Saturday, 20 June 2009

Diprotodont Evolution

Marsupials are one of the three major lineages of extant mammals: they are known for their extremely short gestation periods, so short that the young are born undeveloped and have to be reared by the mother in a pouch (in most forms). Marsupials do not make up a single order of mammals; however, in fact, there are seven extant orders and several more extinct ones. Those orders are: Didelphimorphia (opossums), Paucituberculata (shrew opossums, also known as rat opossums, flap-lips or caenolestids), Microbiotheria (the colocolo or monito del monte), Dasyuromorphia (marsupial carnivores, including the dunnarts, quolls, Tasmanian devil, thylacine and numbat), Notorycterimorphia (the marsupial moles), Peramelemorphia (bandicoots and bilbies), and the Diprotodontia.

Long-nosed potoroo
Potorous tridactylus (Kerr, 1792)
Family Potoroidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

Diprotodontia are the largest extant marsupial order, and contain a wide range of well-known animals, from the tiny sugar glider and honey possum to the giant red and grey kangaroos, to the unusual wombats and koalas. The study by Meredith et al. (2009) in June’s edition of Molecular Phylogeny and Evolution is the first to incorporate every extant genus in the order, uncovering relationships which had been hitherto unknown.

Honey possum
Tarsipes rostratus Gervais & Verreaux, 1842
Family Tarsipedidae; clade Petauroidea; clade Phalangerida; order Diprotodontia

Diprotodonts (meaning ‘two front teeth’, after the procumbent, or sticking out, lower medial incisors) are generally herbivorous, but there are some specialised nectarivores (i.e. the honey possum), folivores (leaf-eaters, like the koala) and insect-omnivores (such as rat kangaroos). They inhabit a wide range of ecological niches, from arboreal koalas to bipedal kangaroos, and from semi-fossorial wombats to gliding petaurids. The procumbent incisors are a synapomorphy (unique derived feature) of the order, but there are others, mostly to do with internal anatomy. A feature shared with the Peramelemorphia, the bandicoots and bilbies, is syndactylous hind feet, where the second and third digits are fused, appearing to be one digit with two claws. It is still unclear whether this commonality makes the Peramelemorphia the sister group to Diprotodontia.

Common brushtail possum
Trichosurus vulpecula (Kerr, 1792)
Family Phalangeridae; clade Phalangeroidea; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

Meredith et al. carried out a molecular analysis on every extant genus of diprotodont, and members of every other extant marsupial family, and also carried out a molecular clock analysis. This estimates the dates of divergence of each clade so it can be worked out when certain lineages first arose and therefore this can be linked with climatological evidence.

Mountain pygmy possum
Burramys parvus Broom, 1896
Family Burramyidae; clade Phalangeroidea; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

Among the polyprotodont marsupials (the name given to those marsupials not belonging to the Diprotodontia, having four or more upper incisors), one true monophyletic clade (a grouping with a common ancestor) was found, which also includes the Diprotodontia: the Australidelphia. The Australidelphia, as the name suggests, originated in Australia, most likely whilst it was still connected to South America (where the other marsupials are still found) via Antarctica. All but one of the australidelphians are found in Australasia today; that one is the monito del monte (Dromiciops gliroides) from Chile, of its own order, Microbiotheria. This poses a biogeographical problem: did the Microbiotheria disperse back, or did the australidelphians disperse multiple times? The dilemma is as yet unresolved. What we do know, however, that at least by the Eocene, marsupials were present in Australia, as proven by the earliest australidelphian, Djarthia murgonensis Beck et al., 2008.

The other two orders from the Americas, Didelphimorphia and Paucituberculata, were thought to make up a clade, Ameridelphia (with their origins in America, obviously), but the study found that group to be paraphyletic (not sharing a unique common ancestor). Classically, the two clades were described on the basis of differences in the ankle joint.

Bear cuscus
Ailurops ursinus (Temminck, 1824)
Family Phalangeridae; clade Phalangeroidea; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The diprotodonts are basally split between the Vombatiformes (wombats and koalas) and Phalangerida (all the rest; don’t confuse with Phalangeridae). The previously established suborder Phalangeriformes, thought to include all possums, is proven paraphyletic, as the Macropodiformes (kangaroos and their kin) are nested neatly in between the two groups containing ‘possums’. Supposed apomorphies (derived characters) shared by all members of the ‘Phalangeriformes’ are either convergent, or were secondarily lost by the Macropodiformes.

Musky rat kangaroo
Hypsiprymnodon moschatus Ramsay, 1876
Family Hypsiprymnodontidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The musky rat kangaroo (Hypsiprymnodon moschatus) is definitely in a family of its own, from the inferred date of its divergence (28 million years ago) and several other factors, but it was previously thought to be the sister taxon to the Potoroidae (potoroos, bettongs and rat kangaroos). It is instead the sister to both Potoroidae + Macropodidae. This suggests that bipedal hopping and reduction of litter size evolved only once, at the root of the Macropodiformes.

Banded hare wallaby
Lagostrophus fasciatus (Peron & Lesueur, 1807)
Subfamily Sthenurinae; family Macropodidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The banded rat kangaroo (Lagostrophus fasciatus) has usually been placed close to the rat kangaroos of the genus Lagorchestes. They fall on different sides of the Macropodidae to each other, to the extent of strengthening the hypothesis that Lagostrophus is the only surviving member of the otherwise extinct subfamily Sthenurinae (containing the infamous giant kangaroos of the Pleistocene), whilst Lagorchestes is close to the larger kangaroos and wallabies of the genus Macropus. Lagostrophus diverged c. 15 million years ago in the mid-Miocene. Of course, being supposedly the only member of an extinct taxon, it is impossible to carry out more molecular studies to confirm the relationship, unless late Pleistocene proteins or DNA are found.

Common spotted cuscus
Spilocuscus maculatus (E. Geoffroy, 1803)
Family Phalangeridae; clade Phalangeroidea; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The name for a new clade has been given, Australoplagiaulacoida (what a mouthful!) to the grouping of Macropodiformes and Phalangeroidea. The derived feature is serrated premolars in the (supposed) primitive members of each family in the clade.

Red kangaroo
Macropus rufus Desmarest, 1882
Family Macropodidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The large and familiar genus Macropus, containing the largest living diprotodonts, has recently been split into several genera. The analysis defends neither corner, but the authors consider the genus to consist of subgenera. The swamp wallaby (previously Wallabia bicolor) is nested between the agile and western brush wallabies (Macropus agilis & M. irma) and the rest of Macropus, thus proving it is actually in that genus, but as a distinctive subgenus.

Striped possum
Dactylopsila trivirgata Gray, 1858
Family Petauridae; clade Petauroidea; clade Phalangerida; order Diprotodontia

Within the Petauridae, there is a weakly supported (by bootstrap methods – too complex to explain here, and even I don’t understand the methodology too well) grouping of the Leadbeater’s possum (Gymnobelideus leadbeateri) with the striped possum and trioks (Dactylopsila spp.). One would think, based on appearances alone, that the Leadbeater’s possum, which looks like a ‘flightless’ sugar glider, would pair with the genus Petaurus. Because the clade is weakly supported, this may be taken with a pinch of salt.

Common ringtail possum
Pseudocheirus peregrinus (Boddaert, 1785)
Family Pseudocheiridae; clade Petauroidea; clade Phalangerida; order Diprotodontia

The Pseudocheiridae (ringtail possums and greater glider) have traditionally been grouped into three lineages: Hemibelideus + Petauroides; Pseudocheirus + Pseudochirulus; and Pseudochirops + Petropseudes. The monophyly of these groups have been confirmed, but Pseudochirops has been found to be paraphyletic. The rock ringtail possum (formerly Petropseudes dahli), despite being ecomorphogically disparate to the arboreal members of Pseudochirops, was found to be nested within that genus.

Green ringtail possum
Pseudochirops archeri (Collett, 1884)
Family Pseudocheiridae; clade Petauroidea; clade Phalangerida; order Diprotodontia

The genus Strigocuscus, formed of two species of phalanger from eastern Indonesia, has been proven paraphyletic, with S. pelengensis finding itself closer to the genus Phalanger, so the species should probably move to that genus. Strigocuscus celebensis can stay where it is, for now.

Eastern common cuscus
Phalanger intercastellanus Thomas, 1895
Family Phalangeridae; clade Phalangeroidea; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The so-called molecular clock revealed the approximate times of divergence of every marsupial clade. Some interesting dates from the Cretaceous and Palaeogene include Vombatiformes diverging from the Phalangerida approximately 53 million years ago in the early Eocene; Australidelphia separating from the rest of the marsupials 63 million years ago in the early Palaeocene; and (extant) marsupials as a whole diverging from all other mammalian clades about 77 million years ago, predating the Cretaceous-Tertiary mass extinction. Most of the families as we know them diverged before the end of the Oligocene, with the Potoroidae and Macropodidae diverging after 23 million years ago.

Bennett’s tree kangaroo
Dendrolagus bennettianus De Vis, 1887
Family Macropodidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

Meredith et al. also looked at “ecological venue”, meaning the habitat and niche that the ancestral diprotodont and the ancestral marsupial would have occupied, linking this to the information derived from the molecular clock. It appears that the common ancestor of all diprotodonts was arboreal, using maximum likelihood. Going further back, however, it becomes more difficult to assign an ecological venue for the ancestor of all marsupials: the maximum likelihood seems to be arboreal, but could also be terrestrial, fossorial (digging), or semi-fossorial. The tree kangaroos (Dendrolagus spp.) are secondarily arboreal, as all of their close kin are terrestrial, but their distant ancestors lived in trees. The original transition from arboreal to terrestrial macropodiforms occurred between 45 and 27 million years ago, with tree kangaroos reverting to the ancestral condition much more recently, in the Pleistocene.

Phascolarctos cinereus (Goldfuss, 1817)
Family Phascolarctidae; clade Vombatiformes; order Diprotodontia

If we look at the Vombatiformes, consisting of three semi-fossorial members and one arboreal one, it is tempting to say that they started out terrestrial and only the koala took to the trees. It is more likely that the reverse is true: the direct ancestors of wombats lived in trees. The koalas diverged before the wombats, but both families kept similar adaptations for arboreality, such as the distinctive hands and fingers that unite the wombats and the koalas (not to mention the extinct giant wombats such as Diprotodon, who gave its name to the order). It is difficult to say when exactly the wombats became semi-fossorial, as it obviously happened after the divergence of the koalas, but before the different genera of wombat diverged from each other. The earliest vombatid, Rhizophascolomus crowcrofti, of the late Oligocene/early Miocene, is known only from cheek teeth; postcranial information, which is crucial to ascertain its ecological venue, is lacking.

Southern hairy-nosed wombat
Lasiorhinus latifrons (Owen, 1845)
Family Vombatidae; clade Vombatiformes; order Diprotodontia

In the Palaeocene, Australia was dominated by warm, wet rainforests. Gymnosperms (conifers, cycads and their relatives) were still dominant, with the new-fangled angiosperms (flowering plants) becoming more widespread by the start of the Eocene. Temperatures then decreased and the podocarp (coniferous gymnosperm) forests were replaced by more diverse southern beech (Nothofagus spp.) forests. The lack of divergence of marsupial groups prior to this time is put down to lack of available niches in such homogeneous podocarp forests. Marsupials were arboreal then, as we have seen, because little habitat was available at ground level. Imagine a tropical-forest dwelling wombat trying to dig its way through mulchy or pine-needle covered strata. Recall that the earliest Australian marsupial dates to the Eocene, with no marsupials recorded from the Palaeocene; this probably has much to do with this lack of niches. Djarthia possesses traits which liken it to arboreal animals, matching the molecular clock and palaeobotanical data.

Proserpine rock wallaby
Petrogale persephone Maynes, 1982
Family Macropodidae; clade Macropodiformes; clade Australoplagiaulacoida; clade Phalangerida; order Diprotodontia

The lineages leading to the three major australidelphian orders, Peramelemorphia, Dasyuromorphia and Diprotodontia, were established prior to 55 million years ago, but no extant families had yet been established. The early Oligocene brought a significant decrease in rainfall in Australia, mainly caused by the formation of an ice-cap in Antarctica. The Australian plate collided with the Asian plate, causing the island of New Guinea to rise, complete with mountains. The mountains created a rain shadow, resulting in a dry continent. By the end of the Oligocene, new habitats consisting of sclerophyll (Mediterranean-like) vegetation, with scattered sedge and reed swamps, and by this time every terrestrial Australian marsupial family had been established. The rainforest canopy had thus opened, creating far more terrestrial niches. By the time of the Pliocene, grasslands were prevalent across much of Australia, with the current climatic conditions reached by the Pleistocene. The divergence of the Macropus genus coincided well with the spread of grassland in this epoch.

Greater glider
Petauroides volans (Kerr, 1792)
Family Pseudocheiridae; clade Petauroidea; clade Phalangerida; order Diprotodontia

The study also looked at the evolution of gliding membranes. Gliding has arisen in three eutherian taxa: Petauristinae (flying squirrels); Anomaluridae (scaly-tailed squirrels); and Dermoptera (colugos). The authors confirmed the suspicion that gliding evolved, again, three times independently in the marsupials: once in the Acrobatidae; once in the Petauridae; and once in the Pseudocheiridae. Within those groups, not all members possess the gliding membranes. The feathertail possum (Distoechurus pennatus) does not possess the membrane that the feathertail glider does; only the members of the genus Petaurus can glide, unlike the genera Dactylopsila and Gymnobelideus; and only one species out of seventeen of Pseudocheiridae, the greater glider (Petauroides volans) can glide.

Feathertail glider
Acrobates pygmaeus (Shaw, 1793)
Family Acrobatidae; clade Petauroidea; clade Phalangerida; order Diprotodontia

The specific anatomy of the membrane differs between the three families, showing that they are not completely homologous to each other: the patagium of Petaurus stretches from the hands to the ankles; in the greater glider it stretches from the elbow to the ankle; in the feathertail glider it stretches from the elbow to below the knee joint. The latter species’ tail, which is fringed with stiffened hairs, also assists in gliding. This feature is shared with the other member of the family, the feathertail possum, which is incapable of flight. The greater glider’s sister taxon, the lemurine ringtail possum (Hemibelideus lemuroides) has flaps of skin where the membrane should be. These flaps could either be incipient membranes (i.e. haven’t formed), or formerly present but now are vestigial.

Sugar glider
Petaurus breviceps Waterhouse, 1839
Family Petauridae; clade Petauroidea; clade Phalangerida; order Diprotodontia

Gliding on the whole amongst diprotodonts coincided with the drying out of Australia’s forests at the end of the Oligocene. This transition of habitat may have been the impetus for various groups of petauroid marsupial to evolve patagia.

I leave you with a diagram, modified from the Bayesian tree (Fig. 1) from Meredith et al.’s paper, featuring some of the diprotodonts mentioned in the study and this review. The four main groups, top to bottom being Petauroidea, Macropodiformes, Phalangeroidea and Vombatiformes, are quite visible.

Beck, R. M. D., H. Godthelp, V. Weisbecker, M. Archer & S. J. Hand (2008) Australia's Oldest Marsupial Fossils and their Biogeographical Implications. In: PLoS One 3(3): e1858. doi: 10. 1371/ journal. pone. 0001858

Meredith, R. W., M. Westerman & M. S. Springer (2009) A phylogeny of Diprotodontia (Marsupialia) based on sequences for five nuclear genes. In: Molecular Phylogenetics and Evolution 51:554-71. doi: 10. 1016/ j. ympev. 2009. 02. 009
All illustrations:
Graphite pencil
By Mo Hassan
June 2009


Allen Hazen said...

Thanks for that really nice account of the Diprotodonta!
On a tangential topic... the estimate of the number of independent developments of gliding in mammals needs to be increased. There's apparently a second case among rodents: some years back a fossil was described with features suggesting gliding (I think it may even have been one of those extraordinarily preserved specimens that showed traces of the patagium), but5 related to dormice and not to flying squirrels. Then of course there was Volaticotherium. And PROBABLY the ancestor of bats went through a gliding phase.

And given that Ptilodus has been reconstructed with arboreal adacptations, if asked to bet I'd put a little money on there having been at least one gliding Multituberculate, but so far no actual evidence on that one!

Fuzhong! said...

very interesting post