Fossil Octopuses


Note: Phil welcomes discussion on this article in the Cephalopod Fossils forum.

Fossil Octopuses
Fossils of octopuses are by far the most enigmatic and mysterious of all the ancient groups of cephalopods. Due to their delicate structure fossils of these animals are exceptionally rare, as the soft-bodied nature of the animal does not lend itself to fossilisation. They are so unusual that there is just one known from Illinois, one from France, a handful from Lebanon and a couple of jaw fragments from Japan and Vancouver Island. In almost three hundred million years of octopod existence, the fossil record currently comprises just eight species in six genera - our entire record would fit inside a suitcase!

Very little is known about ancient octopus history, how they evolved and developed, or their lifestyle. Following is a brief look at some of the theories surrounding them, the octopus fossils themselves and the sites they were found in. These ancient forms almost certainly do not represent a single line of descent.

A little background may be of assistance in trying to understand these fossils. If we travel back in time to the Late Devonian period (408-360m) practically all the cephalopods we would find were the externally-shelled nautiloids. The very earliest ammonoids were starting to appear in forms known as goniatites, and living alongside them were a strange off shoot of the nautiloids known as the bactritids. These bactritids had long straight conical shells and share similarities with not only the ammonoids, but also the belemnoids; it is from this peculiar transitional group that all later forms of cephalopod including the coleoids (i.e. belemnites, vampyromorphs, squid, octopus, cuttlefish and argonauts), and ammonites are believed to have evolved. As common ancestor of the coleoids, it therefore seems likely that at least in some forms, the shell of these bactritids may have been internal.

All these differing forms of coleoid had another very important feature in common; they all had ten arms. Each group took this in a separate direction; the belemnites, which probably represent the most 'primitive' condition, had ten equal arms covered in hooks. The vampyromorphs probably slowly reduced one arm pair into long filaments as demonstrated in the modern deep-water relic the Vampire Squid (Vampyroteuthis). The squid, spirulas, other teuthids and ultimately the cuttlefish adapted one arm pair to form tentacles and the octopus, the most 'advanced' form of cephalopod, lost the arm pair completely leaving it with eight arms.

It is widely accepted that the octopods are closer to the vampyromorphs than the squid mainly due to the number of arms and the fusion of the head to the mantle. In addition, studies of the DNA of the coleoid groups have indicated that octopods and Vampyroteuthis have a closer affiliation than comparison with the squid or Spirula. The most likely scenario seems to be that during the Late Devonian the first octopods stemmed from the primitive vampyromorphs almost losing any trace of an internal shell by the time of our first fossil, Pohlsepia.


Precisely how, when and why the earliest octopods lost their shells is unclear. There are a number of theories including the possibility that loss of a shell aided burrowing in order to evade predators and to find forms of food that would remain largely inaccessible to the fish and sharks of the time. Another theory holds that shell loss enabled the octopods to colonise deeper water than the nautiloids and ammonites as they would not have been limited to an implosion depth by a gas filled shell. Though these mechanisms are uncertain it is clear that colonisation of the deep sea was extremely successful producing the interesting and varied forms we have today. Some of these adaptions are quite remarkable, the living cirrate octopod Stauroteuthis has even modified its suckers to produce bioluminescence!

Whether or not the primitive octopods moved into deeper waters as a response to predation pressures (where they presumably lost their shells) is a debated issue. This largely depends on whether the Devonian fish and sharks shared the same diet and occupied a similar niche, but with so few fossils to examine it is very hard to determine what precisely that niche was and if competition was really an important factor. Another theory holds that the octopuses later returned to shallower waters following the end-Cretaceous extinction having colonised the deep sea where they were immune to the worst effects of the extinction event. On the other hand, the octopus is a highly tactile and visual animal with very well developed eyes capable of astounding colour changes; it is unlikely that these features would have originated in the deep where such abilities would be largely redundant.

From the fossils alone it is clearly impossible to tell if these animals had chromatophores and were capable of colour change, but it seems possible. If it is considered that both the modern squid and octopus groups have this ability then it is highly probable that their common ancestor in the Devonian amongst the bactritids could too, otherwise we would be looking at the less-likely scenario of independent evolution of colour change ability in these groups. (It is likely that the only surviving vampyromorph, Vampyroteuthis, has lost its chromatophores due to adaption to a deep-water habitat where they would be of little use).

It is interesting to note that most of the forms of fossilised octopod that have been published bear a physical similarity to some of the modern cirrate (finned) octopod forms that are alive today, mostly confined to the deep sea. In common with modern octopods such as Cirroteuthis, Opisthoteuthis and Grimpoteuthis, they all have short squat bodies and some have powerful fins used to propel the animal whilst swimming. However, unlike these modern cirrates the fossil forms existed in fairly shallow warm or tropical waters, though in common with them they were probably bottom-dwellers. This is not necessarily a contradiction to the deepwater origin theories above, the lack of fossils of deep water octopods does not mean that they were not there, and not all cirrates were necessarily deep water adaptations.

It seems clear that cirrate octopuses are clearly an ancient and primitive form; unfortunately due to the lack of fossils, we can only speculate as to when the more familiar incirrate (non-finned) octopuses evolved. We cannot even be certain whether the incirrate octopuses gave rise to the cirrates, or vice versa, though the latter theory seems more likely given the recent discovery of Pohlsepia.

Syllipsimopodi bideni
In March 2022 the earliest known vampyropod was described by Christopher D. Whalen and Neil H. Landman of the American Museum of Natural History and Yale. Dating back to a staggering 328 million years old, i.e., in the Carboniferous, it hailed from the limestones of Bear Gulch, a site of exceptional preservation in Fergus County, Montana. The specimen had been found in 1988 and donated to the Royal Ontario Museum, but had not been studied to date.


Photo credit: S Thurston, AMNH

Bearing a superficial resemblance to a squid with a torpedo-shape, the specimen was named Syllipsimopodi bideni (Greek for prehensile foot and Joseph Biden). The specimen is 4.7 inches long, has a triangular gladius, a pair of fins, an ink sac and ten arms bearing rows of suckers.

The arm number was hugely significant as it was direct evidence that the ten-armed state was the ancestral coleoid cephalopod form; the squid later adapted two of these arms into tentacles, the vampyromorphs into filaments and the octopods lost them altogether. This was the working theory of many years but until this specimen was described had not been evidenced by any physical fossil. Two of the arms appeared to be longer than the others and may have been used for hunting and pulling prey such as ammonites out of their shells.

Bear Gulch during the Carboniferous was a tropical shallow marine bay and it is believed the exceptionally preserved fossils from the site were buried during seasonal monsoons. Fossils described from there include a large variety of exquisitely preserved arthropods, molluscs, worms and fish.


Art by Karen Whalen, AMNH

This specimen is not without controversy. Paleobiologist Christian Klug of the University of Zurich believes that the fossil is of an early coleoid from Bear Gulch named Gordoniconus beargulchensis, and that the purported gladius of Syllipsimopodi is truly a crushed phragmocone of the other animal.

So far only one specimen of Syllipsimopodi is known, but more examples are needed to determine if it is truly a ‘missing link’ discovery. If the current analysis holds, Syllipsimopodi has pushed back the record of the vampyropods by 82 million years.

Pohlsepia mazonensis (296m)
Pohlsepia mazonensis was named after the person who discovered it, James Pohl, and the location, Mazon Creek. It is the earliest octopod that has been described to date and is approximately 296 million years old. Up until the recent discovery and publication of Pohlsepia in 2000 it was thought that the octopus lineage stemmed from the vampyromorphs sometime in the mid Jurassic, so it is obvious how important this discovery was of a soft-bodied octopod from the Upper Carboniferous (Pennsylvannian) as it pushed the origin of the octopus group back at least 140 million years further. It is important to remember that Pohlsepia clearly had its own ancestors and even at this early date had clearly defined cirrate-octopus features. The true origin of the octopods must have happened a few million years before even this remarkable fossil.

The fossil hails from the Upper Carboniferous deposits at Mazon Creek in Illinois, a source of extensive coal deposits. Many other cephalopods have been found in these deposits including nautiloids and the shelled torpedo-shaped ten-armed coleoid known as Jeletzkya. Specifically Pohlsepia comes from the Francis Creek Shale Member, this site of exceptional preservation consisted of rapid deposition of silt and sediments believed to have been at the mouth of a river delta where it met the sea. It is believed that storm surges following heavy rains swept masses of sediment down the river and out to sea burying coastal and marine animals and vegetation extremely rapidly. Concretions of ironstone then formed around the dead animals very quickly. Pohlsepia originates from the 'Essex' marine deposits and is preserved as a carbon film resembling a compressed stain inside one such nodule; this is typical for most fossils from Mazon Creek.


Just one example of Pohlsepia is known; as it is in a primitive condition the octopod actually has ten arms, two of these were modified but the other eight were approximately of the same length. The animal is small and is estimated to have had a Mantle Length of just 25mm long by 35mm wide. The animal appears to lack an internal shell much as with modern cirrate octopuses. The animal is sack shaped, has no clearly defined head and has very short arms. It also had two fins on its mantle, which are longer than they are wide, again much like modern cirrate octopuses. The fossil has been preserved in a ventral aspect, eyes, a funnel, mandibles and a radula are identifiable and there is an indistinct feature that may represent an ink sac (extant cirrate octopods do not have these). No arm hooks or suckers are present. Peter Doyle and Joanne Kluessendorf published the fossil in 2000 and they have concluded that Pohlsepia should be assigned to the order Cirroctopoda.

Other researchers have been more cautious in assigning the ambiguous Pohlsepia to any modern group citing that the lack of internal structure in the stain of the fossil makes diagnostic features hard to determine. There are other issues, it is possible that the fins actually be remnants of an internal shell and the enormous distance in time between Pohlsepia and the first confirmed cirrate octopus fossils is also problematic. Further specimens are desperately needed to make sense of the evolutionary relationships.

Pohlsepia is housed at the Field Museum of Natural History, Chicago, Illinois.

Whalen and Landman, when describing Syllipsimopodi (see above) added a footnote that Pohlsepia may not even be a cephalopod due to its lack of beak, ink sac, cirri, arm-web, gladius, phragmocone or rostrum. They suggest it could be a cnidarian.

Proteroctopus ribeti (164m)

Moving forward in time 132 million years to the mid Jurassic, at a date of 164 million years ago (Lower Callovian) we find our next ancient octopus, Proteroctopus ribeti. Again, this consists of just one specimen and was discovered in 1982 in the marl deposits at Voulte-on-Rhone in France by the director of the local palaeontology museum, Bernard Riou. Luckily the fossil is exceptionally well preserved in three dimensions allowing detailed studies to be made of its anatomy.

The amazing preservation of Proteroctopus is due to the unique conditions at Voulte. During the mid-Jurassic the whole of France, excepting Brittany, was covered in a shallow tropical sea and Voulte lay in a basin in the seabed. An underwater current propelled animals from surrounding areas into the lower oxygen conditions inside this basin. It is believed that animals were then buried very rapidly thus preventing the rapid decay of tissues by bacterial action. Most unusually, mineralisation consisted of a mixture of apatite, calcite, pyrite and galena, each mineral forming at a different stage slowly replaced the soft tissues. Each mineral replaced a different part of the fossil, and in the case of Voulte, at a molecular level. Thus the bedding planes contain incredibly preserved fish, worms, starfish, sea urchins and lobsters amongst others. Other notable cephalopods from Voulte include the small dibranchate coleoid Gramadella piveteaui that is believed to be on the lineage that led to Spirula and the teuthids, and one of the earliest known vampyromorphs, Vampyronassa rhodanica.


Examining Proteroctopus, we now have more familiar looking animal. Gone is the modified arm pair of Pohlsepia, Proteroctopus has eight arms containing traces of suckers. The sac-like body appears to have been powerfully muscled and the head is not distinct from the body. A funnel has also been identified. In addition there are two large powerful blade-like fins at the rear of the mantle indicating that the animal was probably a powerful swimmer. As with many soft-bodied coleoid fossils the precise position of Proteroctopus is not clear and some researchers have argued that it is actually a vampyromorph rather than an octopod, but this is not a universally accepted view.

Proteroctopus is currently on display at the Musée de Paléontologie de La Voulte-sur-Rhône.

Styletoctopus annae, Keuppia hyperbolaris, Keuppia levanter (95m)
2009 saw the announcement of three new fossil octopuses by Dr. Dirk Fuchs - a total of five specimens displaying two genera and three species. These specimens were all dated to approximately 95 million years (Upper Cenomanian, Cretaceous) and were discovered at Mount Hajoula in Lebanon.

At the time of these octopuses Lebanon enjoyed a hot tropical climate and was covered
Keuppia levante
in a series of swamps, rivers and deltas that met the sea. These animals would have died in a hot shallow coastal region where above the stagnant seabed there existed a clearly defined oxygen-deficient layer. Any marine animal that dived too deeply would quickly find itself in trouble, but it seems likely that most fossils, including our octopuses, were animals that probably drifted down after dying from natural causes and were buried quickly. Animals lying on the seabed would have been largely free from predation and bacterial action due to these environmental conditions helping to explain the exceptional preservation of the fauna. During the Late Cretaceous the whole area was uplifted due to compression effects when the Africa-Arabian plate collided with the Eurasian plate thrusting up the old seabed to eventually form the mountainous regions in Lebanon.

Palaeoctopus (see below) was already known from Mount Hajoula, but the remarkable new fossil discoveries added a wealth of new information. The new specimens were all well preserved and displayed traces of vital internal structures, suckers, musculature, gills and arm crown. None of the three new Lebanese animals have fins but all have traces of fin cartilages that presumably would once have supported them.

Styletoctopus annae displays two small 'stylets' or rods within the mantle, these are a vestige of the 'wings' of the gladius. They can just be made out in the photograph below looking akin to two score marks. These rod structures very closely resemble that of the current genera of Enteroctopus, Benthoctopus and Eledone implying that modern-type octopuses are much older than we had previously thought and may have an origin in the Early Cretaceous or even as far back as the Jurassic. Styletoctopus has therefore been assigned to the Octopodidae. This specimen also implies that the stylet rods developed and fins were lost amongst the incirrate octopods before the Upper Cenomanian.



Keuppia levante and Keuppia hyperbolaris differ from each other in the development of growth lines in the gladius, but otherwise are very similar. The gladius has not been reduced to rods as with Styletoctopus, but still appears to have reduced wings. The fossils are remarkably preserved and display gills, ink sacs and suckers. They have been assigned to the Palaeoctopodidae to join Palaeoctopus due to similarities in the gladius. Keuppia lacks fins but is thought to have been a larger and more powerful swimmer than Palaeoctopus.



Palaeoctopus newboldi (89-71m)
Palaeoctopus newboldi (aka Paleoctopus) is known from a handful of specimens and was the first fossil octopus to be found. It was originally described in 1896 by H. Woodward and was published under the name Calais newboldi. The name was later revised to Palaeoctopus when it was realised that the name Calais had already been assigned to an insect! Unlike Pohlsepia and Proteroctopus, fossils of this animal are not unique but they are still extremely rare. Palaeoctopus is also known from Lebanon from the Mount Hajoula region and has a Late Cretaceous date of roughly 89-71 million years ago.

Palaeoctopus newboldi is preserved as a film, or tissue impression, in sandstone. It is a short squat eight-armed octopus with an indistinct head. Much as with Pohlsepia and Proteroctopus, Palaeoctopus has a pair of triangular fins on either side of its head though these are smaller than Proteroctopus. A faint trace of a web uniting the arms is visible and the presence of suckers on the arms has been identified. Due to its similarity to the cirrate octopuses the specimen has been assigned to the order Cirroctopoda, and granted its own family, the Paleoctopodidae.

Until the 2009 publication of the three Lebanese specimens described above,
was widely believed to be stem-group animal from which later Cirrates evolved, but their discovery at a slightly earlier date with their adapted gladius forms and modern appearance proves that this couldn't be the case.

The Palaeoctopus illustrated here is Woodward's 1896 specimen found at below the Old Covent, Sahel-el-Alma, Mount Lebanon, Lebanon and is in the collection of the British Museum of Natural History in London.

P. pacifica is known from Vancouver Island and latterly from Hokkaido in Japan, the species was named pacifica as specimens have been found on either side of the Pacific ocean in deposits of virtually the same age and environment. The fragments date to the Lower Campanian. A number of the specimens at the Courtney Museum in Vancouver Island had been misidentified for years as ammonite jaw fragments collected by local fossil hunters after been washed down local rivers, but were recognised by Dr. Tanabe of the University of Tokyo as to their true nature during a visit.

P. haggarti is known from Vancouver Island only and examples date from Santonian and Lower Campanian. Like P. pacifica the specimens are preserved in sedimentary rock concretions. It was named after Dr. Jim Haggart, the West coast ammonite expert.

Similarities in the shape of the jaws have assigned the genera to that of the Cirroctopods, although they are much larger than current members of this group. This implies that large, heavily built finned octopods were cruising the north Pacific in the Late Cretaceous. The animal probably would have probably born a close resemblance to a large Dumbo Octopus, and were named after the cirri, or hairs, on the arms of modern cirrate octopods.


Post-Cretaceous Octopuses
Following the Cretaceous extinction our knowledge of octopuses actually gets even worse. Evidence for octopuses consists of drill-holes in marine gastropods and these have been extensively studied. Despite the lack of an actual fossil of the creature in 1993 a name for the missing 'phantom' octopus was raised following extensive work on Miocene-period drill holes in European scallops, Oichnus ovalis. By 2002 at least seven separate Oichnus species have been identified differing from each other by minute variations in the borings. One Caribbean species, Oichnus excavatus has had its unique form of drill holes traced back to the Late Cretaceous in echinoids and probably thrived at least until the Middle Miocene. Such an animal that has been assigned a name for which we only have trace evidence and no body fossil is called a 'ichnogenus/ichnospecies'.

Argonauts (a.k.a Paper Nautiluses)
Finally, a quick word about Argonauts. The researcher Adolf Naef in 1923 suggested that some ancient octopods began to use ammonite shells to hide in, much as some octopuses do today for camouflage purposes. He speculated that they repaired holes in the ammonite shells using specialised glands on their arms, and by the time ammonites became extinct they had fully developed the skill to replicate a full imitation shell.

However, there are a number of fundamental problems with this theory. Although the secreted shell of the female argonaut certainly bears a superficial resemblance to the ammonite, the argonaut shell differs not only in form but also in function and mineral composition. Their shells are mostly composed of calcite whereas the ammonite was mainly made of aragonite (mother of pearl). Additionally argonaut egg cases lack internal chambers and a siphuncle and are only created by the female, unlike those of the ammonite.

Another major problem is the significant gap in the fossil record between the extinction of the ammonites and the earliest fossil Argonaut. Until recently a shell described as Obinautilus was thought to be the oldest such fossil at 29 million years old, which still leaves a 36 million year old gap with the most recent ammonite. However, Obinautilus has been recently re-described as a nautilus. This leaves our oldest fossil as Argonauta absyrtus from Cyprus and dating to a mere 12 million years old, and a few slightly more recent egg cases from
Los Angeles, Japan and elsewhere (Mizohubaris, Izumenauta and Kapal). Of course older fossils may be out there, it just that we haven't found them yet!

One recent theory holds that the argonaut shell was developed to protect the eggs from UV radiation as these little octopods evolved to live in open ocean near suface environment. However, cladistic studies have hinted that argonauts may have an origin as far back as the Jurassic. Of course until fossils turn up, we have no alternative method of verifying this. Nonetheless, it is widely accepted that they stemmed from the incirrate (i.e. non-finned) lineage of octopuses. The origin of the argonauts, and how and when they split from the octopuses, is shrouded in as much mystery much as the origin of the octopuses themselves.


Blissett, DJ, Pickerill, RK. 2003. Oichnus excavatus Donovan and Jagt, 2002 from the Moneague Formation, White Limestone Group, Jamaica. Caribbean Journal of Science, Vol.39, No.2, 221-223. University of Puerto Rico.

Clarkson, ENK. 1998. Invertebrate Palaeontology and Evolution (4th ed). Blackwell.

Kluessendorf J, Doyle P. 2000 Pohlsepia mazonensis, an early "Octopus" from the Carboniferous of Illinois, USA. Palaeontology 43(5): 919-926

MacLeod ,N. 2003 PalaeoBase Macrofossils pt.2: Mollusca. Blackwell.

McCormick, Cameron. The Lord Geekington online blog about fossil octopods: Fossil Octopodes

Whalen, Christopher D. and Landman, Neil H.: "Fossil coleoid cephalopod from the Mississippian Bear Gulch Lagerstätte sheds light on early vampyropod evolution"

Any bits and pieces from many other websites, too numerous to mention!

Useful Links

Official website of the Musée de Paléontologie de La Voulte-sur-Rhône

Tree of Life web pages

Dr Neale Monks' A Broad Brush History of the Cephalopoda

Picture Credits

Proteroctopus and Ammonite by the author

Timeline by the author

Grimpoteuthis from Deep-Sea Photography

Syllipsimopodi bideni photo by S Thurston, AMNH

Syllipsimopodi bideni art by Karen Whalen, AMNH

Pohlsepia adapted from Kluessendorf and Doyle (2000) as above

Proteroctopus from the Musée de Paléontologie de La Voulte-sur-Rhône website

Keuppia levante by Smokybjb (Wikipedia usage permissions given)

Styletoctopus and Keuppia fossil images courtesy Dirk Fuchs (2009)

Palaeoctopus from the British Museum of Natural History

Cirroteuthis courtesy Paul H. Yancey, Whitman College

Fossil argonauts courtesy

Thanks To:

Dave Lindo, TPOTH, Bernard Riou, Paul Yancey, Joanne Kluessendorf Peter Batson (all 2004) and Dirk Fuchs (2009).

Phil Eyden November 2004, updated May 2010, and again in April 2023.
Original publish date
Nov 19, 2004
About the Author
Phil joined the staff in April 2003. He collects fossils as a hobby, frequently plundering a quarry at Folkestone in the U.K. He has a degree in British archaeology and works for a government department at Dover in England.


HI @wtacollins, @Phil may have other insights, but FWIW, based on this recent post on TONMO, I'd say that's probably a sea urchin:
Hi ya. Glad you liked the article. Yes that is a sea urchin, or rather concretion that has broken off the fossil, so it is concave - a 'negative', if you will. If you push some blu-tac into it and carefully pull it out without bending it, you can imaging that is the surface of the sea urchin (that's a highly scientific method for you). Here is a similar one I found in a pebble on a beach in Deal in Kent, UK, when I was very young. It's mid-Cretaceous and from flint. I'd imagine yours is similar.


  • 60132910_862987427375494_1231679574044049408_n.jpg
    33.6 KB · Views: 272

Article information

Article read time
16 min read
Last update

More from Phil