Procynosuchus delaharpeae, a synapsid from the Late Permian (~259-252 mya). Measuring about 60cm long (2′), it was one of the earliest members of the cynodonts, the lineage that would eventually lead to mammals.
Its fossils are mostly known from southern Africa, but similar remains have also been found in Europe and Russia, suggesting it was actually quite widespread across the supercontinent of Pangaea that existed at the time.
It had a long vertically-flattened tail, strong leg muscles, and paddle-like feet – all adaptations that suggest it was a semi-aquatic otter-like animal capable of agile swimming. It also had forward-facing eyes, giving it good binocular vision and depth perception while pursuing fish underwater.
Nicrosaurus kapffi from the Late Triassic of Germany, about 221-205 million years ago. Although rather crocodile-like in appearance, this 4-6m long (13′-19′8″) animal was actually part of an extinct group called phytosaurs – long-snouted heavily-armored reptiles with their nostrils high up on their heads near their eyes.
Phytosaurs’ exact evolutionary relationships are still disputed, with opinions currently going back and forth between them being archosauriformes or an early branch of the croc lineage within the true archosaurs. But either way they weren’t directly ancestral to modern crocodilians, and instead developed a very similar body plan via convergent evolution.
While some phytosaurs had very slender gharial-like snouts and probably fed mostly on fish, others like Nicrosaurus had much more robust jaws and seem to have secondarily adapted to a terrestrial predator lifestyle. They had longer limbs and a more upright posture than their semi-aquatic relatives, and enlarged fangs at the hooked tips of their jaws that may have been used to deliver a powerful stabbing blow to their prey.
Capinatator praetermissus, an arrow worm from the Mid-Cambrian of Canada (~508 mya). Discovered in the famous Burgess Shale fossil deposits, it was one of the earliest known arrow worms and also much larger than most modern forms, measuring around 10cm in length (4″).
Its mouth was surrounded by 50 hooked spines, which could be extended out to grasp onto its prey – probably feeding on whatever smaller animals it could catch – but when not in use these spines would have been kept folded up inside a fleshy “hood” around its head.
It may have been a transitional form between early large-predator arrow worms and the smaller plankton-feeders that the group later became.
Australovenator wintonensis, a megaraptoran dinosaur from the Late Cretaceous of Queensland, Australia (~100-94 mya). It was a medium-sized member of the group, about 6m long (19′8″), and despite only being known from a partial skeleton it’s still one the best-known megaraptorans – and also the most complete predatory dinosaur from Australia.
Megaraptorans were a group of fairly large theropod dinosaurs, currently known from Australia, South America, and Japan (and maybe Egypt). Their relationships to other theropod groups are rather uncertain, with different studies placing them as neovenatorids, tyrannosaurids, or most recently as an early branch of the coelurosaurs.
They had very lightly-built bodies, with bird-like bones full of weight-reducing air spaces, proportionally small heads with long slender snouts, and leg bones adapted for running. But their most distinctive feature was their hands, featuring massively enlarged claws on the first and second fingers, with the third finger being much smaller and somewhat vestigial-looking. While some other theropods like allosaurids and spinosaurids also had big hand claws, megaraptorans’ almost tyrannosaurid-like mostly-two-fingered arrangement is rather odd.
Their arms and fingers were much more flexible than those of most other non-avian dinosaurs, allowing them to reach out, grab onto prey with those claws, and then pull it in close to their bodies, restraining it in a sort of death-hug while their relatively weak jaws finished it off.
A distinctive injury to the second toe of Australovenator also suggests these dinosaurs may have been able to deliver powerful kicks like modern cassowaries.
Although very lizard-like in appearance, this animal was actually part of a completely extinct group known as parareptiles – a diverse group of early sauropsids who were once thought to be the ancestors of turtles, but are now considered to instead be the evolutionary cousins to the true reptiles.
With a total length of about 25cm long (8-10″), the structure and proportions of its limbs suggest it could run fast on its hind legs, making it one of the earliest known examples of bipedal locomotion. Since its teeth were adapted for a herbivorous diet, it wasn’t using its speed to chase down prey but was instead probably sprinting away from predators.
But unlike the sprawling running of some modern lizards, Eudibamus may have been capable of holding its legs in a more upright position directly under its body, convergently evolving a more energy-efficient posture similar to that of later bipedal animals like dinosaurs.
Ergilornis rapidus, a 1.2-1.5m tall bird (4′-5′) from the Early Oligocene of Mongolia (~33-28 mya). Closely related to modern cranes, trumpeters, and limpkins, it was part of an extinct group called eogruids – flightless birds which existed across Eurasia for a large portion of the Cenozoic from roughly 40-3 million years ago.
Although the earliest known eogruids were smaller and less specialized, and may even have still been somewhat capable of flying, later forms like Ergilornis had highly reduced wings, long legs adapted for running, and convergentlyostrich-like feet with only two toes each.
Known from several near-complete fossils that include rare soft tissue impressions, it’s the first varanopid to show preserved skin details – revealing a pattern of very lizard-like rectangular scales. If it is a synapsid this is a pretty big deal, since early synapsids were previously thought to have had scale-less leathery skin.
It also had unusual mosaic-like patches of tiny osteoderms above its eyes, a feature previously known only in some temnospondyl amphibians. Whether this was the result of convergent evolution or the trait actually being ancestral to most tetrapods is unclear.
Its slender body, long digits, and highly curved claws indicate it was an agile climber. It probably mainly lived up in the treetops, feeding on insects, making it one of the earliest known tetrapods specialized for an arboreal lifestyle.
(*Maybe. There’s apparently an upcoming study that suggests varanopids might actually be sauropsids instead.)
Its short toothless snout suggests it was a suction feeder, using water pressure differences to pull small soft-bodied prey straight into its mouth like a syringe. Along with a heavily built body similar to those of hupehsuchians, and a very long tail that made up over half of its 1.6m length (5′3″), it was probably a fairly slow swimmer living in shallow coastal waters.
It was a close relative of Cartorhynchus, and may have been similarly capable of hauling itself onto land like a modern pinniped.
We don’t know exactly when placentals originated. The first definitive fossils come from the start of the Cenozoic, but a few early ancestral forms probably already existed during the Late Cretaceous (estimated up to 90-75 mya) and only got their chance to rapidly diversify immediately after the mass extinction event.
One of the closest fossils we have to the earliest placentals is Purgatorius. Known mainly from teeth from the Early Paleocene of North America (66-63 mya), it’s not entirely clear whether it actually existed in the Mesozoic, but its remains have been found very close to the K-Pg boundary and one fossil might actually be from the end-Cretaceous.
A few foot bones have been associated with some of the fossil teeth, and if they do belong to Purgatorius then they show that it had very flexible ankles, a characteristic typical of tree-climbing animals. It would likely have been a squirrel-like creature, about 15-20cm long (6-8″), eating an omnivorous mixture of insects, seeds, and fruit. It may also have been capable of burrowing similar to modern chipmunks.
It’s often been interpreted as a placental mammal, specifically a very early type of primate, but more recent studies suggest it might not even be a true placental at all – although it was probably still a very close relative of the common ancestor of all living placentals.
A group known as the leptictidans were probably some of the weirdest early eutherians. With their tiny forelegs, big hindlegs, and long counterbalancing tails, they somewhat resembled jerboas or small kangaroos – except they also had long slender snouts that probably ended in sengi-like proboscises, and their feet were structured more like those of running animals than jumping ones. They’re also thought to have been mainly bipedal, convergently evolving a similar posture and movement style to non-avian theropod dinosaurs.
First appearing in the Late Cretaceous, they made it through the end-Cretaceous extinction and survived up until the mid-Cenozoic across the northern hemisphere, going extinct around 33 million years ago. They were probably omnivores, eating a mixture of insects, small vertebrates, and soft plant matter such as fruit and leaves.
Their mix of “primitive” skull features and highly specialized skeletons makes classifying them particularly difficult. They’ve been proposed to be placentals related to primates and rodents or afrotheres, a very early branch of the eutherians, or close to placentals but not quite true members themselves. The latter interpretation currently seems most likely, but they could also be a paraphyletic group at the base of placentals (suggesting that they could even be ancestral to placentals, and therefore all placentals would technically be leptictidans).
Gypsonictops was one of the earliest leptictidans, living during the Late Cretaceous of North America (70-66 mya). Known only from teeth and jaw fragments, we don’t know much about its appearance or full size – although it was probably smaller than its later relatives, perhaps about 35cm long (1′2″).
Any reconstruction of such fragmentary remains is going to be very speculative, but I’ve restored it here as a sort of transitional form, not yet quite as specialized. A more sengi-like animal, mainly quadrupedal but able to run and hop on its hind legs to flee from danger or chase after small fast-moving prey.
