Douglass S Rovinsky, PhD

@dsrovinsky.bsky.social

437 Followers 189 Following

Thylacines, sabretooths, marsupial megafauna | Palaeobiology | Morphology, Evolution, Functional Ecology | Exhibitions Content Specialist at Australian Museum (he/him)

Photo of the March 2026 lunar eclipse, with the Moon low on the horizon, giving the shadowed side a bloody, reddish colour. There is a sliver of bright reflected light just kissing the upper left side of the Moon.
Female klipspringer with impeccable eyeliner and the most fashionably black-and-white striped inner ears delivering a most withering amount of sass from her squinty eyes and frowny-frown mouth.
Male klipspringer (Oreotragus oreotragus) staring down my camera lens with what can only be assumed to be a haughty sense of disdain. He also has twin tiny little pokey horns sticking up like two miniature wizard hats. Maybe that's his problem.
The Sun just visible through smoke and haze, looking like the red-orange eye of an angry god.
The full moon from Wyperfeld National Park, Australia. The brilliant Tycho crater is at top right, with its rays of ejecta starred across the Lunar face. Copernicus, Kepler, and Aristarchus craters are centre-line in the bottom half of the Lunar face. Maria Fecunditatis, Crisium, Tranquillitatis, and Serenitatis form a top-bottom chain on the upper left, with Mare Imbrium and Oceanus Procellarum filling much of the bottom of the Lunar face.
Mike McCready’s 1960 Stratocaster on the left, showing over a half-century of idiosyncratic wear. New Fender “relic” Stratocaster on the right, mimicking the natural wear that no longer happens on all but the highest-end (and thus, nitrocellulose-finished) guitars.
The thylacine (left) does not equal a wolf (right). Math is hard, ok?
Discrimination of prey size in the thylacine. Canonical Variates Analysis (CVA) discrimination scores of preferred prey size (> 45% or < 45% of predator body mass) within the total cranial (top), facial (middle), and neurocranial (bottom) datasets. The thylacine (hot pink) fits closer to the small-prey specialists (blue) that the (green) large-prey specialist overall, and particularly in respect with the face. The neurocranium tells the opposite story though! Curious!
Comparison of cranial shape between the thylacine, wolf, and the 'maximally convergent species' group. The Grey wolf on left, then the average skull shape of the 'maximally convergent group' (Chrysocyon brachyurus, Lupulella adusta, Lu. mesomelas, Lycalopex gymnocercus), then the thylacine. The Euclidean distances of the meshes from the mean thylacine shape are shown by colour, with white areas closer, blue smaller, and red larger. The wolf is much larger & more robust in the face than the thylacine.
Four canids with the strongest signals of convergence with the thylacine. Clockwise from top left: Maned wolf, Side-striped Jackal, Black-backed Jackal, and Pampas Fox.
A phylomorphospace (lineage & shape space) plot of the 3D data, generation via Principle Coordinate Analysis (PCA). Along the X and Y axes are skulls warped to the extremes of each axis, showing the general trends in shape as we move through shapespace. Top left is 'cat' shaped, top right is 'dog' shaped. The bottom (negative PC2 values) generally show longer, less-tucked neurocrania than the positive values. This is where the 'other' groups are, including all marsupials except the thylacine. The thylacine is positioned middle far-right, close to canids, and closest to Lupulella adusta (typo in the image), the side-striped jackal.
Diagram showing the two types of tests used for evolutionary convergence. One was the Stayton, 2015 test that measures the 'distance closed' between two lineages in a phylomorphospace. Basically, do they move closer together than the maximum distance between ancestral nodes? The other test was the Castiglione et al., 2019 test. This calculated phenotypic vectors along lineages and tests whether the angles formed between vectors is more similar than expected, when taking time into consideration.
Skull showing the various landmarks placed on the skull in 3D. There are a total of 381 landmarks divided between fixed points, semilandmarks that were allowed to slide on curves, and semilandmarks that were allowed to slide on surface patches. This protocol let me measure & compare the shape of animals that may not have had many homologous features.
Phylogenetic tree showing examples of skulls from every species sampled, coloured by Family. Counter-clockwise from top left: Canidae (dogs & kin), Mephitidae (skunks), Procyonidae (racoons), Mustelidae (weasels & wolverine), Hyaenidae (hyenas), Eupleridae (fossa), Herpestidae (mongooses), Viverridae (civets), Felidae (cats), Thylacinidae (thylacine), Dasyuridae (quolls & Tasmanian devil), and Didelphidae (opossums)
Photo of me scanning a thylacine skull with an Artec Space Spider 3D scanner. The output is displayed in real-time on the laptop next to me. That laptop has since died and gone to heaven. I had a mustache then and thought I looked like Henry Cavill. I did not.
Diagrams showing patterns of evolution. Top left is Divergence, where lineages evolve into wholly different phenotypes. Top right is Parallel, where lineages all evolve in the same 'direction' but without becoming more (or less) similar. Bottom left is Incomplete convergence, where lineages do evolve to become similar, but actual phenotypic overlap is minimal or non-existent. Bottom right is Convergence, where lineages evolve to be similar enough that their phenotypic space overlaps.
Images of diverse dogs. Top left is Speothos, the bush dog. It looks like a bear and a corgi had a baby. Top right is Rüppell's fox, which looks exactly like an anime fox. Bottom left is the African wild dog, with its beautiful asymmetric coat pattern, long legs, and heavy, almost hyena-like head. Bottom right is the Ethiopian wolf, with a long, delicate snout.
Photo of Benjamin the thylacine in the Beaumaris Zoo, looking to the left. The enclosure is a concrete slab, with a wooden hutch and wire fence. All the behavioural observations we have of thylacines are from animals in situations like these.
Examples of convergent evolution. On the left is a phylogenetic tree, with the top out-group labeled 'Fish', then as a nested clade 'Mammal' and 'Reptile'. On the right, top to bottom, are Xiphias, the modern day swordfish; Eurhinodelphis, an extinct dolphin, and Eurhinosaurus, an extinct ichthyosaur. All the animals have a streamlined body, fins, and most strikingly, a very elongate, spear-like upper jaw that extends well beyond the lower jaw.
Four mammals, all in side-on view and facing left. Top left, thylacine, a marsupial. All others are canids, which are placental mammals. Top right, grey wolf, bottom left, dingo, bottom right, side-striped jackal. It is very obvious and striking how similar the general body plan of the thylacine is to the placental canids.
Parade of thylacines - digital sculpts and 3D scans of taxidermy specimens. Human and gray wolf on the left for scale, showing their relative size.
Graph showing predator mass (X-axis on bottom) and prey mass (Y-axis at left). A grey dashed line runs diagonally from low left to high right, this is the line of isometry (where the predator:prey mass ratio is 1:1). There is a grey vertical bar covering the predator mass range from 14.5-21 kg. This is the 'switch point' where above this range, predators almost always prey on animals equal to or larger than themselves. Below this, they almost always prey on things that are smaller than they are, often by a large degree. The thylacine, both female (orange vertical dashed line) and male (purple vertical dashed line) fall within or below this size range, indicating they likely took small prey.
Top panels show the mass estimates from the scanned whole body and skeleton mount specimens, comparing how well each estimate type matches. All of them overlap - except for the tooth-based estimates. Bottom panel shows the mass of the entire dataset of thylacine remains, using equations built from the whole-body specimens. Left is the entire dataset, showing a mean mass of 16.7 kg. Right are the masses of all of the known-sex specimens, showing that females' mean (in orange) is 13.7 kg, males' mean (in purple) is 19.7 kg.
General digital workflow, starting top left with a 3D scanned skeleton. Then, a convex hull could be 'shrink wrapped' around it (shown top right), or a life-like fleshed model sculpted around it (left). For aesthetics these sculptures were then fully textured and painted (bottom middle and right).
Digital scans of the mounted thylacine specimens. Top row, left to right: skeleton of Melbourne Museums specimen, skeletons (female and male) of Tasmanian Museum and Art Gallery specimens. Bottom row, left to right: Taxidermy of Melbourne Museums and Tasmanian Museum and Art Gallery specimens.
3D surface scan of the only whole thylacine wet specimen that still exists - a female housed at the Museum of Natural History in Stockholm, Sweden. Her body condition is extremely poor, visibly emaciated.
A moody low-light photo of a taxidermy thylacine at Museums Victoria being 3D surface scanned.
A cartoon pink thylacine upside-down and with its eyes crossed-out, indicating that it is extinct. It rests on a scale, being weighed. Note that this is not really how it is done.
Section of a 19th C periodical report of a giant thylacine which killed 8 sheep in one night and broke a double-barrel shotgun to pieces when confronted.
Jimmy McMillan and his outstanding facial hair proclaiming that the weight of this thylacine is too damn high. I agree!
Small but growing collection of Australian-region DVDs and BluRay discs
Guitar effects pedals velcroed to a PedalTrain pedalboard. The pedals are, in retrospect, a surprisingly monochrome bunch, with only the two Boss pedals (blue and magenta) and the Behringer pedal (orange) - itself a Boss clone, really standing out. At least the wah pedal has polka-dots.
Reconstruction of the short-faced kangaroo Simosthenurus occidentalis. Coat pattern is based on that of the swamp wallaby and the banded-hare wallaby. This was done as a nod to the phylogenetic position of Sthenurinae (between Lagostrophinae and Macropodinae), and to the potentially less-open habitat preference of Simosthenurus compared to modern hopping kangaroos. Simo here has a fetching dark eyemask, fierce white eyebrow patches, and a set of haute dark stripes over the hips. And of course, black gloves.
Simosthenurus occidentalis reconstruction, basic level. Articulated skeleton with silhouette illustrated around it to indicate possible in life appearance. It looks like a theropod dinosaur. No mammal should look like this.
The Green Waterhole Cave Simosthenurus occidentalis specimen - an extinct 'short-faced kangaroo' - housed at the South Australian Museum. The specimen has been photographed and digitally articulated and posed to try to reconstruct the animal. These kangaroos have large, heavy, short heads. Long arms with four-fingered hands, of which the index and middle fingers are very long with exceptionally long claws. Very short back legs for a kangaroo, with rather short feet with only one giant toe ending in a hoof-like claw. They are very mysterious, with analyses and trackways suggesting they may have walked on two legs instead of hopped. If only they had stayed alive for another 50,000 years we could have asked them.
Pencil illustration of a Thylacine, sketched from the photo of one of the Bronx Zoo specimens. Illustration by Douglass S Rovinsky
Entirely legit photo of a human and two chimpanzees. Definitely not a Simpsons screenshot.
Image of a Norway Brown Rat, looking very cute and not at all like they would hide under someone's hat and start a restaurant.
Old illustration of the Christmas Island rat, with the timeline detailing its discovery by Western science to its extinction: from 1887 to 1908. A large factor in their extinction seems to have come from introduced black rats coming in on mining ships, and the diseases they brought.
Zhora, an absolutely adorable grey-and-white kitten with soulful golden eyes. She has a dark heart-shaped patch over her nose, too. Very cute, 10 out of 10 would cuddle again.
An old photo of a meerkat, sitting up on its back legs, looking off to the left with its cute little arms folded neatly in its lap. Maybe they are watching over you, maybe they are just watching you be the best version of you that you can be. Or maybe they are watching you eat the whole boxful of Mac'n'Cheese for dinner, right out of the saucepan.
Photo of Jupiter and its four Galilean moons. Left to right, the image shows Europa, Jupiter, Io, Ganymede, and Callisto. One of Jupiter's darker equatorial bands is faintly visible. Io is visibly redder than the other Galilean moons, and Ganymede (the largest moon in the solar system) is visibly brighter than the others.
Map of the world, with the 17 countries that I have visited or lived in. That's only 7% of the countries out there. I have major blank spots including all of Asia, the Middle East, and South America.
Expanded phylogeny of Dasyuromorphia showing the placement of the fat-tailed dunnart in orange. They are a member of Dasyuridae along with the quolls and Tassi devil; they share a MRCA to the exclusion of the numbat and thylacine.
Photo showing the numbat being adorable and sticking its long, sticky tongue out. They certainly do not represent the plesiomorphic, underived state of Dasyuromorphia!
Simplified phylogeny showing that all modern non-thylacinid dasyuromorphs are equally related to the thylacine, none are closer than any other.
The simplified tree with the branches rotated around their internal nodes. Now the quoll is physically closer to the thylacine than anything else. This shows the principle of free rotation around nodes, and why the MRCA (and thus clades) are more important than physical proximity, which is affected by tip choice and researcher's desire. Often, the tree is internally rotated to place the study branch on either one of the outside edges of the tree, with the other nodes rotated to "ladderise" the tree in the same direction (e.g., either towards the top or bottom of the tree)
The simplified tree showing the MRCA between the thylacine and the rest of the dasyuromorphs lived over 30 million years ago.
The same phylogenetic tree as above, showing that the nodes represent the Most Recent Common Ancestor (MRCA) of the tipward evolutionary lineages (branches). Branches that share a MRCA are more closely related to each other than ones that don't share that MRCA.