Potanichthys xingyiensis, a fish from the Middle Triassic of China, living around 235-242 million years ago.Measuring about 15cm long (6″), it was one of the oldest known fish capable of aerial gliding – possessing a “four-winged” body plan with enlarged pectoral and pelvic fins, and an asymmetrical tail with a long lower lobe. It was also almost completely scale-less, which may have helped to reduce drag and make it more aerodynamic.
Living about 30 million years ago in shallow coastal waters around the southeast USA, in what is now South Carolina, it was a member of one of the very earliest groups of toothed whales known as the xenorophids. Although only very distantly related to modern forms, xenorophids show evidence of being able to echolocate, suggesting the ability was probably ancestral to all toothed whales.
Estimated to have measured about 1m long (3′3″), Inermorostrum had a very short downturned snout and was completely toothless – specialized adaptations for suction feeding on small soft-bodied creatures on the seafloor.
Unusually for a toothed whale it also had proportionally large infraorbital foramina, openings in the bones of its snout for blood vessels and nerves to pass through. This suggests the presence of well-developed fleshy lips and possibly whiskers (as illustrated here), or maybe even an electroreceptive sense similar to some modern dolphins.
Ampelomeryx ginsburgi, a palaeomerycid ungulate from the Early Miocene of France (~17 mya). About the size of a deer, around 1m tall at the shoulder (3′3″), it was a distant relative of modern giraffids.
Males sported three distinctive ossicone-like ‘horns’ – two over their eyes and a third forked one at the back of the skull – and protruding tusks like some modern deer, which probably served a similar purpose in fights against each other.
Titanoboa cerrejonensis, a boine snake from the Mid-to-Late Paleocene of Colombia, South America (~60-58 mya). Estimated to have reached lengths of up to 12-14m (39-46′) it was one of the largest known snakes of all time, about twice the length of the biggest modern anacondas and pythons. It was probably able to reach such a huge size due to a combination of factors – mainly a very warm climate and the absence of large terrestrial predators immediately following the K-Pg extinction a few million years earlier.
Despite frequently being depicted eating dyrosaurid crocodiles, the anatomy of Titanoboa’s skull suggests it primarily fed on fish. Considering that some of the fish in its tropical riverine habitat were some of the largest available prey in the area, reaching around 3m in length (10′), a piscivorous diet would actually make a lot of sense for a such a big snake.
Dinomischus isolatus, an enigmatic animal from the mid-Cambrian Burgess Shale Formation in British Columbia, Canada (~505 mya). Only about 2cm (0.8″) in total length, it had a soft cup-shaped body topped with a whorl of about 20 solid plate-like “petals”, and lived attached to the seafloor by a thin stalk.
Impressions of its internal anatomy show the presence of a U-shaped gut, with its mouth and anus positioned next to each other in the center of the “petals”. It probably fed in a similar manner to crinoids, filtering small particles of food from the surrounding sea water.
But what type of creature it actually was is still unknown. Although comparisons have been made with several different groups – particularly the tiny entoprocts – Dinomischus doesn’t seem to quite fit in anywhere.
Despite this ongoing mystery, a few other similar fossils have been found that seem to be its relatives. Specimens of another species of Dinomischus from slightly older deposits in China show different “petal” shapes, and have been named as D. venustum. Another Burgess Shale animal called Siphusauctum gregarium may also be closely related.
An early relative of kangaroos, Balbaroo fangaroo. Known from a couple of partial skulls discovered at the Riversleigh World Heritage Area in Queensland, Australia, it lived during the Early Miocene (~23-16 mya) and was probably about the size of a cat, around 45-60cm long (18-24″) not including the tail.
Based on the skeletons of other closely related species, it probably wasn’t able to hop. Instead it would have moved around quadrupedally, and the shape of its feet suggest it was also capable of climbing like a modern tree kangaroo.
Mirischia asymmetrica, a theropod dinosaur from the Early Cretaceous of Brazil (~112-99 mya). Although known only from its hips and a few other partial bones, these pieces were so well-preserved that it was given a genus name that translates to “wonderful pelvis”.
In life it would have been about 2m long (6′6″), but since the known fossil represents a subadult its full-grown size may have been a little bit larger. It was probably a member of the compsognathids, closely related to Compsognathus and Aristosuchus – which would make it the only representative of that family currently known from the Americas.
The ischium bones of Mirischia’s pelvis were oddly asymmetrical, hence the species name ‘asymmetrica’, with one side featuring a hole and the other side only having a notch in the same position. The fossil specimen also had thin-walled bird-like bones, and soft-tissue impressions of intestines and a posterior air sac.
Living during the Late Triassic of Colorado, USA, (~220 mya), this 30cm long (1′) amphibian had a skull showing a mixture of features shared with both temnospondyls and modern caecilians – providing a vital “missing link” in their evolutionary history. Previously the oldest known caecilian-relative was the Jurassic-aged Eocaecilia, which already had much more modified anatomy making it harder to definitively link to other known groups.
Chinlestegophis seems to have been part of the stereospondyl branch of the temnospondyls – and an unexpected side effect of adding caecilians into this group is that many temnospondyls could now potentially also be classified as true members of Lissamphibia.
Of course, this is still just one hypothesis of amphibian evolution among several other competing ideas. Maybe it’s right, maybe it isn’t – as always, we need more fossil evidence! – but it’s certainly an interesting and surprising new development in the ongoing saga of “what are lissamphibians? we just don’t know”.
2023 Update: …But the discovery of Funcusvermis suggests Chinlestegophis isn’t a caecilian at all, but instead an unrelated case of convergent evolution!
Pakasuchus kapilimai, a notosuchian crocodyliform from the mid-Cretaceous of Tanzania (~105 mya). This 50cm long animal (1′8″) had an elongated body and relatively long limbs, and would have been an active terrestrial predator chasing after fast-moving small prey like insects.
The bony osteoderms on its body were much smaller and sparser than those found on most of its relatives – except for its tail, which was still heavily armored.
It also had some of the most complex teeth of all known crocodilians, with surprisingly mammal-like ‘canines’ and ‘molars’ that gave it the ability to chew its food.