Friday, December 23, 2022

Obscure controversies in Cenozoic marine vertebrate paleontology 3: Was there an early Miocene extinction in pelagic sharks?

A return to my previously abandoned series on lesser known controversial topics and disagreements in Cenozoic marine vertebrate paleontology - I avoid topics that I am personally opinionated about but am the only critic, as I don't want to use this blog as an alternative to publishing my own opinions. Instead, this series aims to summarize published issues between working groups. This one covers a fascinating new study and two comments/replies about a purported extinction in pelagic sharks discovered in observations of shark denticles (scales) in deep sea core samples.


    Last year (2021) Elizabeth Sibert (Harvard, Massachusetts), and coauthor Leah Rubin (College of the Atlantic, Bar Harbor, Maine) published a surprisingly provocative paper on assemblages of shark denticles from the deep sea floor. Sibert has published some pretty incredible papers over the past few years applying methods typically employed by micropaleontologists studying fossil plankton – typically critters like foraminifera – to shark and ray denticles from deep sea core samples. Deep sea cores are a critical source of biostratigraphic data for the marine record, especially for the Cenozoic – in my own research I am frequently double checking microfossil chronologies based on deep sea core samples. Core samples, for the uninitiated, are cylinders of sediment and rock drilled by a special circular drill bit. These provide a record of plankton faunal change but also can be studied from a paleomagnetic perspective and also yield a wealth of isotopic data that can be used to trace the history of climate change (e.g. water temperature) in the oceans as well as primary productivity (evaluated through the study of opal and phosphorus). Microfossils occur in such vast numbers that statistically significant samples can be collected systematically, unlike the vast majority of vertebrate fossil assemblages on land* where we have single specimens or only a handful of a particular taxon in question.

*I still think much more can be done with shark teeth in continental shelf marine rocks than just ‘here’s a list of the species we found in this layer’, but I digress.

    Elizabeth Sibert’s prior projects including with her mentor Dick Norris* have focused on the K/Pg extinction in sharks and fish and found that sharks became numerically less common right after the extinction, previously outnumbering fish teeth, but after the extinction fish teeth become more common than shark denticles (Sibert and Norris, 2015). In another study she found that pelagic fish “ichthyoliths” (fish teeth in deep sea sediments) become numerically less common in the Tethys Sea for 3-4 million years after the K/Pg extinction but in the mid-Pacific, fish “ichthyoliths” remain constant and suggest only minor perturbation of the food chain, indicating that marine ecosystems were not uniformly affected by the extinction despite loss of primary producers like zooplankton (Sibert et al., 2014). A study of fish teeth and shark denticles from core samples from across the North and South Pacific found three major different long-term faunas from the past 85 million years: 1) A Cretaceous fauna (85-65 myo) dominated by shark denticles, 2) a Paleogene fauna (65-20 myo) where sharks were slightly less common than fish but both were relatively abundant and 3) the modern fauna (20-0 myo) where denticles are very rare, perhaps 1 for every 50 fish teeth, but also smaller amounts of ichthyoliths that are more variable in vertical distribution within the core samples (Sibert et al., 2016). A useful overview of the methods developed for these studies (and modified from micropaleontology) is provided by Sibert et al. (2017).

    One caveat is that shark teeth are generally considered much more diagnostic and the majority of the shark fossil record and taxonomy of extinct sharks has been based on isolated teeth – not denticles. In this regard, denticles are used as a bit of a rough proxy for diversity – and quite a bit rougher than if only teeth were used. More on this below.

A variety of the denticles recovered from deep sea cores collected by Sibert and Rubin (2021A), as compiled by Naylor et al. (2021).

Methods and Marine Vertebrate Taphonomy

    So how do you study ‘ichthyoliths’? They need to be removed from the sediment within the core sample. In order to facilitate removal, the sample sediment is typically rinsed in dilute acetic acid (5% or 10%) unless preserved in clay, in which case it is just rinsed in water. After this, the sediment is rinsed in deionized water and or bleach; staining may be needed to make picking easier. “Ichthyoliths” are then picked one by one from the disaggregated sediment samples. Floating the fossils using heavy liquid (e.g. kerosene) might be needed in some cases to separate them from other non-fossil grains. The use of acid means that no foraminifera or other calcareous microfossils will be recovered since they will be completely dissolved. After picking, they are identified to any number of morphotypes; in the case of the Sibert and Rubin (2021A) study, they assigned denticles to 85 different morphotypes. Many of these are not diagnostic and cannot be identified to a particular genus – many sharks have similar denticles and multiple denticle types (we will come back to this in gory detail) and most fish have little conical teeth. But some sense of the diversity can be gleaned by dividing these into purely descriptive morphotypes that are generally free of judgemental taxonomic baggage. Once identified, you can count the number per morphotype – or just the number of shark denticles or fish teeth altogether – per vertical measurement (e.g. centimeters) of the core sample. Age models are produced by analyzing strontium isotope ratios, matching to the global strontium curve, and estimating sedimentation rate for the remainder of the core sample – generally thought to be slow but continuous in the deep sea, however sedimentation rate can slow down or cease completely. Once ages are assigned to the different parts of the column, an “accumulation rate” for the fossils can be assessed – X number of teeth/denticles deposited in some volume of rock over a time period. For example, denticles were being deposited at a rate of 10-30 denticles per cm^(-2) per year about 20 million years ago, just before the purported extinction. More discussion of the micropaleontological methods adapted for vertebrate microfossils (ichthyoliths) were presented in Sibert et al. (2017).

*Norris was a student of prominent marine taphonomist Susan Kidwell and his Master’s thesis research focused on the fossil invertebrates of the Purisima Formation near Santa Cruz and Half Moon Bay, California – his study, published in 1986, was a major starting point for my own work on the vertebrate taphonomy of the unit which also sought to reconcile Purisima sedimentology and stratigraphy within a sequence stratigraphic framework. Norris has since been working at Scripps in La Jolla on Paleoceanographic research based on studies of deep sea plankton from core samples.

    Now – some comments on taphonomy before we continue – there is an inverse relationship between sedimentation rate* and the abundance of vertebrate fossils. Vertebrate fossils are generally rare compared to invertebrates like mollusks and plankton in marine deposits; vertebrate fossils only seem to be common in strata deposited during periods of slow sedimentation rate. When sedimentation rate slows down, there are more vertebrate fossils per unit volume of sediment. This is why at fast sedimentation rates, vertebrate fossils are frequently so rare that they are nearly impossible to find during fieldwork. Only assemblages that were deposited during slow sedimentation rates are worth targeting for quarrying or screening, and include marine bonebeds. Another issue here is that in most marine strata, vertebrate fossils are so rare outside bonebeds that accumulation rates – how many teeth/denticles/bones per unit volume of sediment/year are fossilized in the strata – cannot be calculated, so we have generally assumed a constant accumulation rate for vertebrates. Studies of marine invertebrates, largely by Susan Kidwell, demonstrated that changes in sedimentation rate AND mollusk accumulation rate can both change. For the purposes of my taphonomic research, the lack of such a high resolution age model and rarity of vertebrate specimens outside bonebeds meant such an approach was not really feasible. But, those problems aren’t an issue with deep sea core samples. To summarize: in marine vertebrate taphonomy we usually assume that increases in the abundance of vertebrate fossils correlates most strongly with slow-downs in sedimentation rate. Bonebeds rich in teeth and bones form when sedimentation slows down or pauses (hiatal concentrations – e.g. formed during a sedimentary hiatus, or a condensed section) and lag deposits can form when erosion scours away sediment and leaves the heavier fossils behind (erosion is a ‘negative’ sedimentation rate).

*Sedimentation rate is simply measuring the vertical thickness of sediments deposited per unit time, usually in millions of years. Deposition rate is something else that avoids the problem of non-deposition, but it’s been 15 years since I took my last stratigraphy class and can’t remember precisely. Whatever it was it was such an esoteric argument that a prolonged google search is failing me – in any event, my advanced stratigraphy professor was a bit of a dick who had an axe to grind against sequence stratigraphy and the argument might have just been one of his ‘pet’ arguments anyway (of which he had many which were ultimately meaningless in the contemporary landscape of sedimentary geology).

Linear denticles (left) and geometric denticles (right) studied by Sibert and Rubin (2021A).


Findings of Sibert and Rubin (2021A)

    The study by Sibert and Rubin, published in Science, found several lines of evidence suggesting that an extinction or something similar affected pelagic shark and fish communities. Recall already that prior studies by Elizabeth Sibert had found evidence of a major discontinuity in the shark/fish ichthyolith record right about 20 myo. The first line of evidence is that many of the denticle morphotypes disappear right around 19 myo – many other morphotypes disappear earlier, with notable terminations of multiple morphotypes during the middle Oligocene at 30 myo, the Eocene-Oligocene boundary at 34 myo (associated with the onset of Antarctic glaciation), and at the Oligocene-Miocene boundary at 23 myo – if their figure 1 is interpreted literally. Instead, these authors suggest that many of these lineages disappear at the 19 myo mark and some earlier terminations might be better explained by signor-lipps effect, which suggests that at “low” sample sizes in the fossil record, a mass extinction event might be smeared out, with many species appearing to go extinct before the extinction event owing to the vagaries of fossil preservation. According to these authors, around 70% of denticle morphotype diversity disappears during the early Miocene, which is higher than the extinction rate (~40%) at the K/Pg boundary based on prior studies by Sibert using the same methods. Of the two broad denticle types – linear, characterized by parallel ridges, and geometric, characterized by more complicated shapes – both suffered considerably (66% and 88% extinction of morphotypes, respectively) but geometric denticle types declined in abundance, from 35% of denticles prior to 19 myo to only 3% of all denticles after 19 myo. Most sharks today have linear denticles, and geometric denticles are only present in small deep sea squaliform sharks like the lantern shark Etmopterus and the cookiecutter shark Isistius

The pattern of extinction and denticle accumulation and abundance through time from Sibert and Rubin (2021A): on the left is a chart showing stratigraphic occurrence of different denticle morphotypes, the grey bar indicating the 19 myo interval; note how many denticle morphotypes disappear at this horizon. Also, note the difference in accumulation rate before and after 19 my all the way on the right hand side.


    As a result of these findings, Sibert and Rubin interpreted this to indicate an extinction that selectively affected deep sea or pelagic sharks. These authors do point out that there is a considerable hiatus in sedimentation at many deep sea drilling sites, and few core samples have a continuous record through the early Miocene. Another issue, explored below, is that extant sharks exhibit a multitude of denticle types even within a single individual shark, and it would be a pretty monumental task to survey all denticle types of all extant shark species. Sibert and Rubin considered most modern shark species to only exhibit 1 or 2 denticle morphotypes, and they surveyed 152 species – which is a daunting task, no doubt, but a small proportion of extant species. It is conceivable, even if every family is represented in their extant denticle catalog, that some morphotypes might be missed altogether given that there are some 1,000+ species of elasmobranchs and about 27% of them were studied (27% according to Feichtinger et al.).

Challenges to the Early Miocene Extinction Hypothesis: rebuttals by Feichtinger et al. & Naylor et al.

    Within a few months of publication, two rebuttals by two different teams of ichthyologists and paleoichthyologists were published – the first is largely a paleontological response by Iris Feichtinger et al. with many well-known paleoichthyologists, including Sylvain Adnet, Gilles Cuny, Jurgen Kriwet, Juergen Pollerspock, Kenshu Shimada, Charlie Underwood, Romain Vullo, and others. The second rebuttal is by Gavin Naylor and colleagues (including famous collector and shark expert Gordon Hubbell) – Naylor, by the way, worked here at CofC before leaving for greener pastures a few years ago – and this study is largely ichthyological in nature.

Different denticle shapes in the lantern shark Etmopterus - from Feichtinger et al. (2021).


Different denticle shapes in the daggernose shark Isogomphodon - from Naylor et al. (2021).


    One major concern issued by both studies, and accompanied by some pretty amazing figures, is the alleged underestimate of denticle types by Sibert and Rubin – each of the two rebuttals notes that many denticle types can be present in a single individual shark; Feichtinger et al. show the denticles of a lantern shark, Etmopterus, and Naylor et al. show regional differences in the denticles of a daggernose shark, Isogomphodon. Each figure is intended to showcase differences in denticle morphology across the body of one individual shark – but no attempt was made by either set of authors to try to assign denticles from their individual sharks to the morphotypes of Sibert and Rubin. There are some differences in the denticles of Feichtinger’s Etmopterus, but I can imagine Sibert and Rubin may have considered these to be lumped into 2-3 ‘geometric’ morphotypes. However, this is admittedly much less possible with the Isogomphodon, which, from my mostly ignorant perspective, would have at least 5-6 ‘linear’ morphotypes. Regardless, had I been a reviewer of either rebuttal, I would have expected Feichtinger et al. or Naylor et al. to replicate the methods and assign denticles of a single individual modern shark to the morphotypes; neither set of authors can hardly claim the method is opaque, as Sibert and Rubin have a very detailed appendix with example denticles for each morphotype well-illustrated. Feichtinger et al. claim that nearly all of the denticles that disappear around 19 myo could have belonged to a single shark species, and just a single species going extinct could produce the pattern of denticle extinction. Naylor et al. make a similar argument, that the regional extirpation of just a few species of deep sea sharks could have produced the same pattern.

Reanalysis of core sample denticle records by Feichtinger et al. (2021).


    Feichtinger et al. also make the point that the purported extinction event corresponds to a major discontinuity in sedimentation rate – with a major increase in sedimentation rate beginning around 19 myo, and likely explaining the lesser abundance of fossils. They also point out a decrease in the abundance of bony fish fossils during this interval, which Sibert and Rubin did not dwell on. Feichtinger et al. argue that the much lower sample size after 19 myo in the samples resulted in the difference in the shark denticle:fish teeth ratio and is best explained as a sampling artifact. When focusing only on samples that could reasonably be expected to include at least one denticle – those having at least 6 fish teeth – only a few samples are informative.

    Naylor et al. also make a claim which I find to be in ignorance of the fossil record – it’s a fair claim to be made by a biologist, but one not founded in paleontological familiarity. They argue that the lack of teeth – and dominance of denticles in the core samples – means the samples are from a “biased depositional environment-one that evidently favors the preservation of bony fish teeth and shark denticles, but strangely, not shark’s teeth.” This is a bit of a problem, as teeth are far more diagnostic than denticles. I have no data on the relative abundance of shark teeth and denticles, but even in tooth-rich stratigraphically thin (and therefore time-rich) bonebeds, shark teeth are rarely ever more common than a few miniscule teeth per kilogram of sediment. The drills used for Sibert & Rubin’s deep sea sampling are approximately 6 cm in diameter (~4.5 inches), but hitting a shark tooth with a drill in deep sea sediments would be like finding the proverbial needle in the haystack. Sharks famously produce and shed thousands of teeth in their lifespan – 10,000 is a commonly found number on the web – but have many thousands of denticles for each shark. The denticles are shed just like teeth (and our skin) and must be several orders of magnitude more common in deep sea sediments as a result. And of course, let us return to the sample size of the original Sibert and Rubin study: a total of 1263 denticles from two different core samples. Even if we assume that sharks ONLY have 1,000 denticles for every tooth, which I think is probably still being extremely generous, we’d expect a single tooth to have been found for this sample*. In sum, Naylor et al. either made this argument in bad faith (e.g. the tooth:denticle ratio in sharks) or are just not aware of the realities of the shark fossil record: from a tooth/sediment volume perspective, shark teeth are extremely rare fossils even when they are concentrated into tooth-rich bonebeds. Vertebrate macrofossils are rarely encountered in core samples, to the point that micropaleontologists usually treat them as oddities; I can think of only two mammal fossils, for example, ever found in core samples – one was a pinniped tooth in a Holocene-Pleistocene core sample from Antarctica someone emailed me about around a decade ago, and another is a ?Paleocene mammal in a core sample from Louisiana I recently learned about on Twitter.

*Sibert and Rubin responded to this (see below) and I wrote this prior to re-reading their response! Glad to know I was right on the money.

    Returning to shark teeth – Naylor et al. point out that no extinctions are known in fossil shark assemblages during the early Miocene, and that all known shark families persist through the 19 my interval. They do allow however that undocumented pelagic shark taxa might be affected by this event – the implication being that they may not be expected to be found in continental shelf rocks. Truth be told, most continental shelf marine deposits were deposited at shallow or mid shelf depths over the past 30 million years, and contain sharks that are largely shallow marine – many of which still live on our shelves. Where do you need to go to even find deep-sea shark fossils? I cannot think of any on the east coast, aside from a handful of cookiecutter shark teeth (Isistius) from a single creek in Florida – and on the west coast, there is only a few Neogene deep sea shark assemblages, like the Skooner Gulch Formation in Mendocino County. Others include mid-Cenozoic assemblages in the Pacific Northwest like the Makah and Pysht formations of Washington, more famous for their early marine mammal fossils. One critical issue is that in these deep-sea deposits, shark teeth are only rarely if ever found at the outcrop, owing to rapid and continuous sedimentation rates. Many are found during the preparation of other marine vertebrates – I recently found something like an Etmopterus tooth while acid-prepping an early cetacean skull from the Lincoln Creek Formation.

    There are no shortage of deep marine shark assemblages from the Mesozoic, BUT this is because these rocks are older and have had plenty of time to be uplifted: the same cannot really be said for Miocene assemblages even on active margins on the Pacific coast, where deep sea deposits are frequently still buried. On a passive continental margin like the Atlantic coast, the continental shelf has never been deeply immersed enough for deep sea sharks to shed their denticles and teeth into sediments that we can collect from. This in a way is a backwards ‘pull of the recent’ – our terrestrial fossil record gets better and better the closer we get to the Pleistocene and Holocene, but our marine record gets terrible. Why is that? It’s because most of the marine rocks are still under the ocean, aside from only the shallowest deposits. I’ve made this argument before in comparing the marine mammal record of ‘lagerstatten’ with marine reptiles – there just aren’t that many deep marine deposits of laminated, black shales in comparison with the Mesozoic because these deposits are probably still under the ocean or way way out on the continental shelf (aside from the greater oxygenation of Cenozoic seas and lack of stratified, anoxic seas).

Sibert and Rubin respond to criticism from Feichtinger et al. and Naylor et al.

    One of the most critical points brought up by Feichtinger et al. is the major change in sedimentation rate at approximately the same horizon as the inferred extinction. However, Sibert and Rubin point out that their accumulation rate normalizes/accounts for changes in sedimentation rate. In other words, the analysis should be robust to changes in sedimentation rate. The low sample sizes are perhaps still a problem, but accumulation rate is pretty widely used in micropaleontology. However, I’m not familiar enough with the method to comment further – though perhaps the same can be said for Feichtinger et al. As for the fish tooth to denticle ratio, Sibert and Rubin indicate that in well-sampled horizons, denticles are outnumbered by fish teeth by 1:37 to as much as 1:70, as opposed to a ratio of 1:5 before the purported extinction – suggestive of some biological phenomenon at minimum, and perhaps indicating that the rarity of denticles is not necessarily a sampling artifact. Rarefaction analyses by Sibert and Rubin (2021B) indicate that further sampling would not have changed the pattern of denticle extinction.

Rarefaction analyses indicating that additional sampling of extant species would not likely reveal additional extant denticle morphotypes, from Sibert and Rubin (2021B). Caption left in for full context since I'd probably screw something up!

    What about the concern addressed by both Naylor et al. and Feichtinger et al. that Sibert and Rubin underestimated denticle variation in sharks, and that the disappearance of just a couple of species (through a relatively minor extinction or extirpation) could explain the pattern of denticles without necessitating some ocean-wide pelagic extinction event. Sibert and Rubin admit that their catalog is not exhaustive, but assert that it is good enough: they also performed a second rarefaction analysis (Sibert and Rubin 2021C) indicating that additional sampling of modern denticles would not have resulted in the recognition of many more denticle morphotypes. This is to a degree based on some simulations, and I do not have the background to judge this. For the uninitiated – rarefaction is a type of analysis used in ecology to figure out how many species can be expected to be encountered based on the sample size, given that small sample sizes/limited number of observations will only produce a small number of species, and the more observations are made, the more species will be recorded. At some point, the number of observations will result in diminishing returns and the number of observations should be “good enough”, so to speak – which Sibert and Rubin (2021C) assert their analysis supports. I am skeptical, to be honest, given the very different assessment of denticle morphotypes present in a single shark species by Sibert and Rubin (2021A; 1-2 morphotypes per species) versus Feichtinger et al. and Naylor et al. (not quantified, but to my estimation at least 3-5 morphotypes in their figured examples, conservatively speaking). With an estimated 27% of modern species sampled, there is a LOT of territory left; many shark species sampled by Sibert and Rubin include only a few skin samples, and based on the Naylor et al. image of Isogomphodon, perhaps not enough anatomical regions of sharks were sampled by Sibert and Rubin. If the initial catalog underestimates morphotype diversity, might that not mean that the rarefaction analyses and simulations may be inadequate? Again, I don’t know enough of the ins and outs of these analyses to comment further.

Rarefaction analysis and simulated relationship between denticle morphotype extinction and genus/species extinction, from Sibert and Rubin (2021C). I left the caption in for the benefit of the reader, given how complicated these sorts of analyses are.


    In another analysis, Sibert and Rubin (2021C) simulated the relationship between denticle morphotype extinction rate and species extinction rate, and found that for 70% of morphotypes to go extinct, perhaps 90% of species/genera might have to go extinct (recall that 70% is the observed rate of extinction of denticle morphotypes about 19 myo in Sibert and Rubin’s (2021A) analyses. However, if I understand this correctly, this is all based on the integrity and exhaustiveness of the original catalog.

    Sibert and Rubin point out* that denticles outnumber teeth in an individual shark by three to five orders of magnitude – in other words, denticles are 1,000 to 100,000 times more common. Sibert and Rubin similarly point out (as I did above) that out of a sample under 2,000 denticles, only 1-2 teeth might be expected to be found.

*I wrote the above section where I criticized the Naylor et al. comments about not finding teeth before I read the response by Sibert and Rubin and am pretty chuffed that I was right on the money.

Is the extinction real? Some corollary evidence from marine mammal paleontology – or coincidences?

    Both of the replies by Feichtinger et al. and Naylor et al. assert, perhaps correctly, that there are no known extinctions of sharks that we know of in marine fossil assemblages found in continental shelf sediments (e.g. on land). I’d wager that, based on what I know of Neogene shark paleontology, there’s quite a bit of fossils out there that have not yet actually been published so there is a helluva ‘black box’ remaining to be opened up. There are some differences between well-known Oligocene shark assemblages and those from younger Miocene rocks, but the timing of transitions and/or extinctions remains poorly established, with notable exceptions like excellent studies of Oligo-Miocene change in the snaggletooth shark Hemipristis (Chandler et al. 2006) and the loss of cusplets in Carcharocles (Perez et al., 2019). Lastly, don’t even get me started on Oligocene sharks: very, very few Oligocene assemblages are published so any claim that they’re well-studied is laughable.

    Well, that’s enough about sharks: what about marine mammals? A couple major things happen during the early Miocene right around the time that the proposed extinction takes place. The first, and best established, is the widespread distribution of phocid seals throughout the Mediterranean and North Atlantic. Many of the oldest phocid seals are Burdigalian in age and Aquitanian specimens are rare or have dubious/poorly established ages. In the Pacific, pinnipeds diversify in the Burdigalian, with a sudden appearance of the first otariid fur seals like Eotaria and the first imagotariine walruses, most of the earliest desmatophocid seals, and some of the last-surviving enaliarctine ‘proto-seals’. The second event are changes in odontocete fossil assemblages, with the first common occurrences of early delphinoids like the typically small-bodied porpoise-like kentriodontids, which about ten million years later would give rise to the first belugas, true porpoises, and oceanic dolphins. In addition, sperm whales become common and arguably more diverse in the Burdigalian – including many large macrophagous species - though it is unclear how sudden or ‘blurry’ this was. The third, and perhaps most important and enigmatic, is the global rarity or outright absence of baleen whales in any earliest Miocene (23-20 myo, Aquitanian stage) deposits as documented by Marx et al. (2019). There are tons of toothed mysticetes, eomysticetids, and Mauicetus-grade whales (NZ only) in late Oligocene strata of the Pacific Northwest, Japan, Australia, New Zealand, and here in South Carolina – and pretty much only dolphins in Aquitanian-correlative deposits in the Pacific Northwest, Chesapeake Group in Maryland/Virginia, and the well-sampled Belluno Sandstones of Italy. Then, in the Burdigalian, the mysticete fossil record picks back up with a record dominated by numerous “kelloggitheres” – a moniker for Parietobalaena-grade early crown mysticetes which Remington Kellogg, the father of paleocetology, devoted no small time to studying as skeleton after skeleton kept on being discovered at Calvert Cliffs in Maryland. Marx et al. (2019) suggested that perhaps chaeomysticetes (toothless mysticetes) were predominantly offshore/pelagic during the Aquitanian, perhaps explaining their rarity in continental shelf strata. During the Aquitanian, there was a drop in sea level worldwide as well as a substantial decrease in the abundance of diatoms.

    Changes in the pinniped fauna as well as the odontocete fauna correspond in a rough sense, but the sudden and widespread appearance of baleen whales after the Aquitanian I feel is perhaps more profound given the feeding ecology of baleen whales – they’re tied to prey low in the food chain (zooplankton). In my opinion, there’s plenty we do *not* know about the early Miocene, and there are just enough coincidences with faunal change in the marine mammal record to suggest that something happened in pelagic ecosystems.

Closing Thoughts

    Altogether, I think this was a pretty fascinating back and forth over deep sea denticles. I’m not fully convinced that an extinction occurred – but I’m not convinced that it’s not “nothing”, either, and there’s still a lot we don’t know about marine vertebrate faunal evolution across the Oligocene-Miocene transition. I think further sampling is needed – both of deep sea denticles and modern shark denticles – and these remain pretty serious issues for the Sibert and Rubin analyses. In my opinion, the problem of undersampled extant shark denticle morphotypes is the single biggest issue with the analysis, and I am less concerned with sedimentation rate v. accumulation rate, as that strikes me as something that micropaleontologists sorted out decades ago. On a parting note, I’m really impressed by the sorts of groundbreaking work by Elizabeth Sibert and colleagues and am eagerly awaiting her next scientific contribution – it’s a new way of studying the shark and fish fossil record and it’s already made ‘big waves’ (hah hah) in marine vertebrate paleontology, compared to the alternative way we’ve been studying assemblages (one new locality at a time). Who knows what further sampling of modern shark denticles will result in!


Boessenecker and Churchill, 2018.

Feichtinger et al., 2021.

Marx et al., 2019.

Naylor et al, 2021.

Sibert et al., 2014.

Sibert et al., 2016. Eighty-five million years of Pacific Ocean gyre ecosystem structure: long-term stability marked by punctuated change | Proceedings of the Royal Society B: Biological Sciences (

Sibert and Norris, 2015. New Age of Fishes initiated by the Cretaceous−Paleogene mass extinction | PNAS

Sibert and Rubin, 2021A.

Sibert and Rubin, 2021B – response to Feichtinger et al. 2021.

Sibert and Rubin, 2021C – response to Naylor et al.

Sibert et al., 2017.


Saturday, December 3, 2022

Bobby's guide to whale & dolphin earbones 1: introduction

I've realized over the past few years that part of my blogging slowdown can be blamed on having a job that leaves not much time for research, but also, I slowed down when I began posting on twitter - twitter became 'microblogging' in a way, and it definitely scratched the same itch that blogging did. However, I don't think it's quite as useful, and the prophesied "heat death" of twitter driven by Elon Musk's hostile (and laughably incompetent) takeover of twitter has made me think twice about blogging more often. I've cultivated a decent following on twitter - and it might not be going anywhere - but I shouldn't neglect the opportunity for longer format communication/commentary on marine vertebrate paleontology. That, and in the past couple years, finally got the courage (and knowledge base) to write a series of blog posts about a rather daunting subject: whale and dolphin earbones - those weird little bones people find while searching for shark teeth, and don't know how to make heads or tails from them. So - enjoy!

Read Part 2: identifying toothed whale/dolphin periotics here.


Fossil collectors here on the east coast typically start out looking for fossil shark teeth – which are, after all, often some of the easiest fossils to find and recognize. Eventually, however, beginner collectors begin to broaden their search image as they begin to recognize more than just shark teeth. Some of the most unusual specimens are earbones of whales and dolphins – but there is widespread confusion about what the bones are, how they function, and extremely little available to the collector (or, even students/new paleontologists) explaining what the heck these things are and how to identify them. I’ve thought about doing such a useful post for a long time, and I hope to cover some of the basics of anatomy, function, and perhaps most importantly, identification of isolated earbones. Identification will follow this introductory post. Paleocetologists generally don't do much science communication and rarely explain how these earbone things are identified - and many paleocetologists outright avoid studying earbones. Interpreting them is occasionally referred to as a "black art" by some whaleontologists, a sentiment I agreed with when I was still a master's student and just hadn't seen enough of these things. I just love them - the shapes are interesting, finding them is always satisfying, and they're quite informative. I hope that this series of posts will help alleviate much of the ambiguity about earbones and furthermore, help fossil collectors and other whaleontologists out there make some sense of these weird little doodads.

Four periotics and a nice bulla found on a dredge island in Charleston harbor this week - December 2022, after about four hours of looking. Thanks to Ashby Gale and Sarah Boessenecker for searching!

Nuts and Bolts I: Basics of Terrestrial Mammal Ears and Hearing

In order to discuss cetacean earbones, we need to lay the foundation of anatomy of a terrestrial mammal ear. The mammalian ear evolved during the Permian and Triassic from synapsid “mammal like reptiles” – bones from the jaw joint and back of the lower jaw were miniaturized and transitioned into different parts of modern mammalian ear. The mammalian ear is a remarkably complex apparatus cobbled together from these different parts, and the result is a bit of a “Rube Goldberg machine” – made of three major regions: the outer, middle, and inner ear. Hearing of course begins with the external ear itself, called the pinna, which funnels soundwaves into the external ear canal (external auditory meatus). These sound waves travel towards the tympanic membrane, which is a thin layer of skin that separates the middle ear cavity from the outer ear. The tympanic bulla is a thin bony shell connected to rest of the squamosal bone and houses the middle ear cavity inside.

Comparison of the middle ear and ossicles of a non-mammalian reptile (left) and a modern mammal with the three middle ear ossicles (malleus, incus, stapes) bridging the middle ear sinus. (from Tucker, 2017). Note that everything white on here is bone, or other tissues of the skull. Cetaceans differ chiefly in lacking a tympanic membrane and external ear canal, as well as having smaller semicircular canals.

The middle ear is an air-filled cavity – air gets in there through the eustachian tube, which connects to your nasal sinus. The eustachian tube is semi-permanently kinked like a water hose; sufficient pressure inside your nasal sinus or the middle ear will cause the tube to unkink – which is precisely what happens when your ears pop – allowing the pressure to equalize between the middle ear and the ambient air pressure outside your head. Air pressure can also build up if you have an ear infection or a bad cold – which can cause fluid to accumulate and without letting the eustachian tube unkink, worsening the problem. A ruptured eardrum allows air to pass from the outer to the middle ear, and apparently sounds windy. Know someone who had “tubes” in their ears as a kid? These are little tiny tubes stuck in a surgical perforation that allows excess pressure (and fluid… gross) to leave the middle ear (apparently the eustachian tube is narrower and more horizontal in kids causing more frequent ear infections). Because the tympanic membrane is basically skin, the tubes fall out on their own within a year (same reason that some ruptured eardrums will heal on their own).

What happens when a sound wave strikes the ear drum? The ear drum vibrates – and these vibrations are transmitted into the middle ear ossicles – the hammer, anvil, and stirrup. The hammer (malleus) is partially suspended in the membrane. The hammer transmits these vibrations by striking the anvil (incus), which is smaller in mass. The anvil strikes the stirrup (stapes) – which really does look like a stirrup. Each ossicle is smaller than the last, and each vibrates more strongly, but with less movement - long story short, this is a way to translate relatively weak sound waves from the air into sound waves strong enough to move through fluid in the cochlea (see below). This is called 'impedance matching'. Sound waves typically just bounce off of a fluid, so the middle ear ossicles bridge this gap.

A fractured baleen whale periotic I collected in August 2022 from Ridgeville, SC, which shows a great cross-section of the cochlea. Yellow star denotes the fenestra ovalis where the stapes articulates; the blue star denotes the base of the first turn of the cochlea; and the red star denotes the apex of the final turn of the cochlea.

The stirrup fits into a hole in the ‘petrous portion’ of the squamosal – petrous meaning stone like. This is the densest part of the skull, and it houses the fluid-filled inner ear cavities – consisting of the vestibule (opening) in the middle, the semicircular canals (responsible for balance) on one side, and the cochlea on the other. The stapes pushes in and out of the vestibular window, pushing the cochlear fluid – and making little waves. These waves travel up the cochlea and stimulate hairs that detect different frequencies. The coiled cochlea is unique to “Therian” mammals – marsupials and placentals – and is implicated in our greater range of frequencies we can hear versus most other vertebrate groups. Monotremes, as well as ‘reptiles’, have a straight or slightly curved cochlear duct that is much shorter, yielding a much lower maximum frequency.

Directional hearing is accomplished through acoustic impedance: the mismatch in timing between left and right ears. If a sound arrives at the right ear slightly earlier than the left, it’s probably closer to the right ear and therefore on the right side. Terrestrial mammal ears are adapted for hearing relatively slow moving soundwaves in air.

Location of the tympanic bulla, periotic, and cochlea in a modern dolphin. Image from T. Plencner's thesis, U. Otago. (see here).

Nuts and Bolts II: Cetacean Ears and Aquatic Hearing

There are a couple of major problems with mammalian hearing while underwater, and they relate to the different acoustic properties of water. The first, and most obvious, given that most readers are probably aware of it – is that sound travels about four times faster in water than in air. The second one is that sound waves tend to be transmitted through a medium of the same approximate density; when a soundwave hits a denser medium, or a much lighter one for that matter, most of the wave is not transmitted – most of the wave bounces. This is what an echo is. The density difference between air and flesh is enough to focus echoing soundwaves into the funnel-like external ear – but underwater, water, flesh, and even bone are close enough in density that soundwaves travel unimpeded without bouncing. This, combined with the greater velocity of soundwaves underwater, means that as far as the mammalian brain is concerned – the sound arrives at both ears simultaneously. This means if you’re submerged, and someone yells at you underwater – you won’t be able to tell the direction its coming from. So how do whales and dolphins hear the direction of sounds? They must – otherwise, echolocation would not work.

Cetaceans evolved a number of adaptations toward directional hearing. Early on, the different parts of the squamosal stopped fusing, and remained separate portions: the petrous portion became the petrosal, and the tympanic bulla… well, stayed the bulla, but lost its fused bony connection to the rest of the squamosal. These reduced bony connections are present in archaeocetes, and slightly reduce bone-conducted sounds reaching the cochlea. Dolphins (Odontoceti) continue reducing the bony connections, with less and less of the petrosal having any hard connection with the squamosal – aside from receiving a triangular prong unimaginatively called the “spiny process”, which seems to be the minimum connection, and is retained in all cetaceans. The connections thus become primarily made of soft tissue (which is why some dolphin skulls are rarely preserved with the earbones in articulation). Lastly, many modern cetaceans have evolved peribullary sinuses – connected to the pterygoid sinus system further forward in the head. These sinuses are filled with air – and remember, a difference in density causes echoes. These air sinuses cause incoming sounds to bounce away from the earbones – in concert with the reduced bony connections, resulting in earbones that are acoustically isolated from the skull. The extreme reduction in bony connections in toothed whales is what led to the name periotic, rather than petrosal.

I know the heading below says "basic" and this figure is anything but - there are tons of little features that cetologists, and paleocetologists in particular, obsess over. But don't worry, I break down the basics below! The periotic and tympanic bulla of a bottlenose dolphin, Tursiops truncatus. From Mead and Fordyce, 2009.

Basic Anatomy and Function of the Tympanic Bulla

The tympanic bulla is a unique bone in cetaceans – well, all mammals have a bony or cartilaginous bulla, but the medial part is thickened into a structure called the involucrum – this thickened lip is unique to cetaceans. What is also unique is that it is largely disconnected from the rest of the skull, as outlined above. At its most simple, the cetacean bulla is made of three major parts: the thickened involucrum, the shell-like outer lip, and the posterior process. The thickened involucrum is present in even the earliest cetaceans, and is actually what resulted in the recognition that the strange deer-like braincase of Pakicetus was an early whale. This massive swollen part is what is preserved 95% of the time: the outer lip, the shell-like part of the bulla, is the part least modified from a land mammal and owing to its fragility, is typically broken away. The involucrum is often cylindrical and typically wider posteriorly; it can be deep and transversely narrow in right whales, or wide and dorsoventrally shallow instead in many dolphins. The involucrum also typically sports some creases on its medial surface. The tympanic cavity – between the outer lip and the involucrum – is filled with air in life, as it is the middle ear sinus. The greater mass of the involucrum helps each bulla vibrate – and these vibrations are transmitted directly to the malleus, which is fused to the outer lip of the bulla. In this capacity, the bulla functions a lot like the eardrum (it is, however not the eardrum, which is of course a soft tissue structure, and cetaceans don’t even have an eardrum any more anyway; fossil collectors often call these “fossil whale eardrums”, which is not accurate).

The bulla connects to the periotic with two small easily broken ‘pedicles’, at least in Crown Mysticeti  – the anterior pedicle is located on the anterior part of the outer lip, and the posterior pedicle is positioned near the back of the involucrum. In cetaceans without periotic-bulla fusion, there is a tiny bone called the accessory ossicle that articulates in the position of the anterior pedicle, and the posterior process of the bulla is connected to the posterior pedicle. The posterior processes of the periotic and bulla are frequently similar in shape and variable in length among different cetacean lineages; the articular facets of each interlock and are mirror images in an individual cetacean. On the outer lip, a tongue-like knob of bone is present and called the sigmoid process – just next to it is the attachment point for the malleus (‘hammer’), the largest of the three middle ear ossicles.

Cross-section through an idealized/simplified auditory region of an ancestral archaeocete whale. From Luo and Gingerich, 1999. Caption included for abbreviations.

Basic Periotic Anatomy

We’ve already talked about the interior of the periotic, although it’s expensive and difficult to study as CT scanning is required – and not medical-grade scanners you would find in a hospital, but micro-CT scanners which are fewer and further between. Because of this, few cetaceans have more than one or a few individual cochleae scanned, and we’re not too sure about individual variation of the bony labyrinth. That doesn’t matter, to be honest, because virtually all of the information we use to reconstruct evolutionary relationships (phylogenetic analysis) and to identify isolated periotics or diagnose new species of cetaceans are external features of the periotic.

General features of the bulla and periotic of a dolphin - from Marx et al. (2016).

Periotics can basically be divided into four main parts: the body of the periotic is the densest, most solid part, and is just kind of a lump of bone – the body articulates with the squamosal in the base of the skull. The pars cochlearis is a hemispherical prominence that houses the cochlea – hence the name, meaning the ‘cochlear portion’ in latin. The anterior process is a knob to triangular or sheetlike bit of bone that projects anteriorly beyond the pars cochlearis from the body, and the posterior process in the opposite direction. The anterior process is primitively long in odontocetes and becomes shorter and stubbier, and the opposite occurs in baleen whales: it begins somewhat short and stubby and becomes longer in some mysticetes.  

General features of the bulla and periotic of a baleen whale - from Marx et al. (2016).

The body of the periotic is of variable thickness and shape but often ovoid to spherical, and can be smooth or rugose. The pars cochlearis is typically smaller and hemispherical in shape amongst Neoceti – but in archaeocetes, is frequently rectangular. The pars cochlearis is perforated by several holes (foramina and fenestrae), including two apertures for endolymph and perilymph which fill the two different divisions of the bony labyrinth – and are also called the apertures for the cochlear aqueduct and vestibular aqueduct. The fenestra rotunda (round window) is present posteroventrally on the pars cochlearis and has a little membrane that can bow out to accommodate fluid displaced internally when the stapes is pushed into the oval window (fenestra ovalis), which is positioned at the junction of the pars cochlearis and body. Just lateral to this is the ventral opening for the facial canal, which transmits the facial nerve (Cranial nerve #7, out of 12). Dorsally, there is a large hole – sometimes funnel-shaped, sometimes a pair of holes – which include the dorsal opening for the facial nerve canal, and the dorsal opening of the vestibulocochlear nerve (Cranial nerve #8). The ‘endocranial’ openings for cranial nerves 7 and 8 are separated by a small crest called the transverse crest or crista transversa – if the crest is very high, the foramina appear separated. If not, the foramina are confluent within a structure called the ‘internal acoustic meatus’. Lateral to the meatus, in archaic baleen whales and dolphins, as well as archaeocetes, there is a large fossa formed on the dorsal side of the body called the suprameatal fossa – this is bordered laterally by a tall ridge that abuts the squamosal, called the superior ridge. This ridge is present only in archaeocetes and the most primitive dolphins and toothed baleen whales, and is rapidly lost independently in both groups. Ventrally, lateral to the ventral opening for the facial nerve canal, there is a pit for the malleus (mallear fossa) and a laterally directed tuberosity – called the lateral tuberosity, typically lateral to the mallear fossa.

Basic parts of the periotic of an odontocete and a mysticete, color coded - modified from the beautiful illustrations from Kellogg (1931, 1966).

The anterior process is a bit of a simple structure and varies somewhat in shape and whether or not it bears a sulcus medially and laterally for nerves/arteries or a furrow for the tensor tympani muscle (same muscle you use to pop your ears, and if you can wiggle your ears, you often accidentally tense it; it pulls a little bit on the malleus and briefly disengages the ‘ossicular chain’, e.g. malleus/hammer, incus/anvil, stapes/stirrup – so you hear a slight difference in sound). The anterior process has a facet for the bulla ventrally, called the anterior bullar facet; this facet is lost in many baleen whales and some dolphins, and is a bit of a primitive feature. The gap between the anterior process and the pars cochlearis is called the anterior incisure – and the angle at which the pars cochlearis and anterior process join is diagnostic, with a 90 degree angle in archaeocetes, toothed mysticetes and odontocetes, an acute angle in mostly modern-group (crown group) odontocetes, and an obtuse angle in most toothless mysticetes.

The posterior process joins the posterior process of the tympanic bulla in archaeocetes, early mysticetes, and odontocetes, and in most toothless mysticetes, the posterior process of the periotic and bulla are fused together and form a compound posterior process. In whales and dolphins with unfused posterior processes, the periotic has a posterior bullar facet for the posterior process of the bulla; this facet can be convex (typically in early baleen whales and squalodontid dolphins), flat, or concave (typically in dolphins) and smooth or bear ridges and grooves; the facet is typically constructed from very dense bone. The posterior process is typically sandwiched between the squamosal and exoccipital bones. Primitively, in archaeocete whales, the posterior process reaches the lateral side of the skull but in Neoceti – the clade formed by odontocetes and mysticetes – the posterior process is shortened and obscured from view by the bulla and the squamosal bone.

Some of the major differences in the earbones of a dolphin (top) and baleen whale (bottom). From Yamato and Pyenson, 2015.

Major Differences Between Archaeocete, Odontocete, and Mysticete Periotics

We will mostly sidestep archaeocete periotics, with a few exceptions in order to illustrate the common shape that odontocete and mysticete periotics evolved from. Basilosaurid whales have large, rectangular pars cochlearis, a short anterior process, a very tall superior ridge, and a broad, deep suprameatal fossa. Your odds of finding an archaeocete periotic are exceedingly rare, and in the USA, would basically require you to be prospecting in Eocene quarries of the Carolinas or perhaps Alabama or Mississippi. I’ve only seen about a half dozen isolated archaeocete periotics discovered, and they are more frequently found in articulation with a skull – owing to the superior ridge, and broad contact with the squamosal – they rarely become disarticulated.

The biggest difference between odontocete and mysticete periotics is size – baleen whales have big periotics, and dolphins (and all odontocetes, to be honest) have small periotics. There are some fossil dwarf mysticetes out there with small earbones, and many toothed mysticetes are quite small, but the former are still much larger than most extinct odontocete periotics and are similar in size to the largest periotics of modern odontocetes (e.g. sperm whale, killer whale) and the latter are very, very rare and you probably will not ever find one (unless you collect at some choice localities in the Pacific Northwest or Australia). Most fossil odontocetes are still nowhere near as large as orcas or sperm whales. If you find something that is HUGE but looks like a dolphin, you’ve found something interesting! There are some giant dolphins from the rock record, represented only by giant, orca-sized periotics not found in association with a skull (from both coasts of North America). Virtually all fossil odontocete periotics are under 4 cm in length when complete, and even the smallest toothless baleen whales have a periotic (not counting the posterior process) that is usually larger. That being said, the size of the cochlea and the pars cochlearis does correlate strongly with body size (and within some groups of fossil whales and dolphins, the pars cochlearis diameter can be used to estimate body size).

Most odontocete periotics have a proportionally large pars cochlearis and a relatively small, narrow body, and a roughly cylindrical anterior process and a short posterior process; when the pars cochlearis is broken away, the rest of the bone is somewhat sausage-shaped. 

In toothless mysticetes, except for the Eomysticetidae, the posterior processes are fused into a compound process – this can be plug-shaped or long and strap-shaped. In most toothless baleen whales other than right whales, there is a very narrow ‘neck’ of the compound process and as a result it frequently breaks off. Balaenopterid whale posterior processes are long and curved, and in large individuals, about the size and shape of a banana. In gigantic Pliocene and modern mysticetes, the posterior process can be over 30 cm long! Baleen whale periotics are also very rugose and tend to have rough, porous, knobby, and spiky surfaces. Such surfaces are common only in the most gigantic of modern odontocetes and quite rare even in large extinct odontocetes, though some limited rugosity can be seen in the very tips of the anterior and posterior process in very ‘elderly’ odontocetes.

Other features that vary between families are given variable names – you can think of these as topographically positive and negative features. Tuberosites, processes, and tubercles are anything that is mound-shaped or higher. A ridge is precisely that – a long, linear feature with a crest, though crest is sometimes used instead. Anything that is mound-like is usually called a prominence or a tuberosity; anything longer, or consistently present in most species, is called a process. Furrows, fissures, and grooves are all excavated linear features – fissures usually refer to a gap that is closing during bone growth but still remains as a gap between bony parts that fails to fill in; furrows and grooves are somewhat interchangeable and are shallower and typically wider than fissures. Sulci are similar but have smooth walls and clearly form from ossification around a nerve or artery; canals are similar but fully enclosed and refer to holes that transmit one of the cranial nerves or a derivative thereof (e.g. facial nerve canal). Foramina are holes transmitting smaller nerves or arteries of known or unknown identity; they are frequently more variable than cranial nerve canals.

We also use shape terms you may or may not be familiar with, and I’ll try to use only terms that are likely to be self-evident – cylindrical, flattened, hemispherical, rectangular, triangular, etc. I actually prefer such simplicity in my actual scientific writing as well; my Ph.D. adviser was a big fan of single-use words like ‘cordate’ or ‘cordiform’ instead of ‘heart-shaped’, but I find the latter to be more inclusive even if it uses two words instead of one. This blog post is already 13 pages and growing and therefore already guilty of indulgence, so why stop there?

How to Find Fossil Periotics (and Bullae)

There’s not much of a trick to finding fossil periotics aside from 1) knowing what they look like and 2) looking in the right places. Periotics are rare – on the east coast, you will find hundreds of shark teeth for each periotic you find – and they are heavy. You best look for bone-rich layers or beaches with strandlines full of periotic-sized pebbles. This is because water does not care what a fossil belongs to –waves and currents will concentrate objects of similar density and size into discrete bands along shorelines – this is a concept called ‘hydraulic equivalence’. A slightly denser, but smaller, object will be concentrated with slightly larger, less dense objects – to a point. Shells will often be in a different place along a beach, because they are lighter and have a large surface area; you’ll often find small teeth mixed in. Here in the Charleston region, I would look for phosphatic pebble rich strandlines along our waterways and then try and find the ‘diamond in the rough’ by looking for periotic-shaped objects. Even for an expert this can be quite mentally exhausting and I’ve gotten many a headache on coastal field trips here in Charleston trying to spot a periotic amongst a sea of one hundred thousand phosphatic lumps of the same exact color. [That’s usually my cue to sit down for a while and drink some Gatorade].

A nice periotic of Eoplatanista, sitting there in the most easily recognizable orientation (for me, anyway). 

An even nicer periotic of a possible waipatiid or eurhinodelphinid laying dorsal up, but not amongst a sea of gravel.

Looking for periotics or tympanic bullae is no different from looking for shark teeth – it’s just much, much less intuitive given the strange shapes that earbones take on. We joke around about looking for “black triangles” and many folks don’t know how to find shark teeth – but they’re quite intuitive in my opinion, and if they are there on a beach, they’re unmistakable. If you can’t find shark teeth, perhaps stop reading now and come back to this after a year or two of practice: finding earbones is hard work and a bit more of an advanced collecting ‘target’. But, once you’ve mastered the search for shark teeth – and you can readily identify a complete, partially exposed, or fragment of a tooth on a beach or in a cliff – you’ve done so because you can recognize patterns and have memorized the shape and surface texture of many different shark teeth. Finding earbones is no different: the shapes are just… weirder. But once you’re familiar with them, you develop a search image. It’s also good practice to just pick up anything 1) you’re absolutely sure is not a pebble and 2) anything you are pretty confident is a fossil, but cannot immediately identify. This strategy will let you start finding some of the more uncommon odds and ends like prismatic cartilage fragments, large ray denticles, bird bones, and osteoderms (skin bones/armor) of armadillos, turtles, gators, glyptodonts, and more – as well as earbones.

A very difficult to spot dolphin periotic amongst a sea of similarly sized phosphate pebbles. When exploring phosphate spoils, it pays to slow down, get a foam kneepad and go a couple feet at a time. The periotic on the right is not the same specimen, but similar (and flipped over).

One last word of advice: most people have no idea what earbones are. This means that in Charleston, I have a decent shot at finding what I’m REALLY after, no matter where I go collecting or when – if folks are out there on a beach or riverbank before me, odds are, they are only searching for ‘black triangles’. On Thursday we visited a spot on a sand bar that has been open to boaters since mid October – and the three of us found four periotics (three of them totally complete) in a morning, despite hosting hundreds of shark tooth collectors each week for six weeks for the second winter in a row after harbor dredgings were dumped. There is always going to be a fossil collector who is more obsessed and competitive than you are, doing all sorts of crazy shit for the chance to score a bigger shark tooth – I like my mornings too much for that, and quite frankly, the extreme measures some people take begins to make me worry about their mental well-being. Out on one of our barrier islands, I heard of a couple guys who go out there at 4am (by boat, since it’s an island without a bridge/causeway) with spotlights and stay out there until the following low tide. Yeah, that’s a no from me. I’ll get up slightly earlier, drink a coffee or two, and once I’m feeling like I’m among the living again, then I’ll drive out to do some beach walking. I fully understand that someone will almost certainly find larger and better shark teeth than I will – that’s fine! What I’m looking for are more scientifically interesting than shark teeth, and nobody else really looks for them, so the chance is always there. If I find some nice shark teeth, great! But that’s not my goal. Loose, beach-cast shark teeth are fun to find, but are not scientifically significant.

A smattering of dolphin earbones, and one partial mysticete bulla, from various localities we prospected in winter 2019 through summer 2020 - many of these are pandemic finds, made after we were allowed to use boat ramps and go to publicly owned access points again in late April/early May when lockdown was beginning to ease up. Eoplatanista, waipatiids, eurhinodelphinids, squalodelphinids, and delphinids are all present, representing Oligocene, Miocene, and Pliocene dolphins. This also showcases variability in preservation - out of all of these specimens, only three are complete, and half are around 2/3 complete or so. Note how at least one of the the anterior or posterior processes, or the pars cochlearis, is busted off.

Preservation of Fossil Earbones

Tympanic bullae and periotics are dense bones, which means they are quite resistant to abrasion. Their higher density, however, means that they are more susceptible to fracturing than other bones are. In fact, tympanic bullae in particular often have conchoidal fracture – the same sort of spalled fracture you see in flint, obsidian, and glass. This means that earbones are quite susceptible to damage if dropped or packed improperly during fieldwork. If found in the side of a cliff, they are often fractured already and will readily break if not excavated carefully. At some fossil sites this means NO HAMMERING – I was instructed early on by one of my whaleontological mentors that you can carve away at the rock with a knife, but hammering will likely shatter a bulla; periotics are a bit more robust. The posterior process of the bulla is usually the first part of the bulla to be lost, and followed shortly thereafter by the outer lip, which is missing in most fossil bullae. In periotics, the anterior or posterior process frequently have chips missing; sometimes the pars cochlearis is fractured, permitting a view of the cochlea itself.

Because earbones are so dense and resistant to damage, they are often found concentrated in bonebeds where other bones from the rest of the skeleton are frequently highly damaged. In cases like this you can prospect bonebeds to get a sense of the diversity of an ancient marine mammal assemblage without having to find dozens of large, difficult to excavate skulls. Based on the Purisima Formation in coastal California, perhaps half of our whale and dolphin species are known only from earbones, and only the most common species are also represented by skulls. It's a bit different here in Charleston, as confirmed earbones collected isolated from the Ashley and Chandler Bridge formations are rare and harder to find than a skull or skeleton - so most of our earbones are actually found in association with more complete skeletons, but this is entirely because the exposures here are quite different. Bonebeds cannot be surveyed over long periods of time with many visits, as the exposures are ephemeral and only the largest fossils are likely to be found.

Notes on Geographic Scope

This series is focused on fossils from North America, and in particular, localities where loose periotics can be found. As a result, I’m going to gloss over groups known only from the Pacific Northwest, like the Simocetidae, which are generally only found by extremely non-casual collectors and only found in concretions; those periotics can only be freed from the rock through painstaking airscribing or acid preparation. Likewise, I am sadly going to be ignoring much of the rest of the world, but only “sort of” – Miocene cetaceans are generally pretty cosmopolitan, and aside from a few random weirdos out there known only from one locality, virtually all of the groups I’m summarizing are known from fossiliferous localities in Japan, NZ, Australia, South America, Europe, and South Africa. Basically, for Miocene and Pliocene cetaceans, you will find very similar periotics between the east coast and Europe as well as South Africa. Fossils from the west coast share quite a bit with Japan, Peru, and Chile. Oligocene fossils from the Carolinas are similar in some respects to those from New Zealand.

Comments on Cetacean Taxonomy and Periotics

This simple review is going to deal with identifying specimens down to the family level – many readers are likely wondering what that is. Families are one of the three lowest fundamental taxonomic units – species and genera being the lowermost units that fit within families. You can think of a family as a closely related ‘collection’ of genera. These historically have been assigned ranks under traditional Linnean Taxonomy – which is old fashioned and inflexible, and we no longer rely on this method for classifying species. We now use cladistics to reconstruct the tree of life. However, genera and species are still critical taxonomic units, and though some cladists reading this might roll their eyes, so are families – and for the purpose of this writeup, families are actually pretty useful: inclusive enough to lump a bunch of species with similar periotics together so as not to overwhelm the reader, and most of these groups have some cladistic support. There are some groups that were traditionally called families but do not appear to be closely related – such as the “Kelloggitheres”, or Cetotheriidae sensu lato – all of these typically middle Miocene baleen whales Remington Kellogg named in the early to mid 20th century and then kicked out of the Cetotheriidae over the past 15 years. When it comes to their periotics, even though the “Kelloggitheres” are a ‘paraphyletic’ grouping – their earbones are all very similar and will constitute a single entry.

Family tree (cladogram) of whale and dolphin species. From Geisler et al. (2011).

More critically, there are some larger group names I am going to use that you will need to be familiar with, and some taxonomic terms we’ve adopted from advances in cladistics. First, crown v. stem – the crown group is the most inclusive group of species represented by modern lineages; Crown Cetacea therefore is the group including modern odontocetes and mysticetes, but excluding Stem Cetacea – species that branched off before the Crown Cetacea split. Crown Cetacea is typically called Neoceti – and stem cetaceans are the “Archaeocetes” - ancient cetaceans that showcase the land to sea transition (pakicetids, ambulocetids, remingtonocetids, protocetids, and basilosaurids). Crown Mysticeti is the group including right whales, gray whales, balaenopterids, and pygmy right whales (all modern) but excluding eomysticetids and all toothed mysticetes – collectively, these are the stem mysticetes. Toothed mysticetes exclude the eomysticetids. Crown odontocetes include sperm whales, beaked whales, “river dolphins” (four families), delphinoids (porpoises, oceanic dolphins, and belugas), and a couple of extinct groups such the kentriodontids, allodelphinids, eurhinodelphinids, eoplatanistids, and squalodelphinids. Stem odontocetes are mostly known from the Oligocene and early Miocene and mostly include the xenorophids, agorophiids, waipatiids, and squalodontids. Aside from these, there are a few additional names I might use elsewhere in this series of posts: Chaeomysticeti refers to all toothless mysticetes – the Eomysticetidae + Crown Mysticeti group. Delphinida refers to most modern dolphins and extinct allies that are more closely related to the Delphinoidea than they are to either the Ganges river dolphin Platanista, sperm whales, or beaked whales. Delphinoidea all have similar periotics, and include at minimum belugas/narwhal (Monodontidae), true porpoises (Phocoenidae), and the oceanic dolphins (Delphinidae) and also the extinct Albireonidae and likely the Kentriodontidae – perhaps the common ancestors of all delphinoids. Members of the Delphinida that lie outside the Delphinoidea include the Lipotidae (Yangtze River Dolphin and the extinct Parapontoporia) as well as the South American river dolphins, including modern and fossil relatives of the Amazon River dolphin Inia (Iniidae) and the La Plata river dolphin or Franciscana (Pontoporia and Pontoporiidae). A more controversial clade is the Platanistoidea – formerly a wastebasket for all weird crown odontocetes with earbones and snouts similar to the Ganges river dolphin as well as all other river dolphin groups, now a more restricted group certainly including the Platanistidae, Squalodelphinidae, and Allodelphinidae. Squalodontidae, Waipatiidae, and some others are groups formerly assigned to the Platanistoidea.

Periotic bones and to a lesser degree, tympanic bullae, are useful for defining (‘diagnosing’) extinct cetacean species. Within Neoceti, the periotic is often loosely connected to the skull and as a result it is not only frequently found isolated, but also skulls are often found lacking periotics. It is preferable to name species from fossil material preserving periotics, whenever possible. Standards for the quality of type specimens – name-bearing specimens that genera and species are founded upon – have never been set and recommendations rarely made, but Larry Barnes wrote in the 1970s that type specimens should not be founded upon isolated earbones or postcrania, and should include skull material and hopefully earbones. Archaeocete taxonomy is a different beast altogether – many archaeocete whale holotypes are isolated vertebrae or strings of vertebrae, and few consist of skulls. Even more troubling is the emphasis of postcranial features and the apparent timidness of archaeocete paleontologists to describe skull features in detail. So, even some basilosaurid whales with well-preserved skulls have skull descriptions that are shorter than the description of the thoracic (ribcage) vertebrae and often not enough figures of the skull, teeth, and earbones.

One of Remington Kellogg's holotype periotics from the middle Miocene Sharktooth Hill bonebed - the holotype of Grypolithax obscura, recombined in the 1980s as Kentriodon obscurus owing to similarities with Kentriodon pernix and the discovery of a Kentriodon pernix-like skull in the Sharktooth Hill bonebed. Much effort has been required to chase down and establish genus-level synonymies for these problematic taxa.

Not having skulls with associated periotics is admittedly frustrating. Remington Kellogg published a paper in the 1930s which established several new genera and species of odontocetes based on isolated periotics from the middle Miocene Sharktooth Hill Bonebed. He did this because the periotic assemblage revealed more diversity than skull material – and also, at the time, there wasn’t really much skull material preserved at all! Odontocete skulls are relatively fragile, and rarely found intact within the bonebed. Most of these are likely all kentriodontids, some are synonymous with each other, and at least one is the allodephinid Zarhinocetus errabundus – originally named Squalodon errabundus by Kellogg. It took decades of collecting before more complete specimens preserved skulls in association with earbones (e.g. Zarhinocetus errabundus), or before some periotic-species could be assigned to species known from more complete material elsewhere, such as Kentriodon obscurus. Kellogg *almost* started a split off field of ‘parataxonomy’. ‘Parataxonomy’ is a situation that arises from the naming of fossils that may never correspond to the main field of body fossils, which in this case, would be species named from skulls. Ootaxonomy – the naming of fossil eggs – and Ichnotaxonomy – the naming of trace fossils – are two such examples of well-established fields of parataxonomy. We want to avoid naming new species from earbones, mostly because that establishes two different systems or sets of names – though there is always the possibility of discovering skulls with associated periotics to establish which names are synonymous. The proliferation of names of fossil sharks based on different tooth positions is a prime example – the species Cretoxyrhina mantelli is the ultimate senior synonym of over 100 named species of fossil mackerel sharks.

The Significance of Whale and Dolphin Earbones: Why do we Care so Much?

Periotics, and to a much lesser degree, tympanic bullae, of cetaceans preserve quite a bit of information. We do not have a great handle on identifying isolated land mammal earbones - and that's probably because we don't have to: land mammals are generally identifiable and chiefly diagnosed based on their teeth. In most Neoceti, the teeth are either homodont (Crown Odontoceti) or nonexistent and replaced by baleen (Chaeomysticeti) - either way, teeth are not reliable for identifying and diagnosing species throughout the bulk of Cetacea, with the exception of archaeocetes and some early odontocetes and mysticetes. Mammalian paleontology famously relies, perhaps too much, on tooth anatomy - and nonmammal paleontologists joke that mammalian paleontologists only care about teeth, and that Mesozoic mammal evolution could be summarized as what tooth begat what other tooth.

The amazing diversity in earbone shapes out there not only helps us identify and diagnose species, but also contributes to our understanding of evolutionary history and phylogeny. We can get a rough approximation of the significance of earbones by looking at the number of anatomical characters used in cladistic analysis. For the beginner - cladistic analyses assign different anatomical states to a particular feature, and species with more shared derived states (states other than 0 - the primitive state) are grouped together on a phylogenetic tree (or cladogram). For example, all cetaceans with fused posterior processes of the tympanic bulla are grouped into the Crown Mysticeti (roughly; there's a couple of mysticetes from NZ that challenge this, but you get the idea).

In the Marx and Fordyce (2015) matrix of baleen whale relationships, there are 69 different anatomical characters for the bulla and periotic - nearly double that used for the entire postcranial skeleton and soft tissue features (34). In the matrix derived from my Ph.D. thesis studies published in 2017 (also with R.E. Fordyce), I had a total of 148 cranial characters (excluding earbones), 72 characters for the periotic (alone), 31 characters for the tympanic bulla (alone), compared to a total of 69 postcranial and soft tissue characters. For comparison with teeth, I only had a total of 12 dental features - despite there being a fair amount of dental variation amongst toothed mysticetes. This means that earbone morphology, whether we like it or not, contributes considerably to how we reconstruct the tree of extinct cetacean relationships.

Another issue, which I'll bring up as an aside, is that there is a bit of a philosophical dilemma about earbones and how we construct phylogenetic characters. My Ph.D. adviser, R.E. Fordyce, prefers not to seek phylogenetic information from anything that is functional or adaptive - because functional convergence can lead to false cladistic relationships. For example, many of the characters initially supporting the "Platanistoidea" of G.G. Simpson was the super long snout in all modern river dolphins. We now know that these long snouts are either 1) primitive and shared with a common ancestor or 2) evolved convergently, because river dolphins are extremely paraphyletic (e.g. do not have a single common ancestor that only river dolphins diverged from). Therefore, Fordyce would say, we should not use rostral features as they are known to be influenced strongly by feeding ecology. My other mentor, J.H. Geisler, has a different view, which is we shouldn't make any assumptions about characters and try to use as many as possible to make as unbiased and assumption-free matrix as possible. Trying to figure out exactly how much functional convergence 'taints' characters is sort of a path to madness, in otherwords. And, ultimately, is likely not testable for many characters.

One of Fordyce's solutions was to focus on the earbones - a nice little archive of functionally meaningless bumps, grooves, pits, and crests preserving a wealth of phylogenetic data - a skeletal equivalent to genetic data that would influence the tree, seemingly without bias. However - I posit that we actually don't know what the function of many weird little bumps and grooves on earbones are *yet*, and we do not know that they are not function-less. So they may actually be prone to bias, and we would never know.

Aside from that fun, esoteric rabbit hole of disagreement over phylogenetic methods and character selection, earbones are often a major source of assemblage data: they are often identifiable to the family, genus, and occasionally even species level (some paleocetologists, like H. Ichishima, insist they can only be tentatively identified to the family level). As I mentioned above, perhaps half the cetacean species from the Purisima Formation in Santa Cruz are identified only from earbones, and skulls are known only from the more common species, with a few exceptions.

In sum, we care so much about whale and dolphin earbones because 1) we're blessed with species that have either no teeth or simple teeth useless in identification, in a field built on tooth anatomy; 2) these bones help us define species and 3) helps us reconstruct the tree of cetacean relationships, and 4) help us understand broader diversity and geographic trends by supplementing the record known from difficult to collect and study skulls and skeletons.