriidae + Crown Theria) Family Spalacotheriidae Clade Cladotheria (Dryolestoidea + Crown Theria) Superorder Dryolestoidea Order Dryolestida Family Dryolestidae Family Paurodontidae Order incertae sedis

Family Vincelestidae Order Amphitheriida

Family Amphitheriidae Clade Zatheria (Peramuridae + Crown Mammalia)

Order and Family incertae sedis, Arguimus Afriquiamus, Magnimus, Minimus, Nanolestes, Palaeoxonodon (sedis mutabilis) Family Peramuridae Subclass Boreosphenida Order Aegialodontia

Family Aegialodontidae table 1.1. Continued

Infraclass Metatheria Cohort Deltatheroida Cohort Marsupialia

Superorder Asiadelphia Superorder "Ameridelphia" Order "Didelphimorphia" Order Paucituberculata Superorder Australidelphia Infraclass Eutheria

Order and family incertae sedis, Eomaia, Montanalestes, Murtoilestes, Prokennalestes (sedis mutabilis) Clade (Asioryctitheria + Placentalia) Superorder Asioryctitheria Clade Placentalia

Superorder Anagalida Superorder Archonta Superorder Insectivora Order Leptictida Order Lipotyphla Superorder Ferae

Order Cimolesta Superorder Ungulatomorpha Order "Condylarthra" ?Order Notoungulata

Subclass incertae sedis Order Gondwanatheria

1 We tentatively accept in chapter 8 the grouping of Allotheria, including the poorly known Haramiyida and Multituberculata. The relationship between the two allotherian groups is, however, uncertain (see chapter 15 and figures 15.1 and 15.2), and the possibility that haramiyidans and multi-tuberculates are not directly related cannot be excluded. As long as fossils of haramiyidans are (with the exception of Haramiyavia) limited to isolated teeth, this issue cannot be unequivocally resolved. On the other hand, significant differences in mandibular features between Haramiyavia and multituberculates do not support the haramiyidan-multituberculate relationship (see Jenkins et al., 1997). An alternative cladistic classification strictly consistent with the cladogram would be to place the Haramiyida as subclass incertae sedis under the class Mammalia.

their own right and are not related to monotremes, nor to crown therians. The greatest morphological divergence in the mammalian dentition occurred in the Middle to Late Jurassic, including the first of two convergent appearances of the tribosphenic molar pattern (among some southern mammals).

A series of studies in the 1980s and 1990s also demonstrated that many northern tribosphenic mammals from the late Early Cretaceous (Aptian-Albian) could not be reliably placed within either the metatherian or eutherian clades, as previously believed (compare the earlier views of Slaughter, 1968a,b, to those of Kielan-Jaworowska, Eaton, and Bown, 1979; Clemens and Lillegraven, 1986; and Cifelli, 1993b). Rather, these stem boreosphenidans appear to represent a burst of diversification of what became blind evolutionary lineages that diverged from crown Theria near the time of the marsupial-placental dichotomy (Cifelli, 1993b).

These scenarios of Mesozoic mammal evolution are not based solely on molar features, as was the case for most phylogenetic hypotheses before the 1980s; rather, they are supported by analyses of comprehensive datasets that include craniomandibular and postcranial characters, as well as those based on molars (Rowe, 1988, 1993; Lillegraven and Krusat, 1991; Wible, 1991; Rougier, Wible, and Hopson, 1996; Luo, Kielan-Jaworowska, and Cifelli, 2001; Luo et al., 2002).

Our current understanding of Mesozoic mammal evolution (figure 1.1) is significantly different from what was envisioned by Simpson, Romer, or Olson in the first half of the twentieth century, wherein the mammalian grade was achieved independently by various evolving lineages of cynodonts (e.g., Simpson, 1959; Olson 1959) (figure 1.2A). The first significant paradigm shift occurred in the 1960s, when the traditional Simpsonian view of mammalian polyphyly was replaced by the concept of mono-

phyletic origin for mammals (e.g., Hopson and Crompton, 1969). With the abandonment of a polyphyletic origin for Mammalia came general recognition of a fundamental early dichotomy of the class into "prototherian" and "therian" groups (figure 1.2B).

Interpretation of early mammalian evolution underwent two important conceptual changes in the 1970s and 1980s. The traditionally recognized groupings of "Sym-metrodonta" and "Eupantotheria" (then regarded as basal therians or, later, "holotherians") were challenged by McKenna (1975), whose cladistic interpretation suggested instead that they represent paraphyletic grades. McKenna's phylogenetic hierarchies for "therians" were largely corroborated by further studies. In the 1980s, a framework for cladistic relationships of all synapsids was established. A significant conclusion of studies by Kemp and by Rowe was that the "prototherian" grouping widely accepted in the 1970s is paraphyletic. Some aspects of these hypotheses were controversial and relevant morphological evidence has been reevaluated (Sues, 1985b; Hopson, 1991, 1994; Wible, 1991; Luo, 1994). Nonetheless, the contributions by McKenna (1975), Kemp (1983), and Rowe (1988) helped to frame the further inquiry of early mammalian evolution in cladistic terms.

The most recent conceptual shift in understanding of mammalian evolution during the Mesozoic was prompted by discoveries of new, highly derived fossil mammals from the previously blank geographic areas of Gond-wanan landmasses in the late 1990s. One piece of the puzzle, the Australian Early Cretaceous Steropodon, was discovered as early as 1985, but its full significance was not generally appreciated until much later. Cladistic and other comparative studies make it clear that the new fossils represent some previously unknown mammalian radiation(s) (e.g., Bonaparte, 1990; Krause et al., 1997; Rich et al., 1997; Flynn et al., 1999; Rich, Vickers-Rich et al. 2001; Sigogneau-Russell et al., 2001; Rauhut et al., 2002). It is equally clear that these recently discovered taxa cannot be accommodated into the prevalent view— based on evidence limited to Laurasian landmasses—that tribosphenic mammals are monophyletic and originated in the Northern Hemisphere. Beyond this, no consensus has yet emerged. In our view, Southern tribosphenic mammals of the Middle Jurassic-Early Cretaceous are not closely related to marsupials and placentals, but rather represent a radiation endemic to Gondwanan landmasses, with a probable relationship to mono-tremes. Hence, divergence of modern monotremes from modern therians may well be a vicariant event (Luo, Kielan-Jaworowska, and Cifelli, 2001; Sigogneau-Russell et al., 2001; Luo et al., 2002; Rauhut et al., 2002).

0 0

Post a comment