Technical Comments

Comment on “Independent Origins of Middle Ear Bones in Monotremes and Therians” (II)

See allHide authors and affiliations

Science  02 Sep 2005:
Vol. 309, Issue 5740, pp. 1492b-1492
DOI: 10.1126/science.1111294

The malleus, incus, and stapes are the middle ear bones spanning from the eardrum externally to the oval window in the inner ear (1). The malleus is formed in both groups of living mammals (monotremes and therians) from the embryonic prearticular and articular (2, 3), which early in development are attached to the lower jaw (dentary). The prearticular and articular are also so attached as postdentary bones in fossil members of the mammalian lineage that show a progressive reduction in size and attachment (4). Rich et al. (5) argue that the malleus has been independently freed from the dentary in monotremes and in therians and acknowledge that “[a]ssertions of fundamental differences in development and morphology between monotreme and therian ears are no longer supported.” Therefore, fossils like Teinolophos, an extinct relative of Australia's modern platypus, are the only potential sources of conflict. Our recent study of all available Teinolophos specimens suggests that the claim of “independent origins of middle ear bones in monotremes and therians” is not supported.

Rich et al. acknowledge that Teinolophos lacks facets indicating the presence or location of the prearticular-articular. Consequently, there is no direct evidence for the middle ear bones, making their report a discussion of indirect evidence. It is the position and relation of the angular/ectotympanic, which is not a middle ear bone per se, together with the “trough on the posteromedial side of the dentary,” that are the core of the Rich et al. argument.

The identified “facet” for the purported angular/ectotympanic is not a facet. It has no limits, no textural changes, and no indication of a conspicuous area for articulation. The flattened facet is exaggerated in its distinctiveness and medial orientation in figure 2C in (5), which is a hypothetical cross section, not an actual tracing or computed tomography (CT) image. The angular facet is merely the floor of the large mandibular canal, which in broken Teinolophos specimens continues anteriorly inside the dentary at least up to the level of the antepenultimate molar and would transmit the hypertrophied trigeminal system, as in the platypus (6).

Teinolophos jaws NMV P216575, P212933, and P216680 indicate that the lingual edge of the mandibular foramen (the mandibular canal entrance) was posterior to the apex of the dentary's posterointernal angle. Therefore, most of the supposed angular facet is inside the mandibular canal [compare figure 2, A and D, in (5)]. Rich et al. acknowledge that there is no model for the extension of the angular into the mandibular canal among mammaliaforms but note that the Early Triassic cynodont Thrinaxodon shows the angular so placed to the level of the last molar. This comparison is problematic because the condition argued to be present in Teinolophos should be corroborated among other Mammalia, immediate outgroups, or embryology. That the extension of the angular into the mandibular canal could be associated with the great enlargement of the mandibular foramen (5) is an ad hoc explanation of a hypothetical condition that has no support in any mammalian model. Additionally, supposed similarities between the Teinolophos angular facet and Thrinaxodon are superficial, because in Thrinaxodon all the postdentary bones are wholly medial to the dentary and there is no mandibular foramen at the level of the angular (7).

Identification of a mandibular trough for support of postdentary bones in Teinolophos is also problematic. A deep, well-defined groove directed toward the condyle, with a prominent medial flange (8) characteristic of mammaliaforms with known postdentary bones is lacking in Teinolophos; individual facets and ridges inside the trough are also absent. Following the platypus model (6), the likely occupant of the groove in Teinolophos would be V3, the mandibular division of the trigeminal nerve. Rich et al. recognize, as do we (9), that other Mesozoic forms may question the monophyletic origin of the mammalian middle ear, but the uncertain phylogenetic position of these taxa renders the case inconclusive (9-11). Rich et al. (5) choose to interpret their findings under a particular tree topology, because “it is in accord with the polyphyletic origin of the definitive mammalian middle ear but requires the least amount of homoplasy in comparison with other proposed phylogenetic placements of monotremes.” Given that no reference, matrix, or other data is provided in support of this claim, the reasoning seems circular. Finally, the topology of the tree selected by Rich et al. does not support their claim, because upon optimization, the character “free middle ear bones” is equivocal at the base of Monotremata and present at the root of Mammalia.


Navigate This Article