Were Octopuses the largest predators in the Cretaceous seas?

Vertebrates became the top predators in almost all marine ecosystems shortly after the evolution of jaws, about 370 million years ago. This has allowed them to shape the structure of these ecosystems, as they have repeatedly evolved large body sizes, combined with increased strength, mobility, and cognitive abilities. In the Mesozoic Era marine predators such as Fish and Sharks were joined by Marine Reptiles such as Ichthyosaurs, Plesiosaurs, Mosasaurs, and Turtles, which secondarily returned to the seas from the land. In the Tertiary Mammalian predators such as Whales and Seals also followed this route. The evolution of the ability to crush shells and other hard parts (durophagy) among various Vertebrate groups caused a major reshaping of marine ecosystems during the Mesozoic (sometimes known as the 'Mesozoic Marine Revolution') in which many marine Invertebrates became smaller, more heavily shelled, and more cryptic in their habits, in response to this increased predation.

The only Invertebrate group which has apparently been occasionally able to challenge the vertebrates for this top-predator status are the Cephalopods (Octopuses, Cuttlefish, Squid, Nautiluses, and the extinct Ammonoids), free-swimming Molluscs which independently evolved jaws at about the same time as Vertebrates. Modern Octopuses are highly intelligent mid-level predators which have either lost their shells or retain them only as vestigial structures. This loss of a mineralised shell has enabled the Octopuses to become more mobile, develop better eyesight, and much greater intelligence.

The Mesozoic fossil record has produced a number of large Cephalopod jaws which have been interpreted as those of Octobranchians (Octopuses or Vampire Squid) likely to have exceeded 2 m in total length. The owners of these jaws have been assumed to have been 'high-level' predators, but little has actually been determined about them, as there are no cases of soft tissue preservation nor stomach contents from which this could be determined. Furthermore, it is very hard to determine the diet of a Cephalopod from its jaw structure, as, while there is considerable variation in such structure across the group, it does not appear to be related to diet.

In a paper published in the journal Science on 23 April 2026, Shin Ikegami of the Department of Earth and Planetary Sciences at Hokkaido University, Jörg Mutterlose of the Department of Geosciences at Ruhr University Bochum, Kanta Sugiura, also of the Department of Earth and Planetary Sciences at Hokkaido University, Yusuke Takeda of the Spectroscopy and Imaging Division at the Japan Synchrotron Radiation Research Institute, Mehmet Oguz Derin of Morgenrot Inc., Aya Kubota of the Department of Geosciences at Osaka Metropolitan University, Kazuki Tainaka of the Brain Research Institute at Niigata University, Takahiro Harada, also of Morgenrot Inc., Harufumi Nishida of the Department of Biological Sciences at Chuo University, and Yasuhiro Iba, once again of the Department of Earth and Planetary Sciences at Hokkaido University, re-examine the status of large Cephalopod jaws from the Mesozoic.

Ikegami et al. examined wear on the jaws of fossil Cephalopods. All such jaws are made from stiffened chitin in life, and therefore are more prone to cracking and wear in more durophagous species, i.e. those which are using their beaks to crack the skeletons of their prey.

Octopus jaws. (A) The entire body of a finned octopus and the position of their upper and lower jaw. (B) Anatomy of feeding organs. (C) The upper jaw. (D) The lower jaw. Ikegami et al. (2026).

Ikegami et al. examined 15 sets of large Octobranchian jaws which had previously been described from Japan and Vancouver Island. They also discovered a series of 12 further such jaws by using a method which they describe as 'digital fossil mining', in which high resolution tomography combined with an artificial intelligence system was used to search for specimens within Cretaceous rocks from Japan. All 27 specimens came from outer-shelf environments lacking wave or current influence, so that transportation-related abrasion to the jaws is unlikely. 

Precise, automatic segmentation of an Octopus jaw fossil and its fine structures using zero-shot learning AI. Ikegami et al. (2026).

All of the preserved specimens are interpreted as having come from members of the Order Cirrata (Finned Octopuses). The excellent preservation of the specimens and the lack of any potential to have been caused during preparation in specimens which had not been freed from the matrix allowed identification of wear patterns caused by the feeding habits of the living Octopuses with some confidence. Pigmentation patterns within the jaws enabled the reconstruction of growth patterns, and the large dataset being used enabled a reassessment of the taxonomy of Cretaceous Cirratans.

Descriptive terms of the Coleoid lower jaw. (A) Inner and outer lamellae. (B) Terms for the individual parts of the lower jaw. (C) Terms for the specific morphological characteristics of the lower jaw. (D) Morphological terms for the area near the jaw edges. (E) Measurements of lengths. Ikegami et al. (2026).

Previous studies have led to the description of five species of Octobranchians from the Cretaceous, all on the basis of fossil jaws. Upon re-examination of this material, Ikegami et al. conclude that only two of these species are valid, Nanaimoteuthis jeletzkyi and Nanaimoteuthis haggarti. Both are assigned to the genus Nanaimoteuthis, which was formerly thought to be a Vampire Squid, i.e. a member of the Order Vampyromorpha, but which Ikegami et al. reassign to the Winged Octopuses, Suborder Cirrata. Modern Winged Octopuses come in two forms, long-bodied forms, which have broad jaw wings, and short-bodied forms, which have narrow wings on their jaws. Because both species of Nanaimoteuthis have broad wings on their jaws, they are interpreted as long-bodied Winged Octopuses. Taking all the specimens, including the newly discovered ones, into account, Nanaimoteuthis jeletzkyi first appeared in the earliest Cenomanian (about 100 million years ago) and disappeared in the late Campanian (about 72 million years ago), while Nanaimoteuthis haggarti first appeared during the Santonian (about 86 million years ago, and also disappeared in the late Campanian (about 72 million years ago).

Huge lower jaws of fossil Octopuses and of an extant Giant Squid. (A) and (B) The largest lower jaws of the Late Cretaceous Finned Octopus species Nanaimoteuthis jeletzkyi, (A) NMNS DS00042 3LmvTpM, and Nanaimoteuthis haggarti, (B) KMNH IvP 902001. Both specimens show extensive loss of jaw material caused by wear. (C) A lower jaw of the extant Giant Squid Architeuthis dux (NSMT-Mo 85956), a species having the largest jaw among modern Cephalopods. (A) is a digital fossil jaw visualized as a 3D model; (B) is an exceptionally well-preserved non-digital fossil jaw; and (C) is a modern jaw dissected from a carcass of about 10 m total body length. Solid lines indicate the extension of striation on the outer surface of the hood and broken lines show the estimated outline of the rostrum without wear. The hood and lateral walls lost by weathering, shown as shadowed areas, are reconstructed based on the holotype and specimen NMNS DS00182 Ru8pBBo. (A) and (C) are exhibited in a mirrored position. Scale bar is 20 mm. Ikegami et al. (2026).

In Cephalopods, there is a direct relationship between the growth of the jaws and that of the soft body, allowing reliable body-size calculations to be made from the jaws alone. Based upon this, the largest specimens of Nanaimoteuthis jeletzkyi are estimated to have had a mantle length of between 67 cm and 184 cm, with a total body-length (including tentacles) of 2.8-7.7 m, while the largest specimens of Nanaimoteuthis haggarti are estimated to have had a mantle length of between 158 cm and 443 cm, with a total body length of 6.6-18.6 m, making it potentially one of the largest predators in the seas of the Late Cretaceous.

Body size estimation of Late Cretaceous Octopuses. The graph shows an allometric relationship between the length of the jaw and mantle in long-bodied species of extant Finned Octopus. The name of the corresponding species is shown along each growth curve. The sizes of Nanaimoteuthis jeletzkyi and Nanaimoteuthis haggarti are based on their largest specimens and are indicated by black vertical lines. Reconstruction of these two species, the extant Giant Squid, and gigantic Vertebrate predators in the Late Cretaceous are shown with their maximum total length. Ikegami et al. (2026).

The largest specimens of both species have beaks blunted and worn by continuous wear, while in juveniles the beaks tend to be sharp. However, it was possible to reconstruct the unworn beak length from the striations and ornamentation of the jaw surface. Thus the largest specimen of Nanaimoteuthis jeletzkyi has lost about 5.7 mm from the tip of its rostrum, while the largest specimen of Nanaimoteuthis haggarti has lost about 10.6 mm. In both cases this represents about 10% of the total jaw length. In both species the right edge of the jaw is more worn than the left edge. Wear takes the form of chips, scratches, and polishing, with the largest chips exceeding 1 mm in both species. The outer surface of the rostrum is polished, removing the original striations, but show numerous scratches, with these reaching up to 5 mm in length, extending vertically or obliquely from the jaw edge. In Nanaimoteuthis jeletzkyi the inner surface of the oral cavity is more heavily worn than the outer surface, which is indicative of the chewing of food. In Nanaimoteuthis haggarti there are many transverse cracks on the rostrum, these again reaching up to 5 mm in length.

Wear on the lower jaw of a modern Giant Pacific Octopus, Enteroctopus dofleini, (NSMT-Mo 85957). (A) Dorsal view of the entire mouth. (B) to (D) Close-ups of the wear, the positions of which are indicated by boxes in (A). (B) Lateral view of the right jaw edge. (C) Anterior view around the rostrum. (D) Dorsal view of the rostrum. Enteroctopus dofleini is the largest modern octopus species. Scale bars are 10 mm for (A) and 1 mm for (B) to (D). Ikegami et al. (2026).

There are two living suborders of Octopuses, the Finned Octopuses, or Cirrata, which are generally found in deep ocean environments, and the Finless Octopuses, of Incirrata, which are found in coastal environments. The discovery of specimens of the Finned Octopus Nanaimoteuthis jeletzkyi in deposits from the earliest Cenomanian (about 100 million years ago) pushes back the fossil record of the Ciratta by about 15 million years, and of crown-group Octopuses by about 5 million years. The presence of these fossils in Japan and on Vancouver Island indicates that large Finned Octopuses were found in outer shelf environments on either side of the Pacific during the Late Cretaceous (at which time the Pacific was already the world's largest ocean).

Wear on the largest lower jaw of Nanaimoteuthis jeletzkyi. (A) and (B) The entire specimen in (A) dorsal view and (B) anterior view. The broken line in (A) shows the boundary where striations become lost. (C) to (H) Close-ups of wear, of which the positions are indicated by boxes in (A) and (B). (C) Lateral view of the right jaw edge with chips (arrows). (D) Dorsal view of the rounded rostral part with scratches (arrows). (E) Oblique view of the rostral part with chips (arrows). (F) Area showing striations, which extends vertically in this figure (arrows). (G) Polished area where striation is lost, with scratches (arrows). (H) Asymmetric loss of the jaw edges (arrows), indicated by a mirrored image in 50% transparency overlaid beneath the original image. All panels show the specimen NMNS DS00042 3LmvTpM. Scale bars are 10 mm for (A), (B), and (H), and 1 mm for (C) to (G). Ikegami et al. (2026).

The living Cirrata can be split into two morphological groups, long-bodied forms (families Cirroctopodidae, Cirroteuthidae, Grimpoteuthidae, and Stauroteuthidae), and short-bodied forms (Family Opisthoteuthidae). Ikegami et al.'s analysis strongly suggests that both species of Nanaimoteuthis are long-bodied forms, something which is in line with both evidence from previously reported fossils, with more than 50 genera of Mesozoic Octobranchians discovered, all of which are long-bodied forms, and molecular phylogenies, which suggest that the Incerrata arose from within the Cerrata, implying that long-bodied forms arose first and that short-bodied forms evolved from these.

Wear on the largest lower jaw of Nanaimoteuthis haggarti. (A) and (B) The entire specimen in (A) dorsal view and (B) anterior view. (C) to (K) Close-ups of the wear, of which the accurate positions are indicated by boxes in (A), (B), and (G). (C) and (G) Lateral view of the right jaw edge with chips, scratches, and cracks. (E) Dorsal view of the rounded rostral part with scratches and cracks. Schematic drawings of (C), (E), and (G) are shown in (D), (F), and (H), respectively. (I) Area showing striation, which extends vertically in this figure. (J) Polished area where striation is lost. (K) Asymmetric loss of the jaw edges (arrows), indicated by a mirrored image in 50% transparency overlaid beneath the original image. All panels show the specimen KMNH IvP 902001. The broken line in (A), (G), and (H) shows the boundary where striation becomes lost. Scale bars are 10 mm for (A), (B), (E) to (H), and (K), and 1 mm for (C), (D), (I), and (J). Ikegami et al. (2026).

Nanaimoteuthis haggarti is significantly larger than Nanaimoteuthis jeletzkyi and examination of pigment patterns in the jaws of specimens of both species suggests that it grew significantly faster. Since the first known fossils of Nanaimoteuthis haggarti appeared about 86 million years ago, while the first specimens of Nanaimoteuthis jeletzkyi appeared about 100 million years ago, this suggests that these ancient Cirratans went through an evolutionary change enabling the emergence of gigantic forms about 10 million years after they first appeared.

The growth rate of Nanaimoteuthis haggarti and Nanaimoteuthis jeletzkyi based on pigmentation patterns. (A) Schematic drawings for the ontogenetic changes from juvenile to adult jaws, based on modern Octopuses. Unpigmented regions decrease with growth. (B) Nanaimoteuthis haggarti  (NMNS_DS00182_Ru8pBBo.stl), corresponding to the left stage of (A). (C) Nanaimoteuthis jeletzkyi (NMNS_DS00173_P74Doy1.stl), corresponding to the middle stage of (A). The broken lines in (B) and (C) indicate the original outline including unpigmented parts, reconstructed based on the adults of the same species. The posterior and ventral margins of (B) and (C) taper without fractures, indicating that they retain the original pigmentation patterns. The specimen of Nanaimoteuthis haggarti has a weakly pigmented inner lamella compared to that of Nanaimoteuthis jeletzkyi  with a smaller size. This pattern indicates that Nanaimoteuthis haggarti is in an earlier growth stage compared to Nanaimoteuthis jeletzkyi. Therefore, Nanaimoteuthis haggarti grew faster than Nanaimoteuthis jeletzkyi. Scale bars are 1 mm. Ikegami et al. (2026).

The extremely large size of Nanaimoteuthis haggarti makes it larger than the Giant Squid, Architeuthis dux, which has a maximum jaw length of about 80 mm, by about 50%. The Giant Squid can have a mantle length of about 2.5 m and a total length of about 12 m, something which has made it the largest known living or fossil invertebrate until now. It also rivals or exceeds the dimensions of the largest Vertebrates in the Cretaceous Seas, including the Ray-finned Fish, Xiphactinus audax, which reached about 5 m in length, the Lamniform Shark, Ptychodus mortoni, which reached about 10 m, Plesiosaurs of the genus Styxosaurus, which reached about 12 m, and the giant Mosasaur, Mosasaurus hoffmannii, the longest specimens of which may have reached about 17 m. Thus Nanaimoteuthis haggarti appears to have been one of the largest organisms in the Cretaceous oceans.

Most living Cephalopods are generalist carnivores, preying on Crustaceans, shelled Molluscs, other Cephalopods, and Bony Fish. Wear to the tip and edges of the beak is typically present in durophagous forms such as Octopuses and Cuttlefish, but absent in non-durophagous forms such as Squid. Thus the presence of such wear can be used to make judgements about the diet of fossil Cephalopods, such as Nanaimoteuthis jeletzkyi and Nanaimoteuthis haggarti. In these Cretaceous Cirratans it is estimated that about 10% of the jaws of large adult specimens had been worn away, but such wear is absent in juvenile specimens, as it is in contemporaneous Squid. The extent of wear seen is greater than is found in any living Cephalopod, suggesting that these were active carnivores, frequently using their beaks to crush hard shells and bone. The distribution of the wear on the jaws is asymmetric, which suggests lateralized behaviour (i.e. these Octopuses had a preferred side when manipulating prey, similar to handedness in Humans), something which is associated with cognative ability in Cepahlopods, suggesting that these Cretaceous Octopuses were already highly intelligent. 

The jaws of Nanaimoteuthis jeletzkyi and Nanaimoteuthis haggarti are much shorter than the jaws of most Late Cretaceous predatory marine Vertebrates. However, the long lateral walls of these jaws suggest they had powerful jaw muscles, and the cracks and chips on the beaks suggests that they were exerting forces which exceeded the resistance of the strongest parts of the jaw. In the larger Nanaimoteuthis haggarti transverse cracking is also present, probably representing larger shear failures caused by greater forces being applied. Thus these jaws appear to have been used to break up food items of considerable size and resilience. The difference in overall size is likely to have derived from the way in which Octopuses hunt, using their elongated arms to capture and overwhelm prey, rather than their jaws as in most marine Vertebrates.

Vertebrates have been the top predators in the oceans for most of the past 370 million years. The appearance of durophagous predation in a variety of lineages in the Mesozoic enabled a greater diversity of such predators, and drove marine ecosystems towards the structure we have today. Mesozoic Invertebrates have chiefly been viewed as prey during this process, adapting to increased predation pressures by becoming smaller, more heavily armoured, and more cryptic in their habits (better at hiding). Ikegami et al.'s study suggests that some Octopuses did not follow this path, instead becoming giant predators which rose to the top of the food web.

Convergent evolution among marine top predators in the Palaeozoic–Mesozoic. This model shows the acquisition of jaws and the reduction of superficial skeletons in the evolutionary history of marine Vertebrates (top) and Cephalopods (bottom) to become top predators. The grey horizontal bars show the chronological range of some selected groups of Vertebrates and Cephalopods. For Cephalopods, stepwise reductions of skeletons are indicated by the blue background. Ikegami et al. (2026).

Both Vertebrates and Cephalopods first evolved jaws in the Late Silurian or Early Devonian, between 423 and 407 million years ago, something which greatly improved their hunting efficiency. Vertebrates subsequently lost their external bony plates, and in larger species greatly reduced their scaly coverings to achieve smooth skin, while at the same time Cephalopods first internalised and then gradually lost their shells. In both cases, this was associated with increased swimming speeds, size, and intelligence. Vertebrates became top predators in the oceans long before Cephalopods completed this process, but during the Cretaceous some Octopuses were able to evolve a bodyplan which enabled them to compete with the very largest Vertebrates.

See also...

An embryonic Synapsid from the Early Triassic of South Africa.

 The persistence of egg-laying in modern Monotremes has led evolutionary biologists to conclude that this is likely to have been the ancestral state in the Synapsids, the group from which the Mammals arose. However, fossil evidence for this has been surprisingly absent. The earliest known potential fossil amniotic egg comes from the Permian of South America, and has been attributed to a Mesosaurid Sauropsid (a group not closely related to Synapsids of Mammals). This specimen preserves an immature skeleton curled in a position consistent with having been in an egg at the time of death, but no actual eggshell (not altogether surprising, as the earliest amniote eggs are not predicted to have been mineralised). The earliest amniotic egg fossils with both embryonic remains and eggshell come from Sauropodomorph dinosaurs from the Early Jurassic of Gondwana. Some potential eggs associated with Synapsid Pelycosaurs from the Early Permian of North America are not considered to be reliable, as neither embryos nor shell structures are preserved.

The Late Triassic-Early Jurassic Elliot Formation of South Africa's Karoo Basin has produced numerous Dinosaur egg fossils with embryos, as well as the skeletal remains of many non-Mammalian Cynodonts, something which has led to questions about whether Permo-Triassic Synapsids laid eggs at all. This is a serious consideration; Synapsids, particularly groups such as Lystrosaurus and Diictodon, are extremely common in the Permian and Triassic of the Karoo, with perinate specimens (specimens thought to have died around the time of birth or hatching) being found here and elsewhere, but no eggs are known. The preservation of Dinosaur eggs in the Karoo suggests there was no taphonomic process here producing a bias against the preservation of eggs, and palaeontologists have been active in the Karoo Basin for over 180 years, suggesting that if such eggs were present, there should have been a good chance of their being found. Egg-laying and bearing live young are found in closely related Snakes and Lizards, and it appears that this group has been able to switch back-and-forth between these conditions fairly easily. It is therefore conceivably possible that Synapsids developed the ability to bear live young very early in their history, and that Monotremes have secondarily switched back to egg-laying.

However, this has wider implications than Synapsid palaeontology. Current theories on the origin of lactation in Mammals have been built on the assumption that this preceeded the switch to live-birth (largely because Monotremes produce both eggs and milk). It is now generally accepted that the purpose of lactation was not originally to feed the young, but rather started as skin secretions used to either moisturise the eggs, provide nutrients, protect them against fungi and bacterial infections, or for hormonal signalling through the egg membrane. Should it be found that the Synapsids from which Mammals evolved bore live young, then these theories would have to be abandoned.

In a paper published in the journal PLoS One on 9 April 2026, Julien Benoit of the Evolutionary Studies Institute at the University of the Witwatersrand, Vincent Fernandez of the European Synchrotron Radiation Facility, and Jennifer Botha of the Evolutionary Studies Institute and Centre of Excellence in Palaeosciences at the University of the Witwatersrand, describe three perinate specimens of the Dicynodont Synapsid Lystrosaurus from the Early Triassic of Xhariep Municipal District in Free State Province, South Africa, one of which appears to have been preserved within an egg.

The specimens examined are the three smallest specimens attributed to Lystrosaurus. They include BP/1/4011, an isolated skull measuring 43.0 mm, discovered by James Kitching in the upper Palingkloof Member of the Balfour Formation at Orangia on Tweefontein 508, BP/1/9332, an almost complete articulated skeleton with a skull length of 44.0 mm, discovered by Brandon Stuart in the upper Palingkloof Member of the Balfour Formation at Nooitgedacht 68 Farm near Spitskop, and NMQR 3636, a complete skeleton with a skull length of 34.5 mm, found by John Nyaphuli at Rheeboksfontein 5 Farm in 2008, probably from the upper Palingkloof Member of the Balfour Formation or the lower Katberg Formation.Each of these fossils was a scanned at the European Synchrotron Radiation Facility in Grenoble, France, with three dimensional models being reconstructed with the Avizo Software Package.

The isolated skull BP/1/4011 was described by Kitching as the smallest known skull attributed to Lystrosaurus in 1964, and attributed to either Lystrosaurus murrayi or Lystrosaurus curvatus by a study in 2006. Benoit et al. are more cautious, attributing it to Lystrosaurus sp. but suggesting it shows affinities to Lystrosaurus curvatus.

The first of the articulated skeletons, BP/1/9332, is considered to be an early juvenile of Lystrosaurus sp., with affinities to Lystrosaurus murrayi. It is preserved in a splayed out position, similar to that of most larger Lystrosaurus specimens from the Karoo Basin, with most bones perfectly articulated, and synchrotron images show that no loose elements are preserved in the surrounding matrix. It appears to be the most developmentally advanced of the three specimens, because its splenials are co-ossified at the mandibular symphysis, although its occipital and basicranial bones remain loose. From the splayed out position in which it was found, Benoit et al. determine that it had hatched before dying, probably moving some distance from its hatching site before death.

Photograph of BP/1/9332 in dorsal view. Benoit et al. (2026).

The final specimen, NMQR 3636, is also considered by Benoit et al. to be an early juvenile of Lystrosaurus sp., with affinities to Lystrosaurus murrayi. However, unlike BP/1/9332, this specimen is curled into a fetal position, consistent with having been within an egg at the time of death. It also appears to be the most developmentally immature of the specimens, lacking tusk buds in its maxillary alveolae, something present in both the other specimens, or a mesethmoid bone, the structure that supports the olfactory bulbs in life, which is again present in the other two specimens. 

Most notably, the lower jaw of NMQR 3636 has an incompletely co-ossified symphyseal suture between the two paired bones in the lower jaw. This is completely co-ossified in both the other specimens, as well as in modern beaked Amniotes such as Turtles and Birds at the time of hatching. Modern Monotremes do hatch with an unfinished intermandibular symphysis, but these feed on milk provided by their mothers for some time after hatching, something Lystrosaurus is not thought likely to have been able to produce. 

Based upon this, Benoit et al. conclude that the early developmental stage of the skeleton, combined with a posture which would be expected of a perinate prior to hatching and a jaw which had not developed to the stage where it could feed on the hard foodstuffs likely to have been consumed by juvenile Lystrosaurus. is indicative of an Animal which died within the egg and was subsequently preserved, albeit without preservation of the egg itself.

Specimen NMQR 3636 in left lateral view. (a) Photograph of the specimen; (b) 3D digital reconstruction of the segmented bones; (c) live reconstruction by artist Sophie Vrard. Colour code for (b): vertebral elements in shades of green, ribs in blue, forelimb elements in red, femur in yellow, pelvic girdle elements in grey, skull in light red, mandible in light orange. Benoit et al. (2026).

Based upon the position of the embryo, it is estimated that the original egg was 3.65 cm long and 2.75 cm in diameter, with an internal mass of 115 cm³ and a mass of 115 g. While size estimates for adult Lystrosaurus vary, this is clearly larger compared to the size of an adult than either living Monotremes or most non-Avian Reptiles, although comparatively smaller than the eggs of Birds. This is probably indicative of a large yolk, which can feed the embryonic Animal for longer, allowing it to develop further within the egg. 

Modern Monotremes produce small eggs compared to the size of an adult, which contain comparatively little yolk material. This is possible because the young hatch at an early developmental stage, and are then nourished with milk. Interestingly, the Jurassic Tritylodontid Cynodont Kayentatherium produced eggs which were even smaller compared to the size of an adult. While Kayentatherium has been reconstructed as being quite Reptile-like in physiology, the small egg size could be a sign that it was capable of a form of lactation. It has also been suggested that Kayentatherium probably had hair, something which is known to be linked genetically to the formation of mammary glands (which produce milk), and it has also been shown that there is a genetic link between the reduction in egg yolk production and the ability to produce milk. All of which suggests that Kayentatherium may have been more Mammal-like than previously reconstructed, and that the appearance of the ability to produce milk may have been closely linked to the emergence of the Mammaliamorpha.

See also...

Did the earliest Vertebrates have four eyes?

Vertebrates have complex, camera-type eyes which have been a source of interest to evolutionary biologists since the nineteenth century, when this seemed an unusually complex system, which it was difficult to imagine arising through a series of gradual steps. Modern evolutionary biologists are less concerned by this, recognising that even a very simple eye is better than no-eye-at-all, and that therefore a complex eye could arise step-wise from the simplest cluster of light-sensitive cells, but beyond this have been able to give no real explanation of what the eyes of our earliest Chordate ancestors looked like. 

Camera eyes comprise a comprise of a spherical lens, a retina, an iris, and a set of muscles exterior to the main eye structure, which can be used to alter the shape of the lens, enabling it to focus an image on the hemispherical retina, which are detected by the optic nerve, and transmitted to the brain.

Almost all modern Vertebrates have two lateral camera eyes, although some groups have lost these, and, curiously, some Lizards have a third such eye on the top or back of their heads, which is derived from the pineal complex of the brain.

Eyes in Vertebrate fossils are often identified by the preservation of the pigmentation from the retinal epithelium, which is rich in melanin, as dark stains, and/or by impressions left by the hard lens. The oldest purportative fossil Vertebrate eyes are seen in Metasprigginna walcotti, a probable Chordate from the Burgess Shale of Canada, dated to about 505 million years before the present. In these fossils a hemispherical shape has been interpreted as the retina, and an associated circular area as the lens. The earliest known example of melanostomes (the cells which contain the pigment melanin) being preserved in the eye of a Vertebrate is the Devonian Jawless Fish Euphanerops longaevus, from the Escuminac Formation of Canada, which has lateral eyes with abundant such cells, inferring the presence of a retina.

No non-Vertebrate Chordates possess a camera eye. The Lancets, or Amphioxi, have four clusters of photoreceptor cells, but are not thought to be able to produce an image (unsurprising since they also lack a brain). Salps, which are planktonic Tunicates, have a multiple stage life cycle, with an colonial adult phase which reproduces sexually, and a solitary adult phase which reproduces asexually. The larval form of the colonial Salp has three pigment cup eyes, while the larval form of the solitary stage has a single eye. During the embryonic development of Vertebrates, the paired eyes arise from a section of the anterior neural plate which also gives rise to the pineal organ, leading some biologists to speculate that these three organs are analogous to the three eyes of the Salp larvae. 

In a paper published in the journal Nature on 21 January 2026, Xiangtong Lei of the Center for Vertebrate Evolutionary Biology and Institute of Palaeontology at Yunnan University, Sihang Zhang, also of the Center for Vertebrate Evolutionary Biology, and of the State Key Laboratory for Vegetation Structure, Functions and Construction at Yunnan University, Peiyun Cong, also of the Center for Vertebrate Evolutionary Biology, and State Key Laboratory for Vegetation Structure, Functions and Construction at Yunnan University, as well as the Oxford University Museum of Natural History, Jakob Vinther of the Palaeobiology Research Group and School of Biological Sciences at the University of Bristol, Sarah Gabbott of the Centre for Palaeobiology & Biosphere Evolution at the University of Leicester, Fan Wei again of the Center for Vertebrate Evolutionary Biology, and State Key Laboratory for Vegetation Structure, Functions and Construction at Yunnan University, and Xing Xu, once again of the Center for Vertebrate Evolutionary Biology and Institute of Palaeontology at Yunnan University, and of the Institute of Vertebrate Paleontology and Paleoanthropology of the Chinese Academy of Sciences, identify organs which they believe are eyes in two species of Myllokunmingids (Early Chordate Animals which may be ancestral Vertebrates) from the approximately 518-million-year-old Chengjiang Biota of Yunnan Province.

Lei et al. consider Myllokunmingids such as Haikouichthys ercaicunensis and Myllokunmingia fengjiaoa to be the earliest known Vertebrates. For their study they examined six specimens of Haikouichthys ercaicunensis and four slabs which each contained multiple specimens of an as yet unnamed new Myllokunmingid. In both species they found that the head region typically has four black spots, two larger spots being placed laterally on the head, and two smaller spots facing forward. Previous studies have identified the larger of these spots as eyes, while the forward-pointing spots have been identified as nasal sacs. 

General morphology of the lateral eyes and pineal complex with their preserved melanosomes in two species of Myllokunmingidae from the Chengjiang biota. (a)-(b) Haikouichthys ercaicunensis (YNGIP-90281) with its enlarged eye region (b). (c) Carbon (red) and iron (green) element mapping of the same region in (b), arrows denote the position of figured melanosomes in (g) and (h). (d) General morphology of the unnamed Myllokunmingid  (YNGIP 90291-b,). (e) Enlarged eye region of the unnamed Myllokunmingid (YNGIP 90292-a), illustrating lateral eyes (circles in dotted line) and pineal/parapineal organs (arrows). (f) Carbon (red) and iron (green) element mapping of the same region in (e), arrows denote the position of figured melanosomes in (i), (j). (g)-(h) Melanosomes in the eyes (g) and pineal complex (h) of Haikouichthys ercaicunensis. (i)-(j) Melanosomes in the eyes (i) and pineal complex (j) of the unnamed Myllokunmingid. Scale bars are 2 mm (a); 1 mm (d); 200 μm (b), (c), (e), and (f); 500 mm (g)-(j). Lei et al. (2026).

Energy dispersive X-ray, Raman spectroscopy, and X-ray photoelectron spectroscopic analysis of the lateral eyes and the forward facing spots are enriched in organic carbon. Examination under a scanning electron microscope revealed that these organic patches are made up of oblong or cylindrical microbodies, which measure 200-1200 nm in length, and 200-900 nm in width. Most of these microbodies appear deformed or fused together, and they are associated with pyrite minerals and a clay matrix.

In the lateral eyes of Haikouichthys ercaicunensis these microbodies are consistently oval in shape, ranging from 250 to 900 nm in length and from 200 to 800 nm in width. Element mapping suggests that these objects are carbonaceous structures with a small central hole. In the unnamed Myllokunmingid, there are two morphotypes of microstructures present, the first similar to those seen in Haikouichthys ercaicunensis, and the second being cylindrical in shape and between 400 and 1200 nm in length and between 200 nm and 550 nm in width. These structures also have a central hole. Transverse sections of the melanosomes of some living Vertebrates have also shown such a central hole.

Lei et al. next investigated the molecular composition of the microstructures using Time-of-Flight Secondary Ion Mass Spectrometry. This revealed that in both species the microstructures contained the pigments eumelanin and phaeomelanin, both of which are found in living Vertebrates, confirming that these structures are in fact melanosomes. 

The melanosomes in the lateral eyes of Haikouichthys ercaicunensis appear to be largely distributed on the horizontal axis, while those of the unnamed Myllokunmingid are spread along a diagonal axis, with the two types of melanosomes present having different distributions and pigment contents; the cylindrical cells have a higher eumelanin content (which would have made them browner in colour) while the ovoid cells have a higher phaeomelanin content (which would have made them oranger in colour). 

In living Vertebrates, melanosomes are found in the iris, choroid and retinal pigment epithelium, but layers of ovoid and cylindrical melanosomes are found only in the retinal pigment epithelium. The observed structures in the eyes of the unnamed Myllokunmingid are consistent with a retinal pigment epithelium with a similar structure. However, in the six specimens of Haikouichthys ercaicunensis examined only ovoid melanosomes could be observed. However, rather than interpreting this as a more primitive state, Lei et al. note that in the Lamprey Mayomyzon pieckoensis and the Cartilaginous Fish Bandringa rayi from the Carboniferous Mazon Creek Fauna of Illinois, a preponderance of ovoid melanosomes have also been observed in eye structures, and that relatively few living Vertebrates have have been investigated to determine what forms of melanosomes are present in their retinas.

In both Chengjiang Myllokunmingids, the central spots are smaller than the lateral spots, about 160-240 µm in diameter in Haikouichthys, and about 90-120 µm in diameter in the unnamed Myllokunmingid. These were also found to be carbonaceous in composition, and to contain microbodies which appeared to be melanosomes; in each species these were consistent with the bodies found in the larger lateral eyes, with only oval melanosomes in Haikouichthys and both cylindrical and oval forms in the unnamed Myllokunmingid. Based upon this, Lei et al. conclude that these medial organs are also preserved retinas.

Carbonaceous preservation of Myllokunmingids eyes and median dark s pots (a-h). (a)-(b) Haikouichthys ercaicunensis (YNGIP-90285) showing lateral eyes (grey) and pineal eyes (green) with lens (blue). (c) Carbon element map of Haikouichthys ercaicunensis (YNGIP-90285) head. (d)-(e) Haikouichthys ercaicunensis (YNGIP-90296) showing lateral eyes (grey) and pineal eyes (green) with lens (blue). (f) Carbon element map of Haikouichthys ercaicunensis  (YNGIP-90296), arrows indicating left pineal eye. (g)-(i) Eyes of Haikouichthys ercaicunensis showing lens (arrows). (g) YNGIP-90283. (h) YNGIP-90284. (i)  YNGIP-90289. (j), (m) lens in Elonichthys peltigerus (ROM56794). (k), (n) Lens in Platysomus circularis (PF7333). (l), (o) Lens in Bandringa rayi (ROM56789). Scale bars are 200 μm (a)-(f); 50 μm (g)-(i); and 500 mm (m)-(o). Lei et al. (2026).

As well as melanosomes within their retinas, both species show preserved lenses, which are ovoid in structure, and about one fifth of the size of the associated retinas. These structures are preserved as impressions with some relief, suggesting that they represent an original structure which was somewhat decay resistant. This placement, size, and composition is consistent with the interpretation of these structures as eye lenses, which are harder tissue than other components of the eyes, and have been found in other Vertebrate fossils, including the Middle Cambrian vertebrate Metaspriginna walcotti.

The similarity of the lateral eyes of the two Myllokunmingid species from the Chengjiang Fauna to those found in later Vertebrate fossils is taken by Lei et al. to indicate that camera eyes had appeared by the Early Cambrian. The combination of a large retinal pigment epithelium and smaller lens is consistent with a fluid-filled retinal sphere with an iris opening within which the lens is suspended, as seen in living Vertebrates. Such eyes would almost certainly have been capable of image formation, although the quality of such images is impossible to know. 

The median, forward-facing spots on Myllokunmingids have previously been interpreted as nasal sacs, or possibly pineal organs. The former explanation seems unlikely, as nasal sacs otherwise appear to have been quite a late development, not found in many later stem Vertebrates, and probably first evolving in Galeaspids (probable stem Gnathostomes) between 435 and 370 million years ago. Lei et al. report the discovery of melanosome-bearing tissues and lenses in these spots, which are again inconsistent with an interpretation as nasal sacs. They instead interpret them as paired pineal organs functioning as a second pair of camera eyes.

Lei et al. also note that the Middle Cambrian stem Vertebrate Metaspriginna walcotti also has a pair of dark spots between the lateral eyes, preserved as carbonaceous films, and that these also appear to have associated spherical objects, which may also have been lenses, suggesting that this species may also have had a second pair of median eyes.

In Lampreys, the pineal organ is photosensitive, helping the Animal to respond to changes in light levels within the environment. In Mammals, the pineal organ is entirely internal, but it is associated with aligning the neuroendocrine system with the day/night cycle. In Lizards, the pineal organ is also associated with the neuroendocrine system, but in some species retains a photoreceptive capacity. It has therefore previously been suggested that the pineal organ may have developed from some sort of precursor eye, something that has entered popular culture as the 'third-eye' theory. Lei et al. suggest that the pineal organ may have begun as a pair of photosensitive organs acting as additional camera eyes. 

The presence of complex visual systems in the earliest Vertebrates suggests that this sense was of key importance to the success of the group from very early in its history. Both the photoreceptive cells of Vertebrates and the cells of the retinal ganglion arise from nurosensory cell precursors also present in Tunicates. A theoretical model has previously been developed in which the camera eye developed via two rounds of whole-genome duplication, the first allowing for a divergence between the photoreceptor cells and the optical ganglion cells, the second between the pineal complex and the lateral eyes. The apparent presence of a second pair of camera eyes associated with the pineal complex in Early Cambrian Myllokunmingids may represent a transitional stage, in which the genes associated with the development of the eyes have been duplicated, but only just started to evolve towards the modern pineal complex.

Evolutionary scheme of visual system in early Vertebrates. (a) Thalia (Tunicata). (b) Haikouichthys. (c) Euphanerops. (d) Generalised Lamprey. (e) Sacabambaspis. (f) Shuyu. (g) Aphyocharax. Coloured regions show positions of key sensory organs: blue, eyes; red, pineal. Light grey lines represent body outlines. Coloured bars represent the suggested acquisition of key characters. Abbreviations: br, brain; p, pineal; pp, parapineal; TG, total group. Cyclostome represents the Petromyzontidae and Myxinoidea total groups and Gilpichthys, which was recovered in a polytomy with those two groups. Cyclostome and Gnathostome total groups in this topology recovered in a polytomy with Metaspriggina and (Haikouichthys + Myllokunmingia). Lei et al. (2026).

Euphanerops longaevus, an anaspid-like fossil from the Devonian Escuminac Formation of Canada, which has been suggested as a stem-Agnathan (jawless Fish) also has paired median dark patches which have been shown to be carbonaceous films with structures identical to the melanosomes of its lateral eyes. Living Lampreys have a pineal eye and a smaller parapineal eye, both of which have functioning retinas (but not lenses) and are used to detect changes in light conditions. The stem Gnathostome (jawed Fish) Sacabambaspis has two pineal openings, which Lei et al. suggest are analageous to the pineal and parapineal eyes of Lampreys. Later stem Gnathostomes, such as the Galeaspids, only have a single such opening, suggesting a progressive loss of this system. Crown Gnathostomes have lost this opening completely, but some have a preserved pineal window, with an area of thin, semitransparent skull overlaying a pigmented area associated with the pineal complex. Thus an image-forming pineal complex was slowly replaced with a light sensitive organ regulating the production of the hormone melatonin, which regulates sleep patterns. Most crown Vertebrates possess both pineal and parapineal organs, sugesing that this complex was originally paired.

During the Cambrian Explosion, early Animals went through a phase of remarkable morphological innovation, with each new development changing the ecological environment in which all Animals lived, particularly as predation became more common. It has been suggested that higher levels of ultraviolet radiation in shallow waters during the Cambrian may have made the rapid evolution of vision more important, although it is likely that the evolution of predator-prey relationships would have been sufficient to drive this. The appearance of large (for the Cambrian) predators such as Radiodonts, gilled Lobopods, and stem Chaetognaths, all of which developed complex visual systems, would have made it important for smaller, non-predatory Animals such as Myllokunmingids to develop equivalent systems to evade predation and survive. 

See also...

Parafaveoloolithus xixiaensis: A new oospecies of Dinosaur eggs from the Upper Cretaceous of HenanProvince, China.

The Xixia Basin lies in Henan Province, China, to the east of the Qinling Mountains of southern Shaanxi Province, and extends roughly 100 km east-to-west, covering an area of about 518 km². Within this basin a series of Upper Cretaceous fluvial deposits overlie a Devonian basement. These deposits have been divided into three formations, the middle one of which, known as the Zhaoying (or sometimes Majiacun) Formation, comprises a 2120 m thick sequence beginning with motley argillaceous siltstones, sandstones and mudstones at the base, and grading into an upper layer comprising reddish mudstones and sandstones. This sequence, considered to have been laid down in a fluvio-lacustrine basin in an area with a generally arid climate, is noted for its production of preserved Dinosaur eggs, with at least seven described oospecies (because eggs are a record of a part of the life-cycle of an animal, can seldom be related to a species defined from body fossils, they are described under a parataxonomic system as oospecies, which are then organised into oogenera and oofamilies) from the Zhaoying and underlying Zoumagang (or Gaogou) formation, as well as ichnofossils (trace fossils), Dinosaur bones, Turtle eggs, and fossil Bivalves, Gastropods, Ostracods, Spinicaudatans, and Plants.

In a paper published in the journal Acta Palaeontologica Polonica on 17 December 2025, Qing He and Shutong Li of the School of Resources and Environmental Engineering at Anhui University, Shukang Zhang of the Institute of Vertebrate Paleontology and Paleoanthropology of the Chinese Academy of Sciences, Yifan Huang of the Prevention and Control Center for Geological Disasters at the Henan Geological Bureau, Xiqiang Cao of the Henan Scientific Academy of Land and Resources, and Hongqing Li and Mengyuan Zhu, also of the School of Resources and Environmental Engineering at Anhui University, describe a new oospecies of Dinosaur eggs from the Zhaoying Formation of Xixia County.

The new species is placed in the oogenus Parafaveoloolithus, and given the specific name 'xixiaensis' meaning 'from Xixia'. The species is described from a clutch of 13 subspherical eggs arranged in a radial pattern. The individual eggs are 123.3–142.6 mm by 97.2–127.2 mm, with shells 123.3–142.6 mm and 97.2–127.2 mm thick. The shells have a single structural layer with no visible growth lines and a honeycomb structure with straight pore canals. 

A clutch of Dinosaur egg oospecies Parafaveoloolithus xixiaensis. YJYM-01–13 (each egg has unique repository number), from the Upper Cretaceous of the Xixia Basin, Henan Province, China. He et al. (2025).

The oogenus Parafaveoloolithus belongs to the oofamily Faveoloolithidae, which includes six genera from the Late Cretaceous of China, Mongolia, and South Korea. No fossil eggs from outside East Asia have been assigned to the oofamily (some 'Titanosaur eggs' from the Late Cretaceous of Argentina have been suggested as possible members of the family, but this is doubtful), suggesting that the egg-layers had a limited geographical distribution, although they are found in a variety of different palaeoenvironments.

Thin sections (SREE X13-01) of Dinosaur eggshell Parafaveoloolithus xixiaensis, (YJYM-13) from the Upper Cretaceous of the Xixia Basin, Henan Province, China. (A₁) A single structural layer composed of loosely arranged eggshell units and the straight pore canals between eggshell units; arrows indicate the secondary eggshell units. (A₂) A line drawing showing the eggshell units in radial section. (A₃) Enlargement of the gathered egg￾shell units; arrow points to the single eggshell unit. (A₄) Growth centres of the gathered eggshell units; arrows point to the six growth centres. He et al. (2025).

Very few eggs belonging to the Faveoloolithidae have been found in clutches to date, and Parafaveoloolithus xixiaensis is probably the best known example to date. The radial pattern in which the eggs are arranged suggests that this is a true representation of how they were deposited, rather than a result of transportation and redeposition. He et al. suggest that the pattern and porosity of the eggs implies the female Dinosaur would have deposited the eggs in a roughly circular arrangement, before covering them over with sand - something which would also have aided there preservation. 

See also...