* AUSTRALOSPHENID A
Fig. 4.2. Relationships of Mesozoic mammals. The current consensus places haramiyids (including Haramiyavia) and Multituberculata as sister taxa within Mammalia. (Modified after Luo et al., 2002.)
Until these are found, Adelobasileus will remain a taxon of problematic relationships.
Two other families that appeared in the Late Triassic (?Norian-Rhaetic), Haramiyidae and Theroteinidae, could be the oldest known mammals, but are also problematic. Both have been suggested to have possible affinities with multi-tuberculates because of dental resemblance. Theroteini-dae are known only from isolated teeth from France that have complex crowns and preprismatic enamel (Sigogneau-Russell et al., 1986; Hahn et al., 1989). They could belong to primitive mammals or to advanced cynodonts, but similarities to teeth of haramiyids and primitive multituberculates suggest that they are probably mammals. The Middle Jurassic eleutherodontids (Eleutherodon), based on isolated teeth from England and China (Kermack et al., 1998; Butler and Hooker, 2005; Maisch et al., 2005), represent a third family perhaps related to theroteinids or haramiyids; but it is difficult to reach definitive conclusions based on these isolated teeth. Citing similarities in their molariform teeth, Butler (2000) united these three families in the order Haramiyida. If multituberculates originated from within this group, Hara-miyida would be a paraphyletic assemblage. The grouping of Haramiyida and Multituberculata is called Allotheria.
Until recently, haramiyids were also known only from isolated teeth. They resemble multituberculate molars in being relatively low crowned and having two parallel, peripheral rows of cusps, longitudinally arranged and separated by a median furrow. These teeth were presumed to be mo lars, but their orientation and position in the toothrow were uncertain. Their mammalian status did not appear to be in question, however, for they have multiple roots, pre-prismatic enamel, and wear facets indicating precise interlocking occlusion, which in turn suggests diphyodont tooth replacement (Sigogneau-Russell, 1989). Wear patterns even seem to indicate a palinal (longitudinal and backward) chewing stroke, as in multituberculates (Butler and MacIntyre, 1994).
Much more complete haramiyid fossils discovered in Greenland (Jenkins et al., 1997) confirm that the isolated teeth were indeed molars and clarify their orientation. These fossils, named Haramiyavia (Fig. 4.3), also show that the postcranial skeleton was generally similar to that of morganucodonts, the most primitive mammals for which the skeleton is known. The lower dentition further resembles that of multituberculates in having procumbent incisors separated from the cheek teeth by a conspicuous diastema. According to Jenkins et al. (1997), the occlusal relationships of the teeth contradict those observed in other haramiyids and indicate a predominantly orthal (vertical) chewing stroke. This observation might suggest that Haramiyavia, at least, was not as closely related to multituberculates as had been supposed. However, Butler (2000) suggested that limited palinal movement probably did occur during the power stroke of chewing in Haramiyavia, which would support its position as a transitional form leading to multituberculates. Butler (2000) placed Haramiyavia in its own family, Haramiyaviidae.
Haramiyavia retained larger postdentary bones than did morganucodontids, suggesting that haramiyids could be an
D A3 B5
D A3 B5
even earlier offshoot of the mammalian stem. Unfortunately the jaw joint is not preserved in the fossils of Haramiyavia, but the presence of larger postdentary bones suggests that a quadrate-articular joint was still functional. Potentially more significant is the evidence for orthal and palinal jaw movements during chewing in Haramiyida and multituber-culates. This mode of occlusion, which differs fundamentally from that in other mammals, led Butler (2000; see also Butler and Hooker, 2005) to hypothesize that allotheres diverged from other mammals before they evolved unilateral shearing and transverse jaw movements—which would be very early indeed. Primitive haramiyidans document transitional stages in the development of palinal occlusion. Whether haramiyidans are related to multituberculates or represent a separate branch of primitive mammals, or even cynodonts, remains unsettled (e.g., Butler and MacIntyre, 1994; Butler, 2000; Luo et al., 2002).
Morganucodonts, Slnoconodon, and Kuehneotheriids
Prior to these discoveries, the oldest mammals were long held to be those from fissure-fillings in Wales that have usually been considered of Late Triassic (Rhaetic) age.
However, their uncertain age was indicated by the label "Rhaeto-Liassic" often applied to these fossils, and the age of the fissures now appears to be younger rather than older (late Rhaetic-Liassic, or Sinemurian: latest Triassic-Early Jurassic; Kermack et al., 1981; Clemens, 1986; Kielan-Jaworowska, 1992). They are usually assigned to the families Morganucodontidae and Kuehneotheriidae, which have essentially triconodont-like cheek teeth. Morganucodontidae (formerly considered to be basal "triconodonts") are known from teeth and skulls, whereas Kuehneotheriidae (formerly basal "symmetrodonts") are represented only by jaws and teeth. Chinese deposits of similar age (Liassic) have produced additional skulls of Morganucodon and of another primitive form, Sinoconodon (Fig. 4.4A). These taxa, though perhaps not the oldest, are widely considered to be the most primitive mammals (Crompton and Luo, 1993). Eozostrodon, based on two isolated teeth, has often been considered a synonym of Morganucodon, but it is possibly a distinct morganucodon-tid (Kielan-Jaworowska et al., 2004).
Morganucodonts (now considered to include the families Morganucodontidae and Megazostrodontidae) were small shrew- to mouse-sized animals (10-30 g; Jenkins and Crompton, 1979) that were widely distributed during the latest Triassic(?)-Early Jurassic, occurring in Europe, Asia, Africa, and North America (Fig. 4.4). The anatomy of morganucodonts indicates that they occupy a central position at the base of the mammalian radiation. Kermack et al. (1973, 1981) and Jenkins and Partington (1976) detailed the anatomy of morganucodonts, which combines derived mammalian traits with primitive cynodont features. The dental formula of Morganucodon varies within and among species: 3-5.1.4-5.3-4/4-5.1.4-5.3-5 (Kielan-Jaworowska et al., 2004). As in eutriconodonts, the premolars are simple, with one main cusp, and the molar cusps are linearly arranged (Fig. 4.5). Based on this morphology, morganucodonts were previously regarded as primitive triconodonts; however, the similarities are now generally considered to be plesio-morphic. Current consensus separates morganucodonts from eutriconodonts and places them at the base of mammals, whereas eutriconodonts are thought to be closer to therian mammals (e.g., Luo et al., 2002).
The dentary of morganucodonts is mammal-like in having a large coronoid process and a well-developed condylar process that articulated with the squamosal. From the lateral side, this appears to be the only jaw joint, but medial to it a functional quadrate-articular jaw joint was also still present. In Morganucodon the angular process is situated well anterior to that of more derived mammals, which led Jenkins et al. (1983) to identify it as a pseudangular process. The skull of Morganucodon lacks prefrontal and postorbital bones, as in other mammals and advanced cynodonts, but primitively retains a septomaxilla. Although first reported to retain tabular bones, as in cynodonts (Kermack et al., 1981), subsequent studies have concluded that tabulars are absent in Morganucodon (e.g., Luo et al., 2002).
The vertebral column of morganucodonts is more regionally differentiated than in cynodonts and shows other
modifications associated with mammal-like movements, including an essentially mammalian atlas-axis complex and an enlarged cervical canal (reflecting expansion of the spinal cord in the region of the brachial plexus, which in turn suggests more complex neural control of the forelimbs). At the same time, the pectoral girdle is distinctly cynodont-like. The scapula lacks a supraspinous fossa, and there are two coracoids, although only the posterior one contributes to the glenoid fossa. The humerus is rather therian-like at the proximal end, with a hemispherical head and a pair of tuberosities; but distally it has an ulnar condyle as in cyn-odonts, rather than an ulnar trochlea as in advanced mammals. The proximal and distal articulations of the humerus are twisted relative to each other. These humeral features suggest a sprawling stance. The pelvis is derived, as in mam mals and tritylodontids, in having a long ilium with separate gluteal and iliac surfaces and a large obturator foramen. The femur has a spherical head and trochanters arranged as in mammals (and tritylodontids, but not other cynodonts). The ankle, however, shows few mammalian specializations except for the presence of an astragalar foramen. Morphology of the phalanges and a probably abducted hallux suggest grasping ability. Together these features suggest that morganucodonts had rather generalized skeletons that enabled them to climb as well as scramble on the ground.
Sinoconodon has triconodont-like cheek teeth, which differ from those of morganucodonts in lacking well-developed cingula (Fig. 4.4A). As in morganucodonts, there is a bony separation between the orbits, and the jaw joint is between the squamosal and the dentary, but the postdentary bones are more reduced than in morganucodonts (Crompton and Sun, 1985)—a presumably derived condition. In other ways, however, Sinoconodon seems to be more primitive than Mor-ganucodon. Cochlear structure was more primitive (Luo et al., 1995), and the absence of consistent wear facets indicates that Sinoconodon lacked precise molar occlusion. It resembles cynodonts in retaining a large septomaxillary bone and multiple replacement of the incisors and canines; in addition, the posterior molars were replaced once (Zhang et al., 1998). These features suggest that Sinoconodon diverged from the mammalian stem earlier than morganucodontids and could be the sister group of all other mammals (Wible, 1991; Crompton and Luo, 1993; Luo et al., 2002).
Kuehneotherium (Fig. 4.5) was formerly regarded as a basal symmetrodont belonging either to the family Tinodon-tidae or to a separate family, Kuehneotheriidae. The current view is that kuehneotheriids occupy a position near the base of mammals (Cifelli, 2001; Luo et al., 2002). The lower dental formula of Kuehneotherium is 4?.1.6.4 or 5 (Gill, 1974). As in morganucodonts and eutriconodonts, the cheek teeth have three principal cusps, with the central one tallest. The main cusp of the upper molars is probably homologous with the paracone of therians, whereas that of the lowers is thought to be homologous with the protoconid. In contrast to morganucodonts and eutriconodonts, however, the cusps are not directly aligned, but form an obtuse angle, with the front and back cusps rotated slightly lingually on the lower teeth and buccally on the uppers, foreshadowing the arrangement in tribosphenic therians. The upper cusps probably represent the stylocone, paracone, and metacone of therian molars, whereas the lowers are probably the three trigonid cusps (followed by an incipient talonid cusp; Patterson, 1956). Whether this cusp rotation is homologous with that in true symmetrodonts and therians or evolved independently is controversial. The relative position of upper and lower cusps during occlusion also differs from that in Morganucodon, being shifted so that each tooth opposed parts of two others. These progressive dental features suggest that Kuehneotherium could be closer to the stem of the therian radiation than any other Rhaeto-Liassic taxon (Cromp-ton and Jenkins, 1979). Luo et al. (2002), however, recently questioned the supposed close relationship between Kueh-neotherium and extant mammals. Kuehneotherium primitively retained much reduced postdentary bones and a double jaw articulation, but the dentary-squamosal joint was predominant. In true symmetrodonts only the dentary-squamosal joint was present.
Woutersia is a possible relative of Kuehneotherium known from isolated teeth from Rhaetian deposits in France. It differs from Kuehneotherium in having cusps on the lingual cingulum (one on upper molars, two on lowers), thus broadening the teeth, a possible early adaptation for crushing (Sigogneau-Russell and Hahn, 1995). Butler (1997) be
lieves that it may represent a transitional form leading to docodonts.
Luo, Crompton, and Sun (2001) described a slightly younger animal with triconodont-like teeth, Hadrocodium, based on a shrew-sized skull from the Sinemurian (Early Jurassic, 195 Ma) of China (Fig. 4.6). It was one of the smallest known mammals, estimated to have weighed only 2 g. Hadrocodium is significant in showing several unexpectedly progressive features for such an ancient mammal. Its skull is wide posteriorly and the braincase is relatively large. There is a single jaw articulation, between the dentary and the squamosal bones; and there is no post-dentary groove, implying that the middle-ear ossicles were already separate from the lower jaw and attached to the skull. These features, together with wear on the molar teeth, suggest that the single known specimen represents an adult or subadult (hence its small size is not attributable to being a juvenile). Hadrocodium had a primitive incisor count of 5/4 but had a reduced number of premolars and molars (dental formula 22.214.171.124/126.96.36.199). The derived features of Hadrocodium indicate that it is more closely related to crown-group Mammalia than is either Sinoconodon or Morganucodon, but the reduced number of cheek teeth make Hadrocodium too specialized to be on the direct line to therian mammals.
Was this article helpful?