News & Views | Published:

Mammalian evolution

Relationships to chew over

Nature volume 409, pages 2831 (04 January 2001) | Download Citation


Did advanced mammals evolve on the southern continents and then move north? Not according to a new study, which concludes that such mammals evolved in both the south and the north.

There are three groups of living mammals — placentals, marsupials and the monotremes. The first two, along with some mammalian fossil relatives, have so-called 'tribosphenic' teeth, which provide a highly efficient way of chopping and grinding food. Monotreme ancestors are also thought to have possessed such teeth.

On page 53 of this issue, Luo et al.1 argue that mammals with tribosphenic teeth evolved not once but twice, after the supercontinent of Pangaea pulled apart more than 160 million years ago. According to their hypothesis, one lineage radiated across the southern landmass of Gondwana but is represented today by only the platypus and echidnas, which together constitute the monotremes. The other lineage, isolated on northern Laurasia, gave rise to living marsupial and placental mammals (Fig. 1a). Luo and colleagues' taxonomy is based on an extensive analysis of teeth and skeletal remains. But it is radically at odds with other interpretations of fossil data2,3, and also with an evolutionary history reconstructed from sequences of mitochondrial DNA4 (Fig. 1b, c ).

Figure 1: Proposed evolutionary relationships and origins of living mammals — placentals, marsupials and monotremes — and some fossil relatives.
Figure 1

a, Luo et al.1 retain a widely accepted theory of early mammalian biogeography, but propose that complex tribosphenic teeth evolved twice. b, This contrasts with a composite phylogenetic tree inferred from relationships favoured by Rich et al.2and Flynn et al.3, in which the fossil Ausktribosphenos is included among placentals. In this scheme, tribosphenidans would have dispersed northwards from a southern point of origin. Both a and b are based on fossil data. c, Both views conflict with that of Penny and Hasegawa4, who considered molecular data without discussing biogeographical or fossil evidence. The geographical distribution of origins shown in c is that favoured by Gregory for Marsupionta9.

In the past three years, there have been reports2,3 of two significant discoveries of early Gondwanan mammals. In both cases, the fossil animals are surprisingly advanced, given their antiquity. Rich et al.2 placed the 120-million-year-old Ausktribosphenos firmly within the Tribosphenida, a clade of mammals that includes marsupials and placentals, and that is diagnosed by the presence of tribosphenic teeth. Indeed, they suggested that Ausktribosphenos might be an early representative of the placental lineage. This seems arguable, however — Ausktribosphenos has peculiar teeth, and a jaw structure that is in some respects primitive2,5.

The other fossil, Ambondro, known from a 167-million-year-old jaw from Madagascar, was described by Flynn et al.3. They placed Ambondro within the Tribosphenida, too, fuelling speculation that the break-up of Pangaea had not been much of a barrier to mammalian dispersal, or even that Tribosphenida originated in the south. In placing Ausktribosphenos and Ambondro in the Tribosphenida, Rich et al. and Flynn et al. upheld a long-standing view that the tribosphenic dentition evolved once. Their findings contradicted previous fossil evidence for a Laurasian origin of the Tribosphenida.

The extant platypus and echidnas are specialized for feeding on invertebrates and have no teeth. But a toothed monotreme, Steropodon, is known from a partial jaw, which was discovered in an Australian opal mine and dates to between 113 million and 97.5 million years ago6. Analyses that included Steropodon indicated that monotremes were more primitive than mammals with tribosphenic teeth; they have been considered to possess a precursor morphology to the more advanced condition. Now Luo et al.1 have included Ausktribosphenos, Ambondro and Steropodon with other Mesozoic mammals in three phylogenetic analyses that include characters from the teeth and jaws, and also from the rest of the skeleton. Each analysis produces the same elegant result: the Gondwanan animals, including monotremes, are closely bound together in a clade that is widely separated from that of the marsupial and placental mammals. In this new model, mammals evolved tribosphenic teeth in parallel, on either side of a vast and widening gulf — the Tethys Sea that divided Gondwana and Laurasia.

This view of events does not require as unlikely a convergence as it might seem. Early mammals were small and endothermic (loosely speaking, warm-blooded), with high surface-to-volume ratios. They probably had high metabolic rates and correspondingly high nutritional requirements. Living shrews, which face the same constraints, have prodigious appetites. So survival probably depended on efficient food processing, and the tribosphenic dentition provides just that.

As a tribosphenic mammal bites down, a large cusp on the upper molar settles, mortar-like, into a pestle-like basin of the lower molar. Simultaneously, notched shearing crests on the sides of the triangular upper molar scissor against those of the lower molars. This combination of shearing and grinding has long been considered a key innovation in the clade containing marsupial and placental mammals — indeed, as the innovation that was possibly most significant in the spread and diversification of mammals. So advantageous is it that a third, entirely extinct, mammalian lineage, represented by the fossil Shuotherium, seems independently to have evolved a 'pseudo-tribosphenic' system that functioned identically, although the positions of the lower molar basin and shearing crests were reversed7.

Nor is the hypothesis that marsupials and placentals originated on the northern continents startling. Although most living marsupials are restricted to the Southern Hemisphere, the fossil record of the early marsupial lineage (Metatheria) is all Laurasian: the earliest metatherians are Asian, and the most diverse are North American. Marsupials are thought to have spread into Gondwana sometime between 84 and 64 million years ago, moving between Australia and South America by way of Antarctica8.

This view superseded an earlier hypothesis of mammalian relationships, espoused by W. K. Gregory before the acceptance of plate tectonics and before the discovery of toothed monotremes, Ambondro, Ausktribosphenos or many fossil metatherians. Gregory's hypothesis was that monotremes and marsupials are more closely related to each other than they are to placentals, forming a clade dubbed Marsupionta9. Ironically, as new fossil evidence makes the Marsupionta hypothesis ever more obsolete, some4, although not all10, phylogenies based on gene sequencing appear to support it.

With three competing hypotheses on the table, we can only be sure that more genes will be sequenced and more fossils will be found. What Luo et al.1 offer is a well- supported theoretical framework in which everything seems to make sense: unusual morphological features of the early southern mammals are explained in a phylogenetic context, and biogeographical considerations do not require mouse-sized animals to cross the Tethys Sea. This framework is certain to be useful, but it will also engender controversy. And, like Gregory's Marsupionta, it may eventually turn out to be wrong. The Mesozoic of Gondwana is not well known — surprises are no doubt still in store.


  1. 1.

    , & Nature 409, 53– 57 (2001).

  2. 2.

    et al. Science 278, 1438– 1442 (1997).

  3. 3.

    , , , & Nature 401, 57–60 (1999).

  4. 4.

    & Nature 387, 549–550 (1997).

  5. 5.

    , & Lethaia 31, 267– 268 (1998).

  6. 6.

    , , & Nature 318, 363–366 ( 1985).

  7. 7.

    & Austr. Mammal. 5, 127–142 (1982).

  8. 8.

    & J. Mamm. Evol. 3, 121–161 ( 1996).

  9. 9.

    Bull. Am. Mus. Nat. Hist. 88, 1–52 (1947).

  10. 10.

    et al. Mol. Cell 5, 707– 716 (2000).

Download references

Author information

Corresponding author

Correspondence to Anne Weil.

About this article

Publication history




  1. Search for Anne Weil in:

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing