Evolution of tribosphenic molars—uppers with three principal cusps arranged in a triangle (the trigon), the buc-
cal paracone and metacone and a lingual protocone; and lowers with a three-cusped triangular trigonid and a basined talonid for occlusion with the protocone (see Fig. 2.2)—was one of the most important anatomical innovations in mammalian history. It laid the stage for the great diversity in dentitions of therian mammals. Tribosphenic molars can grind as well as shear food. With this dental structure, mammals were able to expand widely into omnivorous, herbivorous, and other specialized dietary niches.
Until quite recently it was assumed that tribosphenic molars evolved only once in mammals, first appearing in the Cretaceous Northern Hemisphere aegialodontids and eventually leading to marsupials and placentals. This established dogma has been challenged by the discovery of several apparent tribosphenic mammals from southern continents, which do not seem to fit the widely held model. In contrast to the conventional view, which assumes a monophyletic northern origin of all tribosphenic mammals, two competing hypotheses have been advanced: that tribosphenic mammals evolved first in the Southern Hemisphere, much earlier than previously thought (e.g., Flynn et al., 1999; Sigogneau-Russell et al., 2001; Woodburne et al., 2003), or that tribosphenic mammals evolved independently in the Northern and Southern Hemispheres (e.g., Luo, Cifelli, and Kielan-Jaworowska, 2001; Luo et al., 2002; Rauhut et al., 2002; Kielan-Jaworowska et al., 2004). According to the latter hypothesis, marsupials and placentals derive from the northern radiation, whereas monotremes are the only extant remnants of the southern radiation. It is not yet clear which, if either, hypothesis is correct.
Southern Tribosphenic Mammals: Australosphenida (Monotremes and Extinct Relatives)
Several very ancient tribosphenic mammals, all based on lower dentitions, have recently been reported from the Southern Hemisphere, complicating what had seemed to be a relatively straightforward record of the origin of tribosphenic mammals on northern continents. They are regarded as tribosphenic because their lower molars have three-cusped trigonids and the talonids have three peripheral cusps bordering a talonid basin, implying the presence of an upper molar protocone. There is currently no consensus on whether the tribosphenic condition in these mammals is homologous with that of Holarctic therians. Luo and colleagues (Luo, Cifelli, and Kielan-Jaworowska, 2001; Luo et al., 2002) offered the intriguing hypothesis that these animals— Ambondro, Ausktribosphenos, Steropodon (Fig. 4.18), and related forms including extant monotremes—belong to an independent, southern radiation of tribosphenic mammals, which they have called Australosphenida.
Ausktribosphenos is based on a lower jaw from the Early Cretaceous of Australia that shows a precociously hedgehoglike molar pattern (Rich et al., 1997), leading its describers
to suggest that it could be one of the oldest known placen-tals. Others contend that Ausktribosphenos is not a placental because of the position of the mandibular foramen (in contact with the Meckelian groove), presence of a posteromedial depression presumably for postdentary bones, and different structure of the angular process than in therians—features suggesting that Ausktribosphenos might have been derived independently from symmetrodonts (Kielan-Jaworowska et al., 1998). The shape of the dentary and the last premolar are symmetrodont-like, but the molars are rather different from those of symmetrodonts. Subsequent discovery of a second ausktribosphenid jaw showed that an angular process is, in fact, present at the back of the dentary, as in therians (Rich et al., 1999). These authors continue to argue that, based on that feature and the presence of five premolars and three molars, including a submolariform last premolar and molars with a talonid basin and low trigonid, Ausktribosphenos was, after all, a primitive placental (Rich et al., 1999, 2002; Woodburne et al., 2003). One of the problems with this interpretation is that Ausktribosphenos does not resemble known Cretaceous eutherians; rather, it bears a (probably superficial) resemblance to middle and later Tertiary erina-ceids. Recently, Rich, Flannery et al. (2001) named another ausktribosphenid, Bishops, from the late Early Cretaceous of Australia. They classified it, too, as a placental, even though it has six premolars, in contrast to any known placental. The affinities of ausktribosphenids are puzzling, but few paleontologists have accepted placental ties. A relationship to monotremes seems more likely than to placentals (e.g., Sigogneau-Russell et al., 2001; Luo et al., 2002).
Even more unexpected was the discovery of a jaw with three tribosphenic molars from the Middle Jurassic (about 167 Ma) of Madagascar (Flynn et al., 1999). Named Ambon-dro mahabo, it is 25 million years older than Tribotherium, the oldest tribosphenic mammal known from the African continent. It is primitive in having an open molar trigonid, but already has a fully developed talonid basin.
The first South American australosphenidan was recently described from late Middle Jurassic strata of Argentina (Rauhut et al., 2002; Martin and Rauhut, 2005). Asfaltomylos is based on a mandibular fragment with several teeth, including three tribosphenic molars. The dentary is primitive in having an anteriorly placed mandibular foramen at the front of a postdentary trough, implying the presence of small postdentary bones. Asfaltomylos appears to be closely allied with, but slightly more primitive than, Ambondro and is therefore the most basal australosphenidan. The discovery of this South American form reinforces the notion of a widespread Mesozoic radiation of australosphenidans.
Steropodon, based on a jaw containing three molar teeth from the late Early Cretaceous Lightning Ridge local fauna of Australia, has been interpreted as a monotreme (Archer et al., 1985). Its molars have a mesiodistally compressed trigonid separated by a deep notch from a short lophlike talonid. The configuration is reminiscent of the tribosphenic cusp pattern of marsupials and placentals, but differs in the mode of wear, which suggests that Steropodon may not have had a protocone on the upper molars (which are unknown). Based on this resemblance, Steropodon and monotremes were considered aberrant therian mammals, representing a lineage separate from that leading to marsupials + placentals (Kielan-Jaworowska et al., 1987a).
Upon restudy of these fossils, however, Luo, Cifelli, and Kielan-Jaworowska (2001) found that Steropodon shares derived dental features (an anterolingual cingulid, a low trigo-nid, and a twinned paraconid and metaconid) with Ausktribosphenos and Ambondro. On this basis they postulated that these taxa, together with other monotremes, represent an endemic Gondwanan radiation of mammals, Australo-sphenida, that evolved tribosphenic molars independently from the Holarctic Boreosphenida. According to this hypothesis, the tribosphenic molar arose at least twice and apparently evolved earlier in the Southern Hemisphere than in Laurasia. It also suggests that monotremes are the sole survivors of the australosphenidan radiation, whereas marsupials and placentals represent the boreosphenidans.
This interpretation has been questioned by some authors. Sigogneau-Russell et al. (2001) suggested that the presence of an anterolingual cingulid is not restricted to australo-sphenidans, but rather is a widespread primitive feature in early mammals. In their view, Ambondro could be as closely related to boreosphenidans as it is to australosphenidans. They therefore postulated that the tribosphenic molar had a Gondwanan origin, spreading from there to other parts of Gondwana as well as to Laurasia (and diversifying differently in these two regions). Finally, Woodburne et al. (2003) considered Ambondro, Asfaltomylos, and Bishops, as well as Ausktribosphenos, to be eutherians, and suggested a single, presumably southern, origin of tribosphenic mammals. Additional fossils obviously will be important in testing these novel hypotheses.
The living platypus (Ornithorhynchidae) and echidnas (Tachyglossidae) of Australia and New Guinea are all that remain of the Monotremata, a group that diverged from other mammals during the Mesozoic. Because they are en-dothermic, have hair, suckle their young, and have several other mammalian synapomorphies, they are undisputed mammals. Nonetheless, their primitive nature and long separation from other living mammals are underscored by their unique characteristic of laying eggs. Monotremes are more primitive than other extant mammals (and resemble morganucodonts) in possessing cervical ribs, a therapsid-like shoulder girdle with an interclavicle and both coracoids, and a sprawling forelimb posture. The scapula lacks a spine and a supraspinous fossa. In addition, the skull retains a septomaxilla, as in cynodonts and morganucodonts, and the cochlea is only bent (not coiled as in marsupials and pla-centals) and lacks the bony laminae within the cochlear canal found in marsupials and placentals. The pelvic girdle is more therian-like than the shoulder girdle, and there are large epipubic bones in both sexes.
Superimposed on this basically primitive structure are several specialized (autapomorphous) features. The skull is usually described as birdlike, the rostrum being either narrow and beaklike (in Tachyglossidae) or shaped like a duck's bill (in Ornithorhynchidae). Adult monotremes are edentulous, but juvenile platypus retains milk "molars." The jugal bones are small or absent, and the zygomae are formed from processes of the maxilla and squamous temporal.
The primitive nature of monotremes indicates that they must have diverged from the mammalian stem quite early in the history of mammals, yet the fossil record of monotremes is extremely poor and is almost entirely limited to Australia. Possible derivation from peramurid or dryolestoid eupantotheres has been suggested, but no transitional fossils have been found. As discussed above, it is possible that monotremes instead are part of a Gondwanan tribosphenic radiation; but we are just getting our first glimpses of this supposed australosphenidan clade, and its origin remains obscure. Recently Woodburne (2003) advanced the hypothesis that monotremes are relicts of a Mesozoic, mainly Gondwanan radiation of pretribosphenic mammals, unrelated to australosphenidans.
Pleistocene remains of both living families are known, as is a large Miocene echidna (Zaglossus), but they are similar to living forms and contribute little to understanding either the relationships or evolution of Monotremata. A Miocene ornithorhynchid, Obdurodon, was initially described from isolated teeth that resemble the transient teeth of juvenile platypus (Woodburne and Tedford, 1975). Both Obdurodon and juvenile Ornithorhynchus have vaguely bilophodont "molar" teeth comprised of two parts separated by a transverse valley. Each part consists of a high internal cusp joined by transverse crests to one or two labial cusps. Subsequent discoveries of Obdurodon, including a complete skull, confirm that it is a fossil platypus (Archer, Murray, et al., 1993). The realization that these Miocene teeth belonged to fossil ornithorhynchids proved critical to identification of still older monotremes—all of which are known only from teeth or jaws.
Since the mid-1980s several intriguing, much older specimens thought to be monotremes have come to light. Three of them are based on lower jaw fragments from the late Early Cretaceous (Aptian) of Australia. They have been placed in two different families. Steropodon, whose molar structure resembles that of Obdurodon, was discussed in the preceding section. A second Steropodon-like form, Teinolophos, was recently reported from the late Early Cretaceous Flat Rocks locality in Australia (Rich et al., 1999; Rich, Vickers-Rich, et al., 2001). It was initially interpreted as a eupantothere but is now considered to be a monotreme. Significantly, Teino-lophos retains a trough on the medial surface of the mandible, a primitive feature reflecting the presence of accessory (postdentary) bones. The retention of postdentary bones, in turn, implies that the middle-ear ossicles were not yet separate from the lower jaw and, therefore, that the three-ossicle chain evolved independently at least twice in mammals (Rich et al., 2005; Martin and Luo, 2005).
Kollikodon, from Lightning Ridge, Australia (the same locality as Steropodon), was initially based on a lower jaw fragment with three teeth. The name, meaning "bun tooth," alludes to its peculiar, quadrate lower molars with four rounded and inflated cusps (perhaps adapted to feeding on crustaceans)—which reminded its describers (Flannery et al., 1995) of hot cross buns! The upper teeth are also multi-cusped and very bunodont (Kielan-Jaworowska et al., 2004). This dental morphology is strikingly different from that of other monotremes, and its monotreme affinity remains to be demonstrated. If it is indeed a monotreme, it indicates that two quite divergent lineages already existed in the Early Cretaceous.
The only known Early Tertiary monotreme, Monotrema-tum (Fig. 4.18), comes from the early Paleocene ("Peligran") of southern Argentina (Pascual et al., 1992, 2002). It is also the only monotreme known from outside the Australian region. Although only two teeth have been found, it can be securely identified as a platypus based on its close resemblance to Obdurodon. Presence of a monotreme in South America indicates that the present-day restriction of the group to Australia is but a remnant of a once much wider geographic distribution.
Northern Tribosphenic Mammals: Boreosphenida (Metatherians, Eutherians, and Related Therians)
From the late Early Cretaceous come two closely allied (perhaps synonymous) genera that have been widely considered to lie close to the stem of Holarctic tribosphenic mammals, or Boreosphenida (essentially =Tribosphenida of McKenna and Bell, 1997, or Theria + Aegialodontia of general usage), or at least to represent a critical stage in the origin of tribosphenic molars (Crompton, 1971). Included here are Aegialodon (Lower Cretaceous Wealden beds of England) and Kielantherium (late Early or early Late Cretaceous from Khovboor, Mongolia). Kielantherium (Fig. 4.16C) is advanced beyond such eupantotheres as Arguimus (Fig. 4.16D) in having a talonid basin and an entocristid. Recently Sigogneau-Russell et al. (2001) described a third, even older aegialodont genus, Tribactonodon, from the earliest Cretaceous (Berriasian) Purbeck Group of England. Tribactonodon is based on a lower molar with a tall trigonid and a long, narrow talonid with three cusps.
Although aegialodont upper teeth are unknown, the structure and wear pattern of their lower molars indicate that a protocone was present on the upper molars, making aegialodonts the most primitive known truly tribosphenic mammals from Laurasia. The lower jaw held four or possibly five premolars and four molars (Dashzeveg and Kielan-Jaworowska, 1984). Disagreements persist over the correct homologies of the postcanines and, therefore, the dental formula of Aegialodontia and Peramura. Kielan-Jaworowska (1992) argued that Aegialodontia was ancestral to Metatheria but not Eutheria, but subsequently aegialodonts have been depicted as the sister taxon of Theria (Metatheria + Eutheria; e.g., Luo et al., 2002).
Tribotheria was proposed to encompass primitive tri-bosphenic forms whose teeth lack diagnostic traits of Metatheria or Eutheria (Butler, 1978; Sigogneau-Russell, 1991). Consequently "Tribotheria" is an evolutionary grade rather than a clade. Nonetheless, the term "tribothere" remains a useful informal name for these fossils—most known only from isolated teeth or fragmentary jaws—which are often difficult to assign and are otherwise given the cumbersome moniker "therians of metatherian-eutherian grade." Most are probably basal members of Boreosphenida; however, it may eventually be possible to assign many of them to either Metatheria or Eutheria when they are more completely known. A variety of "tribotheres" have been described from the Cretaceous of North America, including Comanchea, Picopsis, Potamotelses, Kermackia, Pappotherium, and Holo-clemensia (Fig. 4.19), the last two originally considered to be basal eutherian and metatherian, respectively. Substantially older tribotheres are now known from Africa, and their antiquity has raised significant questions regarding the time of origin of tribosphenic molars. Tribotherium, based on an upper molar from the earliest Cretaceous of Morocco, pushed back the first occurrence of mammals with tribosphenic molars by 30 million years (Sigogneau-Russell, 1991).
Although the three groups of living mammals (mono-tremes, marsupials, and placentals) are differentiated primarily by reproductive strategy, it has long been held that differences in dental pattern and morphology also distinguish primitive marsupials and placentals and can be used to assign fossils to one group or the other. With the discovery of an increasing number and variety of Cretaceous mammalian remains, however, it has become clear that the "defining" dental characters were acquired in mosaic fashion. Consequently, it is often not possible to identify a fossil taxon as definitively metatherian or eutherian based on teeth alone, and it is even possible that some fossils represent tribosphenic clades that diverged before the metatherian-eutherian dichotomy. Once the dichotomy occurred, probably during or before the Early Cretaceous, nearly all higher therians seem to lie closer to either Metatheria or Eutheria; but this inference has often been based on characters that are not demonstrably derived and may well be primitive. Experts still disagree on whether certain early forms represent true marsupials or other metatherian clades and on how to recognize the oldest true placentals.
Was this article helpful?