The marsupial moles are rare and poorly understood burrowing mammals of the deserts of Western Australia. There are thought to be two species: the Southern Marsupial Mole (Notoryctes typhlops and the Northern Marsupial Mole (Notoryctes caurinus), so similar to one another that they cannot be reliably told apart in the field.
Marsupial moles spend most of their time underground, coming to the surface only occasionally, probably mostly after rains. They are blind, their eyes having become reduced to vestigal lenses under the skin, and they have no external ears, just a pair of tiny holes hidden under thick hair.
The head is cone-shaped with a leathery shield over the muzzle; the body tubular; the tail a short, bald stub. They are between 12 and 16 cm long, weigh 40 to 60 grams, and are uniformly covered in fairly short, very fine pale cream to white hair with an iridescent golden sheen. The pouch faces to the rear.
Marsupial moles provide a remarkable example of convergent evolution, with moles generally, and with the golden moles of Africa in particular. Although only related to other moles in that all are mammals, the external similarity is an extraordinary reflection of the similar evolutionary paths they have followed.
For many years the place of the marsupial moles within the marsupials was hotly debated, some workers regarding them as an offshoot of the Diprotodontia (the order to which most living marsupials belong), others noting similarities to a variety of other creatures, and making suggestions that in hindsight appear bizarre. A 1989 review of the early literature, slightly paraphrased, states:
- When Stirling (1888) initially was unable to find the epipubic bones in marsupial moles, speculation was rife: the marsupial mole was a monotreme, it was the link between monotremes and marsupials, it had it closest affinities with the (placental) golden moles, it was convergent with edentates, it was a polyprotodont diprotodont, and so on. 
The mystery was not helped by the complete silence of the fossil record. On the basis that marsupial moles have some characteristics in common with almost all other marsupials, they were eventually classified as an entirely separate order: the Notoryctemorphia. Molecular level analysis in the early 1980s showed that the marsupial moles are not closely related to any of the living marsupials, and that they appear to have followed a separate line of development for a very long time, at least 50 million years.
In 1985, the vast newly discovered limestone fossil deposits at Riverseigh in Queensland's Gulf Country yielded a major surprise: marsupial mole fossils between 15 and 20 million years old, which were by no means identical to the living species but clearly related, and possibly even of a direct ancestor.
In itself, the discovery of a Miocene marsupial mole presented no great mysteries. Just like the modern forms, it had many of the features that are assumed to be adaptations for a life burrowing in desert sands, in particular the powerful, spadelike forelimbs. The Riversleigh fossil deposits, however, are from an environment that was not remotely desert-like: in the Miocene, the Riversleigh area was a tropical rainforest.
One suggestion advanced was that the Miocene marsupial mole used its limbs for swimming rather than burrowing, but the mainstream view is that it probably specialised in burrowing through a thick layer of moss, roots, and fallen leaf litter on the rainforest floor, and thus, when the continent began its long, slow desertification, the marsupial moles were already equipped with the basic tools that they now use to burrow in the sand dunes of the Western Australian desert.
- Carina Dennis, A mole in hand
- Archer, Hand & Godthelp, Australia's lost world: Riversleigh, World Heritage Site, Reed New Holland, 1991.