#multituberculata

[harpagornis]

New animal for Multituberculate Earth: https://multituberculateearth.wordpress.com/2022/04/03/mammals-at-sea/

[obscurefossils]

Iberica hahni

Iberica was a genus of multituberculate mammal from the Early Cretaceous Period. Its type and only species is I. hahni. Known specimens were found in the El Castellar Formation in Galve, Spain. While the fossils assigned to Iberica have been known since the 1960s, they were not officially named until 2011.

The name Iberica comes from the fossil's location on the Iberian Peninsula. The specific name hahni was named in honor of Gerhard and Renate Hahn for their research on Iberian Peninsula multituberculates.

Known fossils of I. hahni include seven P1/3s (premolars), which are the type specimen, as well as the referred material of a P4 fragment (premolar) and two M2s (molars). Its autapomorphies include the four cusps on its premolars, as compared to three cusps from related genera, as well as cuspules on the mesial and distal margins.

The material assigned to I. hahni, especially the referred P4 and M2s, cannot be certainly assigned to either Eobaataridae or Plagiaulacidae. The type premolars are similar to that of Parabolodon elongatus and Eobaatar? pajaronensis. The P4 fragment is also very similar to that of Cheruscodon balvensis. The tentative assignment to Eobaataridae seems somewhat unlikely.

References: Original description by Ainara Badiola, José Ignacio Canudo, and Gloria Cuenca-Bescós; cladistic assessment by Thomas Martin et al.

Wikipedia article: Iberica hahni

[alejandrodelorenzo]

The small pachycephalosaurid Stegoceras and the primitive mammal Meniscoessus. I don't know exactly what they're doing here. 🤷😂

[harpagornis]

New Multituberculate Earth content: https://multituberculateearth.wordpress.com/2022/03/29/kogaionidae-old-dogs-do-learn-new-tricks/

[harpagornis]

Did multituberculates and co have a corpus callosum?

Filikomys reconstruction by Misaki Ouchida. Multituberculates are now thought to have been intelligent mammals, so this discussion is worth having.

Modern placental mammals have something marsupials and monotremes lack: a corpus callosum. This allows for connectivity between the brain hemispheres and thus higher intelligence (though marsupials might be more clever than we give them credit for, and birds lack a corpus callosum and are next in line for civilization). Its thus worth to ask: did extinct non-placental groups have a corpus callosum?

On one hand, the fact that only placentals out of all amniotes have one would imply that this is a feature exclusive to them. On the other hand, marsupials and monotremes both produce undeveloped, fetal young, so they could have secondarily lost the corpus callosum as part of their simplified development. Extinct groups like multituberculates not only were as intelligent as placental mammals but had parts of their braincase not seen in other mammal groups, so it's not unreasonable to assume either the corpus callosum was a mammal synapomorphy and lost on multiple groups or it developed independently. Or maybe multies and co didn't have one and simply got smart via other ways like birds.

Hopefully future finds (or genetic and developmental studies in living mammals) answer this.

[harpagornis]

Multituberculate Earth: Djadochtatheroidea: The Children of the Desert

Djadochtatherioidea (also spelled Djadochtatheroidea or Djadochatheroida in publications because an already complicated name needs a more labyrinthine approach obviously) is probably the second best understood multituberculate clade due to the amount of complete skeletons. These animals evolved in the Cretaceous of Asia and were in fact the dominant multituberculates in the continent, to the point the only unambiguous non-djadochtatheroidean allothere from the Late Cretaceous of Asia are the taeniolabidoids Erythrobaatar and Yubaatar. They occupied a myriad of ecological niches from the predatory Kryptobaatar to the herbivorous Catopsbaatar, though they all seem to share adaptations for jerboa-like hopping and digging.

Traditionally, it is thought that djadochtatheroideans met their end in the KT event, since they are conspicuously absent from the Asian Paleocene. However, recently two Paleocene groups, Eucosmodontidae and Boffiidae, have been recovered as nested within this group; similarities have been noted between at least the former and classical Asian taxa for years now so it doesn’t come out of nowhere and I’m inclined to believe it. Both groups are still conspicuously absent from Asia, where lambdopsalid taeniolabidoids dominate, so either the Asian species became extinct in the KT event or were quickly displaced by the lambdopsalids. Considering these animals evolved in desert environments, its clear that the wetter forest world of the Paleocene didn’t do them many wonders.

Both eucosmodontids and boffiids continue the trend of djadochateroidean diversity, the former tending towards carnivory with seed side dishes and the latter being a large herbivore, one of the largest mammals of the Belgian Paleocene. In our timeline they actually endured for quite a bit until the PETM, and in this one this is no different at all, the main difference being that they survived this event, having a higher diversity as placentals declined.

Eucosmodontidae

Khusuurbaatarelegans by Dave García.One of various flying euscosmodontid lineages, this one lasted across the Eocene, with fossils found in Asia, Europe, North America, Australia and Antarctica. It was a genus of fast aerial insectivores, this species being the largest known with a wingspan of 60 centimeters.

Relatively basal members of Djadochtatheroidea, eucosmodontids possess large plagiaulacoids, apt for their primarily carnivorous diets with occasional granivory and frugivory. Across the Paleocene and Eocene they mostly diversified as small sized carnivores and omnivores, occupying niches similar to those of muroid rodents, tarsiers and cats, with some otter and desman-like swimmers. Like their Cretaceous relatives they are well adapted to hopping, and larger species resemble carnivorous kangaroos that act like dogs, hopping after prey for long distances. Also like their Cretaceous ancestors (and most multituberculates for that matter) they possess tarsal spurs, though they are rarely venomous as they use that energy and resources for hopping.

Myypbaatar eocursor, a civet or cat like eucosmodontid from the Eocene of Mongolia. Like many of its peers it was a predator, but also an opportunist fruit and seed-eater. By Dave Garcia.

In terms of predatory niches, eucosmodontids carefully remained in the background as other clades like ptilodontoideans and microcosmodontids duked it out. This lasted until the Miocene, when a combination of expanding grasslands and the decline of other multituberculate predators allowed them to sky rocket both as mouse analogues and as carnivoran analogues, mimicking the ecological radiation of those clades in our timeline.

And they were quick to turn to another ecological goldmine: the skies. And in this a group of mostly unassuming hoppers produced some of the most spectacular mammals of this timeline.

Hopping animals, eucosmodontids were prime material for developing powered flight. Most species were only semi arboreal, but in the rainforest world of the Eocene it paid off to be able to use both resources on the trees and the ground. Hopping offered the begins of a flight stroke in the extension of the forelimbs before landing, and as the animals developed various airfoils different wing types evolved. Some like Plummobaatar developed “feathered” wings similar to those of pteroectypodids, while others like Khusuurbaatar developed flying squirrel-like styliform bones on the wrist, that quickly elongated to form a pterosaur-like wing. Others still just developed bat-like wings.

This innovation proved quite useful. Like all flying mammals and pterosaurs, acamapichtliids launch quadrupedally, vaulting using the forelimbs. The backwards-pressing styliform bone adds more power, acting like a spring; likely an ossification of the triceps tendon, it is pulled backwards by the humerus muscles, and against the ulna by wrist tendons. As such it is suffused by tendons inside and out, extending or flexing the bone, preventing it from breaking by distributing pressure and adding additional energy as they move. Combined with the respiratory benefits of the epipubic bones (to which are attached muscle systems responsible for lung ventilation), this allowed acamapichtliid flight to be energy efficient, and they quickly attained massive sizes.

Already by the Rupelian/Chattian boundary these animals attained wingspans of 7 meters, comparable to those of the largest flying birds and contemporary insulonycteriids and surpassed only by the long gone azhdachid pterosaurs; this might represent the current size limit as the wing length is limited due to the styliform folding against the lower arm when in disuse, though derived species are getting around this problem by developing bent wing tips, digitigrade forelimbs and a more flexible yet strong styliform with various ossified tissue and tendon arrangements, all allowing for a potentially longer wing with top-tier folding.

Still, the low aspect ratios are mostly favoured by inland soarers, which is exactly what these animals specialised as. Occupying a niche similar to those of azhdarchid pterosaurs, they stalk their prey in the ground, though their tends to be proportionally larger since all terrestrial mammalian carnivores above 30 kg need to subsist on prey about the same size or larger. This is mostly not a problem the smaller members of this group with wingspans from 2-4 meters, which tend to be under 30 kg, but for the titans they are the first great raptorial flyers, competing both with birds of prey and insulonycteriids as well as terrestrial carnivores like microcosmodontids, ptilodontoideans and flightless eucosmodontids. On island ecosystems acamapichtliids can very well be the apex predators, though even here they retain the ability to fly.

Besides the styliform, acamapichtliids have collagen membranes fibers to keep their wings from fluttering, similar to pterosaur aktinofibrils; these fibers also further help disperse stress and strengthen the styliform when launching. Their uropatagium is supported by a modified calcar-like tarsal spur but it is otherwise small, the long tail being free from it. It is instead used as a display device, for these animals are rather social, gathering in massive flocks when not hunting.

Unlike most djadochtatheroideans, which have fast breeding cycles, acamapichtliids are K strategists. At birth the young can already walk and run, but like all flying mammals they can only take to their air when they’re close to adult size and the young are rather small in proportion to the mother in order to save weight (i.e. a species with a seven meter wingspan produces a pup the size of a domestic cat). This leaves them at least one year on the ground, so to lessen vulnerability most species form creches like those of flamingos, even located in less hospitable environments like salt flats to deter predators. Both parents take care of the young, males capable of producing milk much like in some bats. Even after growing large enough to fly the young may stay in the vicinity of the family group, with males being the most likely to leave.

Flightless eucosmodontids are mostly retricted to the northern continents, but both acamapichtliids and some Eocene flyers have managed to reach as far south as Antartica. In the southern continents they frequently meet their distant relatives, the boffiids.

Tauutus peon, a derived bipedal boffiid from the Late Oligocene of Africa. These animals tended towards fast running herbivorous niches, and eventually graduated from kangaroo-like hopping to theropod-like running. By Dave García.

More derived and related to forms like Catopsbaatar, boffiids debuted as the large herbivore Boffius splendidus and indeed most members of this group are herbivorous, even losing the plagiaulacoid. They seem to have been displaced from Asia by lambdopsalids in both timelines, and in this one they kept diversifying in Europe’s Eocene, where they produced a myriad of hopping herbivores, some as large as a red kangaroo.

Messelboffius atraxa, a red-kangaroo sized hopper from the Eocene of Europe. By Dr Spooky.

The Grand Coupure led to the end of these forest-dwelling hoppers in Europe, but a lineage managed to raft its way to Afro-Arabia during the mid-Eocene. Here they diversified, there being not only no other djadochtatheroideans besides them but also no other fast-running small herbivores and omnivores, galulatheriids being specialised herbivores, kogaionids hypercarnivorous and afroptilodontoideans stuck in the trees (barring some terrestrial seed eaters). As such, they underwent a massive adaptative radiation, giving rise to a variety of kangaroo and springhare-like hoppers, anomalure-like tree climbers and even tree-kangaroo-like folivores as leaves were too icky for afroptilodontoideans. Most notably, one lineage became bipedal runners, likely evolving in a similar manner as our timeline’s sthenurine kangaroos by being forced to walk bipedally while browsing, though unlike them they were truly agile animals like small dinosaurs. A number of species also appear to have been in an intermediary state, between bipedal runners and hoppers; in general, hoppers preffered open environments and were grazers while runners were forest animals and browsers, with both groups having mixed feeders.

Diagram of Messelboffius atraxa, showcasing its musculature. By Dr. Spooky

Two lineages rafted forth from Africa in the Eocene/Oligocene boundary: one to Madagascar and one to South America. The former diversified in a way similar as they did in Africa, but the latter were met with an essentially overcrowded continent with all manner of mammal lineages. For now, they simply occupy small hopping herbivore niches similar to that of our timeline’s Argyrolagus, but the decline of local omnivores and carnivores by the late Miocene might imply an expansion, particular as eucosmodontids haven’t reached South America yet…

[harpagornis]

Multituberculate Earth: Eocene

After the PETM, the world became a hothouse planet. Tropical rainforests stretched from pole to pole and sea levels rose exponentially. In our world, these paradisiacal conditions were actually fairly short lived, due to the Azolla Event, but there this does not happen in this timeline for reasons you’re about to find out. Thus, the Eocene truly is a tropical earth from beginning to end, and in these prolonged warm conditions live thrives.

Reptiles most assuredly benefit from this tropical world. Though squamates aside from snakes are no longer here with us, allocaudates and sphenodontians diversify like crazy, some of the latter even attaining megafaunal sizes. Crocodylomorphs diversity both as terrestrial sebecians, planocraniids and mekosuchines as well as various aquatic forms, including marine dyrosaurids and gharials. Turtles do much the same, the seas having several lineages while tortoises debut on land. Birds, the sole living dinosaurs, diversify in ways similar to our world, with large flightless gastornithiforms, ratites, penguins and plotopterids as well as massive flying pelagornithids and lithornithids, but the specific make-up of their diversity will change drastically as the cooling conditions that killed off some groups and allowed some modern groups to thrive never came about.

Mammals are no different. Therians may be reduced to a few relics soon to die out, but gondwanatheres, dryolestoids, monotremes and of course cimolodont multituberculates thrive.

In the northern continents rising sea levels decrease the formation of land bridges and thus four unique faunas develop:

North America is probably the closest to “Paleocene classic”, with taeniolabidoids as the largest land herbivores, ferugliotheriid gondwanatheres as small to mid-sized herbivores and carnivorous ptilodontoideans, eucosmodontids and microcosmodontids occupying various carnivore/omnivore tiers. But adalatheriids and lambdopsalids have arrived from Asia, meniscoessids are gone, and microcosmodontids now threaten the supremacy of ptilodontoideans, some being the de facto apex predators.

Asia has changed drastically due to the collision with India. Gondwanatheres now run the show as far as herbivorous niches are concerned, with only lambdopsalids remaining otherwise. Ptilodontoideans, eucosmodontids and microcosmodontids all vie within predatory guilds, the former two also forming the bulk of arboreal forms. Relictual kogaionids cling to India and Southeast Asia.

Europe is now an island continent. It retains some Paleocene groups (taeniolabidids, ptilodontoideans, microcosmodontids, eucosmodontids, boffiids) alongside new arrivals from North America (ferugliotheriids) and Africa (galulatheriids). Overall, mammals are rarely the largest land animals, that title going to birds like gastornithids.

Balkanatolia is also an island continent, but dominated by kogaionids in predatory roles and meniscoessids and galulatheriids as herbivores.

Africa sees the arrival of ptilodontoideans and boffiids, occupying primate and rodent like niches respectively. However, galulatheriids and kogaionids still rule the show, composing most of the megafauna. The neighbouring Madagascar shares a similar faunal composition, albeit with sudamericids and adalatheriids instead of galulatheriids.

South America, Antarctica and Australia share a similar faunal block with ferugliotheriid, greniodontid and sudamericid gondwanatheres as herbivores, monotremes as piscivores and molluscivores, ptilodontoideans as arboreal omnivores and carnivores and dryolestoids in a myriad of roles from small insectivores to giant herbivores and apex carnivores. Recently, galulatheriids and African ptilodontoideans have arrived to South America, resulting in some ecological turnovers as the natives adjust to these newcomers.

In this tropical world, mammals aren’t just satisfied with the ground. For the first time multituberculates and dryolestoids take to the skies, mirroring the evolution of bats in our world, albeit instead of one lineage of flying mammals there are several. The first marine mammal is the taeniolabidoid Iqiqquq, an animal which will leave no long term descendents but will have drastic ramifications for this timeline. Other mammals will venture into the seas, but for now it is the marine reptiles that rule.

The Eocene hothouse conditions will keep going until the very end. Carbon was sequestered far more slowly than in our world, but eventually enough was that, in combination with Antarctica’s isolation, it causes a rapid drop in temperatures. Combined with the collision of Europe, Balkanatolia and Asia and subsequent faunal exchange, this will cause another but larger extinction event, the Grand Coupure.

Table of contents:

Azolla Shmolla

The Last Placental

Example Site: Uzunçarşıdere Formation

The Grand Coupure.

[harpagornis]

What is a multituberculate?

Multituberculates were in our timeline a highly diverse group of non-therian mammals. They first appeared in the Jurassic, likely splitting off from our therian ancestors as well as the ancestors of monotremes around this time, and came to dominate Cretaceous mammal faunas, in some sites even more common fossils than dinosaurs; they reached an even higher peak during the Paleocene, before plunging to a few relics that lasted for a few tens of million years longer. In this project’s timeline, the last part didn’t happen.

Multituberculates are not as flashy as dinosaurs, pterosaurs or even some other stem-mammals like Dimetrodon or the gorgonopsians, but they are insanely interesting in their own way. This project might as well be an excuse to explain why I think so.

The first multituberculate ever described was Ptilodus mediaevus, by famous time-waster and petty eccentric Edward Drinker Cope, of the famous Bone Wars. While many of his interpretations of extinct animals are hilariously wrong, his initial assumption, that this was some kind of marsupial, wasn’t reasonably wrong at the time. It certainly wasn’t a placental mammal, and had some vaguely marsupial-esque traits like the presence of epipubic bones.

Yet, as more fossils came in and we got a clearer picture at both multituberculates and mammalian evolution in general it became extremely clear these were not marsupials, and indeed not part of Theria. That’s pretty much it, because multituberculates are seriously strange animals, as we will see in a bit.

There are three main interpretations of where multituberculates fit in the tree of life:

As mammals (slightly) more closely related to therians than to monotremes

As mammals (slightly) more closely related to monotremes than to therians

As essentially non-mammals but instead closely related mammaliform synapsids.

In general the first option is favoured by classical studies, but recent ones have shown that alleged straits securing this position are even less strong than previously thought. Therefore, all bets are off until we somehow get DNA samples, which probably will never happen (but I’m willing to do anything to obtain regardless).

Stranger still is if you factor two other extinct groups into the mix, the haramiyidans and the gondwanatheres. They are in some respects very similar to multituberculates, and sometimes they all get lumped together into the broader Allotheria, but several studies have suggested these similarities are the result of convergent evolution. More confusing yet is that haramiyidans first debuted in the Triassic; given that most genetic studies seem to suggest a Jurassic origin for the last common ancestors between monotremes and therians, if multituberculates are closer to haramiyidans then we can rule them out as true mammals altogether. Not helping is the possibility that gondwanatheres may very well nest within Multituberculata, making them an additionally problematic group to figure out.

The most recent studies on these three groups seem to propose a kind of compromise: multituberculates, gondwanatheres and Jurassic haramiyidans (in the proposed clade Euharamiyida) are all true mammals in a branch slightly closer to therians than to monotremes, while Triassic “haramiyidans” are more basal synapsids. But as seen above, the traits used for this conclusion are not guaranteed.

Trivia: if multituberculates are all by themselves then Allotheria essentially becomes synonimous with Multituberculata. Therefore “dead uncle Allotheria” technically still describes multituberculates in some capacity.

Though superficially rodent like, multituberculates are insanely alien in terms of skeletal anatomy. Unlike modern mammals, which chew in orthal (up and down, “normal”), lateral (side to side; think cows) and propalinal (back to front; mostly rodents and elephants) ways, multituberculates were only capable of engaging in a chewing style with no modern analogue, the palinal stroke, which involves extending the jaw forward and pushing it back. They were capable of using one or both rows of molars to chew food, but they could not engage in any of the other forms of chewing mentioned previously (most modern mammals are capable of using at least orthal + lateral/propalinal, like for example your pet hamster switching between all free or you trying to replicate them right now).

To these ends, the multituberculate skull is modified in rather odd ways. Scar attachment sites are pretty much inverted from the “normal” condition in most mammals, resulting in strange shapes and angles. The molar teeth not only have cusps pointing backwards but instead of the tribosphenic shape you’re probably more familiar with they instead are brick-like and have many tubercules (hence the group name). While their large incisors may remind you of rodents (and even then most have four upper incisors instead of just two), they are not ever growing and were deciduous like other teeth; the main gnawing tooth is instead a large modified lower premolar known as a plagiaulacoid, used to cut through food as the jaw moves backwards. In earlier species all four lower premolars were modified into a plagiaulacoid saw, but in more derived members only one plagiaulacoid remained with relictual peg-like ones in front of it; this single blade-tooth was non-replaceable and lasted throught the animal’s lifetime.

Most multituberculates had a modern mammalian ear bone structure, which evolved multiple times among mammaliaform synapsids anyways. Less familar are odd parts of the braincase, not seen in modern mammals; in general, multituberculates have larger brain sizes than other Mesozoic mammals (though smaller than the average therian’s), which combined with evidence for complex social behaviours seem to suggest several were fairly intelligent. Contrary to therian mammals, multituberculates seem to have a large premaxilla, though it might actually be some sort of entirely new bone. The latter might be of particular relevance in regards to whereas they had muscular lips or a philtrum, as the aforementioned study cites the reduction of the maxilla in favor of the septomaxilla as the origin for the unique muscular face therian mammals have.

Beyond the skull, the oddities didn’t stop. At least a few multituberculates seem to have lumbar ribs (albeit vestigial ones), and the coracoid, a bone vestigial in therian mammals, is well developed. The forelimb is relatively “modern”, with only a few differences from the normal condition in therians, but the hindlimb has a differently shaped femur head, femur dimensions and spurs similar to those of the male platypus (which were fairly common among early mammals and could imply that mammals were ancestrally venomous); this has lead to several debates on whereas multituberculates had sprawling limbs like a reptile or could raise them beneath the body like therian mammals and dinosaurs. Traditionally multituberculates are interpreted as obligate Gollum-walkers, but footprints from other groups once interpreted as sprawlers like phytosaurs and Dimetrodon imply that they might not have been as restricted. Most complete multituberculate skeletons show them spread horizontally rather than laying on the side like most therian fossils, but at least some multituberculates were digitigrade and the gondwanathere Adalatherium has erect limbs (so if gondwanatheres are multituberculates, there’s at least one unambiguous case).

In spite of their problematic relationships, the in-group structure of Multituberculata has mostly been consistent, barring the presence of absence of gondwanatheres within their ranks. A division between the basal ‘Plagiaulacida’ (always understood to be a paraphyletic rather than natural group) and derived Cimolodonta is well understood and consistently recovered. “Plagiaulacidans” are best distinguished for their less specialised teeth, including abundant incisors and plagiaulacoids, while cimolodonts have more derived teeth including the single plagiaulacoid mentioned above and more rodent-like incisors (which still weren’t ever-growing, as also mentioned above).

Some inner specifics may change from study to study. For example a large variety of ‘plagiaulacidans’ are sometimes lumped into Plaugialacidae (which may not be recovered as a natural group in some studies) or Paulchoffatidae, while kogaionids and taeniolabidoids are often recovered as sister taxa but more recently the former are in a basal position among cimolodonts. But beyond this, things are usually stable and consistent.

Multituberculates first debut in the Jurassic, which some take to be a sign of them being true mammals since this is where genetic studies place the division between monotremes and therians – though if they’re not a ghost lineage spanning to the Triassic it is. The earliest unambiguous multituberculates are from the Middle Jurassic (168-166 million years) of Europe and Russia; Indobaatar from the Early Jurassic of India (183 million years give or take) may be even older and imply that the group evolved in Gondwana, but this animal is still of contentious affinities, possibly being a haramiyidan instead. It is possible that multituberculates evolved from haramiyidans, thus making ‘Hamariyida’ paraphyletic in relation to them, though if they’re not related they essentially appear ex nihilo in the fossil reccord, a missing link between them and other mammaliaforms still waiting to be discovered.

‘Plagiaulacidans’ diversified across the northern continents in the Late Jurassic and Early Cretaceous, though they’re typically rare barring a few sites. They can be assumed to have been prey of other groups like the carnivorous eutriconodonts, though hilariously there’s evidence of multituberculates scavenging on them. Regardless, everything changed in the mid-Cretaceous, where the spread of flowering plants greatly altered terrestrial ecosystems; as a result, many mammal groups went extinct. This includes all ‘plagiaulacidans’, but the newly nascent cimolodonts not only survived but became the dominant mammals of the Late Cretaceous in the northern continents (though curiously the oldest cimolodont known comes from Australia). In the southern continents there are a few unambiguous multituberculate remains, but for the most part mammal faunas are dominated by gondwanatheres (which again may or may not be multituberculates) and another mammal group entirely, the dryolestoids. Due to the rather incomplete nature of the fossil reccord of the south hemisphere during the Late Cretaceous it is possible there are many gaps in our current knowledge.

Come the KT extinction event, multituberculates took some time to recover, but once they did they not only kept their dominance but exceeded their Cretaceous relatives in diversity, achieving a golden age for the first few million years of the Paleocene. Some, like the North American Taeniolabis taoensis and the European Boffius splendidus, were the largest land mammals for a time, the former in particular possibly reaching as much as 100 kg. So successful were multituberculates they they seem to actually have constrained the expansion of therian mammals during the early Paleocene.

But all good things come to an end. After the Danian (early Paleocene, 66-61 million years), multituberculate diversity crashes while therians and placentals in particular rise. They still lasted until the Eocene/Oligocene boundary some 33 million years ago, but by then they were reduced to a few small carnivorous forms like Ectypodus. If gondwanatheres are multituberculates, then they lasted until the mid-Miocene 21-17 million years ago, though once again as only relictual forms.

Why fortunes turned on multituberculates has been extensively debated. The traditional interpretation is that they were outcompeted by rodents, but this implies a close ecological role that might be overstated. As shown above, multituberculates seem to have constrained placental evolution until after they were gone, implying that rodents and other placentals diversified in response to multituberculates being cleared out of the way. Additionally, its noted that the last unambiguous multituberculates were generalists, quite the opposite of the usual trends of competitive exclusion where specialised animals are the last to go (though the last gondwanathere, Patagonia peregrina, seems to have been a specialised tuco-tuco-like digging herbivore, which is more in line with that).

Possible explanations may include vegetation and climatic changes in the Danian, in which case it’d be very ironic that the very thing that caused cimolodonts to rise would also do them in.

…we come to an alternate history where this last decline didn’t happen.