This is intended to be the third and final installment in this series, and I'll probably forget something in here - but here goes! The Xenorophus monograph was a monumental task to write and publish, and this "cliff notes" version will probably fail to cover everything. There's a LOT of meat in there, so who knows how many additional posts I'll have to make. To see part 1, click here, and for part 2, click here.
Typological thinking in Paleocetology and Ontogeny (growth) series in an early odontocete
Mammalian paleontologists have never had to worry much about the effects of growth-related changes in skeletal anatomy, because 1) mammals only get two sets of teeth, 2) their skulls do not seem to radically change in ways different than what we observe in modern mammals - as a result, mammalian paleontologists are not faced with the same problems our colleagues in dinosaur paleontology - pretending ontogeny doesn't exist and naming every different growth stage as a different species and then arguing forever about which species are distinct. Jack Horner and Kevin Padian called this "typological thinking" during my time at Montana State - overemphasis on pigeon-holing fossils into cookie-cutter "types". This sort of thinking led to naming different growth stages of tyrannosaurs (Tyrannosaurus rex, Nanotyrannus*), ceratopsians (Triceratops, Torosaurus - still controversial, unfortunately IMO), and even three different species named from 'ontogimorphs' in Pachycephalosaurus (Dracorex, Stygimoloch). This sort of thinking acts as a mental 'blinder' that keeps paleontologists from thinking creatively and flexibly about species concepts and definitions, and prevents consideration of hypotheses that might conflict with their chosen taxonomic framework.
*This one seems to have been settled recently and it seems as though they aren't different growth stages after all.
In marine mammal paleontology, we have different problems that, in my opinion, stem from similar psychological underpinnings, as well as anatomical shortcomings of the taxa we study. To start, certain researchers in the past have overemphasized splitting and holotype-based study. One unnamed scientist told me years ago "stick to holotypes". The un-spoken background here was that some competing researchers had carefully worked with larger samples of the 'species' this scientist had named and declared a bunch of them as synonyms, with new specimens filling in intermediate anatomical conditions. It turned out that the species named by the original researcher all happened to be outliers - and the new specimens instead showed a gradient. If you ignore the middle - the 'referred specimens' - you get to 'keep' all of the species, but only if you ignore data. You can quibble over whether some specimens are correctly identified, but generally speaking, it should always be a red flag whenever a senior scientist tells you to ignore a bunch of data.
Hand-drawn phylogeny of cetaceans, highlighting the diversity of cetacean family-level groups - and perhaps the overemphasis of family level taxa in prior paleocetological thinking. Many of these families are paraphyletic or polyphyletic. Modified from Barnes et al. (1985).
The second psychological problem is overemphasis of family-level groups in cetacean paleontology. I won't point any fingers, but those 'in the know' are probably very familiar with the few key figures. Some of the extinct families in Cetacea are not very well diagnosed in the first place, and frequently some are paraphyletic or even polyphyletic. My coauthor Jonathan Geisler is so averse to this sort of taxonomic thinking he prefers to just follow the phylogenetic tree - the direct signal from the data. We have so many unnamed taxa, especially in the Oligocene, that it is understandable to want to assign every specimen to a particular family - I do it, frequently - occasionally making Jonathan's eye twitch a little bit - but at least in my published writing, try to make it very clear what the definition and diagnosis of each family is. What the hell is a dalpiazinid? I still don't know, and the few fossils considered to be dalpiazinids have since been found to group with Waipatia-like dolphins. I use the term "Waipatiidae" all the time, but I recognize that it very well may be a grade rather than a clade - though some promising analyses by Amber Coste have recently shown it may be a monophyletic group.
The third unique aspect of typology in Paleocetology is the anatomical constraint imposed by the species we study. Most cetaceans simply just do not have particularly diagnostic teeth. Unlike terrestrial mammals, where juvenile and adult skulls can often be easily identified to the same species based on their teeth, dental morphology in cetaceans typically is so variable that species tend to blur together. But, there's more: within Neoceti, there are wildly varying tooth counts, so you cannot make 1:1 comparisons based on the position, with the exception of some toothed mysticetes that have only incipient polydonty - like Coronodon, for example. Oligocene odontocetes tend to have teeth that are readily identified to the family and occasionally genus level, and sometimes have species-level differences that can be distinguished when the teeth are set in skulls - but crown odontocetes from the Miocene and Pliocene have teeth that are frequently not even diagnostic at the family level, with a couple of exceptions here and there. And then there's baleen whales - which, aside from some Oligocene examples, don't even have teeth. All we have to go on is skull and mandibular anatomy, and occasionally vertebral counts for well-preserved skeletons. So, just like dinosaur paleontology, we have 1) teeth that are sort of useless in taxonomy and, in concert with this, 2) large body size associated with some pretty extreme skull shapes that are wildly divergent from the terrestrial mammal 'archetype'. As a result, we frequently have trouble figuring out the identity of juvenile cetacean specimens.
This is all a long-winded explanation of why I focused on reporting ontogenetic changes in Xenorophus. This is pretty much the first ontogenetic series ever reported for a stem odontocete. Sure, there are a couple of Miocene species out there with a few referred specimens and a juvenile or two (the beaked whale Messapicetus gregarius and Brachydelphis come to mind), but this is the first for any archaic dolphin on the odontocete 'stem' (e.g. diverging earlier than the most basal-branching modern species, the sperm whale, Physeter macrocephalus). Why is this critical? Two key reasons: 1) Because there are a couple of archaeocete whales with growth series, and they do not seem to have quite as much anatomical change during growth that Neoceti have; and 2) we don't really have much of a benchmark for assessing how much growth-related change there is in early odontocetes and what sort of features vary between species versus what varies during growth and can be ignored for the purposes of taxonomy.
Growth of the nasal bones in Xenorophus sloanii. Note that they elongate during growth, and in addition to clearly growing posteriorly, they also grow anteriorly - and appear to make the bony nares face more anteriorly than dorsally. From Boessenecker and Geisler (2023).
I already wrote extensively about the teeth in part 1, so now I'll write a bit about the nasal bones and the 'vertex'. This region is not always well-defined, but generally the area including the nasals and the tip of the occipital shield; this ends up including the sagittal crest in archaic Neoceti and archaeocetes, the frontals, and often the posterior tips of the premaxilla and maxilla. The vertex is typically the most informative area used in fossil cetacean taxonomy - and since it's not tooth enamel that is literally static in shape throughout growth, any changes in the anatomy of the vertex should be put under the microscope. One major evolutionary trend in cetacean evolution is cranial telescoping: the facial bones migrate ever further posteriorly in both mysticetes and odontocetes (and archaeocetes for that matter); a separate pattern, typifying mysticetes and probably a couple of odontocete groups, is anterior migration of the occipital shield. This can result in multiple bones being overridden and excluded from the vertex; the parietals are among the first of these to go (but later on in odontocete evolution - all xenorophids have dorsally exposed parietals). As it turns out, telescoping also changes during growth - with less-telescoped skulls in juveniles and more-telescoped skulls in adults. In Xenorophus, the intertemporal region shortens as the occipital shield increases in length (and becomes laterally broader as well). The nasals elongate, especially posteriorly, but also a little bit anteriorly. Muscle attachment crests all over the facial region become more strongly developed and sharply defined. In old adults, the frontal-nasal suture actually nearly disappears as the two bones fuse together. One curious feature is the rotation of the sagittal crest during growth: it is nearly anteroposteriorly directed in juveniles, but in old adults, it is rotated counterclockwise and points off to the left by about 5-10 degrees - more on asymmetry later.
What does this all suggest? We need to be careful about considering which features may be driven by ontogenetic changes rather than differences in species. For starters, some of these 'psychological blinders' prevented Al Sanders, Jonathan, and I - at least for a time - from considering any specimens as belonging to Xenorophus sloanii. Once the supposedly diagnostic features were realized to be related instead to growth, all the dominoes fell into place. Second - the vertex of a fossil dolphin skull can change quite a bit. Alarmingly, these changes continue far past all of the vertebral epiphyses start to close - CCNHM 104, for example, as a very different looking vertex than CCNHM 1077 and ChM PV 5022, but all of these specimens have a mature/adult stage of vertebral epiphyseal closure. In my opinion, based on this information, we should have at least a handful of solid diagnostic features rather than just one or two - because if you focus on just nasal shape, for example, you might even get three different species out of this collection of Xenorophus specimens.
Functional anatomy of Xenorophus: feeding morphology and oral pathology
Oligocene toothed whales have a surprising degree of feeding adaptations - and this is true even within the Xenorophidae - Inermorostrum has no teeth and a short snout, after all. We know already that there are other small bodied narrow-snouted xenorophids like Echovenator, Albertocetus, and Cotylocara in the Oligocene of South Carolina; there a ton of similarly sized "spear-tooothed" waipatiid grade dolphins, and of course, the giant macrophagous dolphins in the genus Ankylorhiza. How does Xenorophus fit in to this exactly? We'll take a closer look at the teeth, how the teeth interlock, tooth wear, snout shape, and pathologies in these specimens.
The teeth of Xenorophus are absolutely large - some of the largest after Ankylorhiza - but it's also a relatively big dolphin. The skull is quite large, about 60 cm long - and the body is up to about 3 meters, about 3/5 the size of Ankylorhiza tiedemani. Xenorophus is the second largest odontocete in the Oligocene assemblage after the 3-4 species of Ankylorhiza (most are unnamed) - which tells us something at minimum about the trophic level of Xenorophus. But, back to teeth. Xenorophus has incipient polydonty - typically about 13 teeth per quadrant, so two additional teeth from the primitive number for placental mammals (11). The dentition is heterodont, with about six to seven conical caniniform anterior teeth, three to four smooth-edged triangular and bilobate to double-rooted postcanine teeth, and three to four cuspate, wide, triangular, and double-rooted 'molariform' teeth; these are fuzzy boundaries, hence the variation in count. Crown height is similar from the front to the back of the dentition, but the crowns become much 'wider' (in lateral view) towards the back of the dentition. The enamel is smooth on the anterior teeth and develops rugose enamel posteriorly in the molariform teeth. Ridges at the base of the crown, called cingula (singular = cingulum) are present in the posterior teeth on both the inner and outer sides.
Xenorophid dolphins are unique amongst odontocetes for retaining deep pits in the mandible and maxilla that accommodate the teeth when the jaws are closed (occluded). This is a primitive feature* amongst cetaceans, inherited from basilosaurid whales - which have comically large teeth, and those teeth need to go somewhere when the animal closes its mouth. The way this works in basilosaurids (and earlier archaeocetes for that matter) is that the anterior caniniform teeth interlock in line with each other, like the teeth of a cartoon bulldog. This is accomplished in two ways: first, the gaps between the teeth are large enough for the opposing tooth to fit, and second, there's a pit - called an embrasure pit - eroded* into the lateral side of the bone in both the upper jaw (maxilla) and lower jaw (mandible). The posterior larger cheek teeth don't have large gaps between them, so the lower teeth actually slide on the 'inner' (tongue) side - aka lingual side - of the upper molars. This means that the anterior teeth typically only have wear facets on the tip of the teeth, wearing from contact with prey during biting, as these teeth don't really touch each other during jaw closing. However, the cheek teeth slide against each other like scissors, so the outer "lip" side of the lower teeth (labial side) gets a shear facet, duplicated on the inner (lingual) side of the upper teeth. These teeth also get worn down at the tips from biting prey. This style of tooth shearing is very similar to the carnassial teeth in mammalian carnivores like dogs and cats.
*The lack of embrasure pits in other odontocete clades makes this a very convenient feature in the formal diagnosis of Xenorophidae and also makes identification of xenorophid rostrum and mandible fragments super easy in the field. At least, here in Charleston, because they basically occur nowhere else.
**I say eroded because when we find archaic cetaceans where such a tooth was lost quickly or did not form, an embrasure pit never forms. In Dorudon atrox, there is a specimen where a tooth was lost evidently early in growth, and on the opposing jaw, there is no embrasure pit. The embrasure pit forms as a bone response to stress imposed by repeated closure of the mouth. Some extant carnivores develop embrasure pits, so I'm sure this could be evaluated in a much larger sample size with modern mammals.
Xenorophus has embrasure pits, but they're a bit different than in basilosaurids. The pits are present in both jaws along the anterior two-thirds of the tooth row, and positioned on the lateral side. This is like the condition in basilosaurid whales - the uppers and lowers are not exactly vertical but splay out slightly. However, there are no embrasure pits on the palate (upper jaw) between the last four upper teeth. Instead, there are deep embrasure pits on the mandible, between the lower teeth, and in line with them. From about the third to fifth lower postcanine tooth, these pits shift from a lateral position to being in-line with the uppers. This means that the upper and lower triangular molariform teeth are in-line with each other, and even more closely resembling a cartoon bulldog mouth! In other words, the mandibular teeth have shifted laterally and they are no longer shearing against each other like archaeocete molars. As a result, the wear on these teeth is chiefly along the front (mesial) and back (distal) cutting edges. However, the tips of these teeth are unworn in most specimens! What's going on there?
What's interesting here is that while this style of occlusion is very different from archaeocete whales - AND modern dolphins. This dentition is functionally adapted for biting prey along the entire length of the jaw, unlike archaeocetes, where only the anterior dentition was likely used for prey capture, and the molars used instead for 'prey processing' (aka chewing). This is basically the way modern dolphin jaws work - despite the embrasure pits. We hypothesized that the embrasure pits deepen so much so that the tips of the teeth never impact enamel from the opposing dentition, maintaining their sharp apex no matter how much of the mesial and distal edges become worn down. In old adults like CCNHM 1077, the embrasure pits are SO deep that the entire crown would have fit into a gum-lined pit and only the roots of the teeth contacted each other - the little wear that did happen to these teeth are a vestige from an earlier stage of growth. What's interesting here is that while the dentition is certainly anatomically heterodont, much of the evidence suggests that it was functionally homodont.
Only a single specimen has worn tooth apices - CCNHM 104 - which is an unusual specimen. CCNHM 104 is osteologically a subadult and it is not large - but it has a rather extreme degree of tooth wear, more extreme than any other specimen. There are a few possibilities that explain this. Perhaps it is actually ontogenetically very old, but had stunted growth, and the gaps between the teeth never became large enough to spare the tips from wear. Another possibility is that this individual consumed prey that was more abrasive, such as sharks - the dermal denticles wear down the teeth of modern offshore killer whales quite rapidly, and right down to the gumline. So, maybe it wore down its teeth rapidly and before the gaps could fully accommodate the teeth.
Missing teeth and a gnarly jaw infection in Xenorophus sloanii. From Boessenecker and Geisler (2023).
There are a couple of interesting pathologies in Xenorophus. Specimen ChM PV 7677 is missing four upper teeth, and CCNHM 168 is missing four lower teeth. In ChM PV 7677, there are ghosts of tooth sockets left, where there's a small irregular patch of bone with a few vascular foramina (holes for blood vessels) present. In CCNHM 168, these missing teeth in the mandible are further associated with what seems to be a healed fracture and bony callous in the right mandible and a pretty disgusting looking pathology with a large open window in the left mandible, right at the same position as the missing teeth, just posterior to the mandibular symphysis. Perhaps both mandibles suffered a hairline fracture here, with a pretty nasty infection - perhaps osteomyelitis - developing on the left side. This infection may have led to the loss of the teeth. We proposed that, between this and the possibility that CCNHM 104 fed upon unusually abrasive food (more likely than stunted growth in my opinion) - that Xenorophus may have engaged in risky feeding behavior. At such a large body size, it may have been tempting to go after prey that was larger and more difficult to capture than regular fish. This risky feeding behavior perhaps contributed to an unusual degree of tooth loss, mandibular fracture, and tooth wear.
Lastly, Xenorophus has a similar rostral proportion index with bottlenose dolphins - an RPI of 2.4-2.6, in other words - a rostrum (excluding the tip made out of just the premaxilla) that is about 2.5 times longer than it is wide at the base. This is suggestive of a generalist diet - not specialized for anything in particular, but permitting feeding on large and small prey. What about suction feeding? Xenorophus clearly has no degree of tooth loss, and an RPI that is too high for specialized suction feeding. A single hyoid bone from CCNHM 168 is rod-like, nearly identical to basilosaurids - differing from the widened, spatulate hyoids of modern delphinoids and ziphiids. While we know that some xenorophids like Inermorostrum were suction feeding specialists - available evidence indicates that Xenorophus was not a specialized suction feeder (despite probably being able to generate some suction like all toothed whales).
Vertebral anatomy and locomotion in Xenorophus
Very few stem odontocetes are known from vertebral columns - and, of the few that are, I've published most of them - Albertocetus, Ankylorhiza, and now, Xenorophus. There are two good Xenorophus specimens with much of the vertebral column preserved - CCNHM 1077 and ChM PV 5022. CCNHM 1077 has more of the caudal and thoracic vertebrae preserved, and ChM PV 5022 has a complete thoracic series. These specimens give us a count of seven cervicals, ten thoracics, ten lumbars, and at least a half dozen caudal vertebrae. The neck of Xenorophus is long - measuring approximately 40-45% of the length of the skull, and suggesting considerable more mobility than in extant odontocetes: the cervicals are not changed very much from the basilosaurid condition.
Skeletal reconstruction of Xenorophus sloanii, based chiefly on CCNHM 168 and CCNHM 1077. Artwork by yours truly. From Boessenecker and Geisler (2023).
In the caudal (tail) vertebrae, we found that there was no evidence of an external narrowed caudal "peduncle". In most modern cetaceans, with a couple of exceptions, the tail stock is very narrow and much higher (dorsoventrally) than wide - in many groups this results in vertebrae that are somewhat narrower than they are tall. This narrow peduncle allows the tail to slice through the water and making most of the force generated by up/down movements of the caudal fluke to be localized on the fluke itself - also permitting greater dorsoventral movement of the flukes. Basilosaurids have vertebrae in this region that are wider than long, so Mark Uhen (2004) suggested that Dorudon atrox may have had a sirenian-like tail stock. The few caudal vertebrae of Albertocetus meffordorum are approximately as wide as tall, and so in 2017 we proposed a similar hypothesis - that xenorophids lacked a narrow peduncle. The very complete vertebral column of Ankylorhiza* confirmed this, and is now confirmed again in Xenorophus. Ankylorhiza is of course not a xenorophid, but a more derived agorophiid-grade odontocete. A narrow peduncle is a derived trait, and so we (Boessenecker et al., 2020) proposed that this adaptation evolved multiple times - or, at least twice: once within mysticetes and once within crown odontocetes. This suggests that Xenorophus was not quite as strong or fast a swimmer as most extant odontocetes.
Observations of the vertebral column don't end there, though. We also took measurements of CCNHM 1077 to produce what I have started lovingly calling "Buchholtz diagrams" after Dr. Emily Buchholtz who started using them 20 years ago. It takes quite a bit of staring at these before you see the forest through the trees, and once you start to get them, they're deceptively simple. It also requires a bit of non-intuitive knowledge of what vertebral dimensions confer to bending of the vertebral column. Shorter, hockey-puck like vertebrae result in a stiffer column despite having more joints - and perhaps the addition of more intervertebral disks results in more stored elastic energy. This confers faster swimming speeds and characterizes many delphinid dolphins, phocoenid porpoises (especially Dall's porpoise, Phocoenoides dalli), as well as lamnid sharks and ichthyosaurs. Elongate vertebrae confer slower swimming speeds and perhaps greater manueverability - this is generally the ancestral condition in all marine tetrapod groups, though it also is present in swordfish, marlins, and sailfish - some of the fastest marine vertebrates (however, as I understand it, they are fast in bursts). These are generalizations, and ignores changes in vertebral proportion throughout the column. So, what happens when vertebrae change shape?
Within cetaceans, Buchholtz (2001) came up with a three-fold division of vertebral dimension patterns. Pattern 1 cetaceans have vertebrae of similar proportion throughout the column, with some anteroposterior shortening in the lumbar/caudal series, no narrowing of the peduncle, and based on modern species, are the least maneuverable; up/down movement (undulation) occurs nearly throughout the vertebral column. Pattern 1 species include modern and fossil mysticetes, the sperm whale (Physeter), archaeocetes, and probably Albertocetus. Pattern 2 cetaceans have a long series of anteroposteriorly compressed vertebrae in the anterior tail and lumbars that is somewhat stiffened, developed as a stiffened 'peduncle'; much of the up/down undulation is localized within the vertebral column to the posterior thoracic vertebrae and anterior lumbars; there is moderate narrowing of the tail stock and a bit of an external narrow peduncle is developed. Pattern 2 species include beaked whales (Ziphiidae), white whales (Monodontidae), the four extant river dolphins, false killer whale (Pseudorca), and of course, Ankylorhiza (though likely lacking a narrow peduncle). Pattern 3 cetaceans have hockey puck shaped vertebrae that are circular to oval and much taller than they are long (anteroposteriorly) and secondary increase in vertebral length in the anterior caudal vertebrae, suggesting that these function as an "elbow" in the vertebral column; these are adaptations for very rapid swimming with a stiffened torso. Pattern 3 species includes most Delphinidae including Orcinus, Steno, Tursiops, Lagenorhynchus, Lissodelphis, as well as extant Phocoenidae.
"Buchholtz diagram" - aka vertebral dimension plot - for CCNHM 1077, the most completely preserved single specimen of Xenorophus sloanii. From Boessenecker and Geisler (2023).
Xenorophus seems to be a Pattern 1 species, though bordering on some Pattern 2 adaptations. The posterior lumbar vertebrae lengthen and also become dorsoventrally deep and large in size, along with the anterior caudals - suggesting a degree of Ankylorhiza-like stiffening of the tail stock. Thoracic vertebrae are wider than long, which is a primitive Pattern 1 feature. Though some of the increases in vertebral length, width, and height are more extreme than in Ankylorhiza, there is no trend towards foreshortening of the vertebrae in the tail stock, a major feature of Pattern 2 swimmers. In comparison to other medium to large size cetaceans from the Oligocene of South Carolina, Xenorophus sloanii was a faster and more efficient swimmer than Coronodon, but less so than Ankylorhiza tiedemani.
Dental asymmetry in Xenorophus sloanii: measurements are in millimeters, whichever side is further anterior. For example: in CCNHM 168, all of the teeth on the right are shifted anteriorly relative to those on the left, by 2.1 to 8.5 mm. From Boessenecker and Geisler (2023).
Cranial, Dental, Mandibular, and Vertebral Asymmetry in Xenorophus
Sometime during the pandemic I was on the phone with my coauthor Jonathan Geisler and we were talking about the oddness of xenorophid skulls, and we thought they were asymmetrical - but it was just a gut feeling. "Something about when you look at their skulls in anterior view... they just look wrong". With this in my mind, I decided to look for more concrete evidence. Sometimes, you just have to start putting lines down over photographs to see if anything is violating basic symmetry. The problem is, it was difficult to even know where these landmarks were supposed to be, given what we eventually found out. At first, I noticed that the palate and teeth were asymmetrical - and these made it into my 2020 SVP talk, which was of course recorded ahead of time and uploaded to the SVP conference webpage rather than delivered in person.* While I was working on the talk, I noticed that the teeth of CCNHM 168 and 104 were all wonky - most of the postcanine teeth right are shifted anterior to the ones on the left, by at least a few millimeters and in some tooth positions over a centimeter! What the hell?! Nobody had ever documented that in a fossil cetacean before. After some more looking, it turns out that many basilosaurids have asymmetrical dentitions - and nobody had really noticed before!** What could drive the asymmetry in these cetaceans? My German colleague Dr. Julia Fahlke published an interesting article in 2012 on asymmetry in Basilosaurus, noting that the rostrum was twisted longitudinally, bent to the left, and that the thinnest parts of the pan bones of the mandibles were asymmetrical, with one shifted further anterior than the other (more on that below). I suspected immediately that the asymmetrical teeth are likely driven by the bending of the rostrum to the left side. But what about Xenorophus? Surely, its rostrum was not also bent - that had been observed only in basilosaurids so far.
*The 2020 conference is actually what killed online conferences
for me. Surely I can do an online conference at the same time I have a
full teaching load! Of course, the majority of my colleagues were in
other states and not back in the classroom - we had just been forced
back into the classroom by the school administration (because red state)
and the Q&A session was scheduled for a time during my first week
of labs. I was already terrified of getting Covid, and I noticed that
virtually everyone on the zoom call was unmasked - because they were at
home. I couldn't speak, because I was in class. I typed out a couple
responses and only had a couple of questions before Q/A moved on. The
remainder of the week I got considerably more stressed as I needed to
tend to my teaching but also felt a bit worried that I was neglecting my
responsibility to attend the conference - and decided to just watch all
of the talks on the weekend. The stress of this stupid experience made
me realize I could never juggle an online conference, and I'll never
attend one again. I know that they're great and everything for folks who
can't travel, but I just can't do both. I'm not teaching any more, so
maybe I can give it another go.
**Doesn't really
surprise me, to be honest. Much of the previous work on basilosaurids,
with some notable exceptions, has not focused on cranial anatomy. And,
in the few basilosaurid whales with good, detailed cranial and dental
descriptions - the skull is quite asymmetrical, and perhaps exaggerated
by burial compaction - e.g. Cynthiacetus peruvianus, described well by Martinez-Caceres et al. 2017.
Asymmetry of the palate in Xenorophus sloanii: note the left-right asymmetry of the maxillopalatine suture, median palatine suture, and the exposure of the vomer. mx = maxilla; pa = palatine; v = vomer. From Boessenecker and Geisler (2023).
Even before this, I had noticed the asymmetry of the palatine bones. The right palatine is always larger than the left in Xenorophus spp. and the maxillopalatine suture is shifted further anteriorly on the right, by 5-15 mm, and the suture is also nowehere close to symmetrical - unlike all other stem odontocetes I've looked at (Ankylorhiza, Simocetidae, Waipatiidae, Squalodon, etc.). This asymmetry of the palatines is shared with Albertocetus, and the right palatine is also longer/larger than the left. It doesn't stop there, though: the vomer is asymmetrical, with a very elongated trapezoidal shape. The vomer exposure in the palate becomes anteroposteriorly longer during growth and the underlying maxilla 'unzips', but asymmetrically so: the left maxilla seems to be emarginated on its lateral side to expose the vomer posteriorly, and anteriorly, it's the right maxilla. But it doesn't stop there! In adults, the median palatine suture is shifted to the left of the midline, and is actually bowed somewhat in specimens like CCNHM 1077.
Skulls of Xenorophus sloanii in anterior view. Something just seemed off... From Boessenecker and Geisler (2023).
Rostral deviation to the left in Xenorophus sloanii. From Boessenecker and Geisler (2023).
I was suspicious about the rostrum given Jonathan's comments over the phone that something just "doesn't look right". When I looked down the tip of the rostrum (in anterior view), you can't really line up everything properly - at no single angle did the skull look symmetrical. In the photo above, I settled on shooting down the axis of the rostrum. So, when I drew a line on these skulls along the midline of the braincase - from the middle of the occipital condyles through the nasals - I noticed that in every specimen, that line went to the right of the midline of the rostrum. Could it be due to burial deformation? What about the specimens being improperly assembled during "puzzle piecing"? Perhaps, but there are several important problems with each of these. First, specimens in the Ashley Formation are typically three dimensionally preserved, and there's not much in the way of structural shearing here. Chandler Bridge Formation specimens are frequently highly fractured with fragments separated from one another. These specimens are cleaned, one fragment at a time (or occasionally screenwashed from goopy sediment), dried, and glued together one joint at a time. This is occasionally done for Ashley Formation specimens when the limestone 'marl' is quite soft. There is one specimen that I think has been deformed naturally: the Xenorophus simplicidens holotype (CCNHM 8720), the rostrum of which is somewhat deformed and actually bends off to the right side by a few degrees - owing to many relaxed fractures infiltrated with silt. However, all other skulls complete enough to evaluate bed about 1-5 degrees to the left. What's noteworthy is that these skulls were preserved and prepared differently. Specimens CCNHM 104, 168, and the X. sloanii holotype (USNM 11049) have some hairline fractures but were basically prepared traditionally through matrix removal and stabilization, and were not fractured into a zillion pieces. The Xenorophus simplicidens paratype, ChM PV 4823, and possibly referred X. sloanii skull ChM PV 7677, were puzzle-pieced together - and both still point off to the left side. Further, in ChM PV 4823, there are some artificial gaps in between the left and right halves of the rostrum, which suggests to me that it was reassembled and the preparator made their best attempt to 'force' it into being symmetrical - and ultimately was still unable to do so.
Growth changes and asymmetry of the vertex in Xenorophus sloanii. Apologies for showing this one again. From Boessenecker and Geisler (2023).
Some other details are evident in the skull. In adult specimens like CCNHM 1077, 168, and ChM PV 5022, the sagittal crest is asymmetrical and actually tracks to the left - as if it's been rotated clockwise. This suggests that during growth the temporalis muscles are asymmetrically developed, with the stress and strain from biting resulting in an asymmetrical muscle attachment ridge. The antorbital fossae - large basins on either side of the rostrum, adjacent to the bony nares - are asymmetrical in Xenorophus. These basins are unique to the xenorophids and are actually a synapomorphy of the family, but there are some subtle ways in which they vary. In Echovenator and Cotylocara, the fossae are symmetrical. But in both species of Xenorophus, they are asymmetrical. In Xenorophus sloanii, the left fossa is more deeply excavated, anteroposteriorly longer, and transversely wider than on the right. Curiously, in the Xenorophus simplicidens paratype (ChM PV 4823), the opposite is true - the right fossa is longer than the left! It's unclear owing to incompleteness if it was wider, however. For a while I thought that perhaps the asymmetry of the palatines was linked to the asymmetrical fossae with a longer palatine underlying the shorter fossa and vice versa in Xenorophus sloanii. However, once I realized that the rostrum was bent to the left, it made sense: the palatine extends further underneath the right half of the rostrum as the entire complex has been rotated counter-clockwise (in dorsal view).
The mandibles of Xenorophus sloanii in dorsal view, showing asymmetry of the lower dentition. Also, you can pretty clearly make out the missing teeth in CCNHM 168. From Boessenecker and Geisler (2023).
Remember the asymmetry of the teeth? What about the rest of the lower jaw? Unfortunately, there's only one specimen with mandibles, and those mandibles do have a couple of pathologies - so it's difficult to reconstruct and measure how offset the teeth are. However, they are quite clearly offset as well. What's far more damning, however, is the relative position of the posteriormost tooth. When I measured from the mandibular condyle (jaw joint) to the back of the last tooth, the distance on the left mandible is an entire centimeter shorter than on the right! Think about it: if you rotate the rostrum to the left, you're stretching the right side and shortening the left.
Rampant asymmetry throughout the skull, mandibles, dentition, and even the vertebral column of Xenorophus sloanii. I wouldn't go so far as to say that the entire vertebral column was asymmetrical - but in this specimen, CCNHM 168, every single vertebra is asymmetrical in ways that cannot be explained by burial (diagenesis).
But we really aren't even close to done with the asymmetry - because it continues into the postcranial skeleton. Yes, even the vertebrae are asymmetrical! Virtually all of the cervical vertebrae have transverse processes that are somewhat differently shaped left to right - having either a different outline, projecting in a different direction, or being longer on one side. This is most extreme in the sixth cervical, frequently identifiable in land mammals and archaic cetaceans because of the long, ventrally directed processes. In THREE different specimens of Xenorophus (CCNHM 168, above, CCNHM 1077, and ChM PV 5022), the left process is more vertical than the right, longer from base to tip, and also more greatly expanded in lateral view (by a half centimeter) - see the figure above. The atlas vertebra is also wedge-shaped - the left side is a half centimeter longer than it is on the right, which means that the anterior and posterior articular surfaces seem to diverge at an angle of over 4 degrees. This suggests that the entire braincase is canted about 4 degrees to the right. Combined with the bending of the rostrum, this suggests that the entire osteological midline is a zig zag. Unfortunately, given that the entire top of the braincase is slightly off kilter, it's difficult to use any landmarks on the skull to test this hypothesis - but, more on this shortly. In addition to this, the neural spines of the thoracic and lumbar vertebrae are also asymmetrical - in ChM PV 5022, CCNHM 168, and CCNHM 1077, each vertebra with complete neural spines is either tilted to the left and/or twisted a little bit along its vertical axis, and in the same direction (clockwise in dorsal view) - and most vertebrae have evidence of both. The transverse processes in the lumbars are mostly broken, but there is a slight difference in orientation in CCNHM 168 where they are well-preserved.
Press release image of the asymmetrical skull of Basilosaurus isis reported by Fahlke et al. (2011), showing the bending of the rostrum to the left (our right) and twisting of it along the long axis. Photo credit: Julia Fahlke.
Asymmetry of the "pan bones" in Basilosaurus isis showing differences in the thickness of the wall of the mandibular fossa/foramen. From Fahlke et al. (2011).
So, what gives? Why is there such an extreme degree of asymmetry? The postcranial asymmetry is odd, but there may be a modern analog: the Ganges river dolphin, Platanista gangetica, engages in side-swimming, and maybe Xenorophus did something similar. However, there's no data available indicating that the postcrania of this species are asymmetrical, and no a priori reason to infer that side swimming is any different than swimming dorsal-up - so I don't really buy that. Since we already know the rostrum is bent, perhaps the vertebral asymmetry is a consequence of cranial asymmetry imparting unequal left/right drag onto the head - pretty obvious if the snout sticks off to one side. So, let's start with the head. Rostral bending and twisting is already documented in basilosaurids and some protocetids by Fahlke et al. (2011), who also determined that the 'pan bones' of the mandible are asymmetrically positioned: the thinnest part of the left mandible is approximately 10 cm further anterior to the thinnest zone on the right, perhaps indicating that the mandibular fat pads are positioned differently.* This study proposed that cranial asymmetry is driven by asymmetrical hearing in late archaeocetes as an adaptation for improved directional hearing - and that this asymmetry was lost in Neoceti as underwater hearing continued to improve. [A quick refresher: sound waves tend to travel through water and bone at more or less the same speed, arriving at each ear simultaneously. Mammalian directional hearing relies on being able to calculate very slight differences in 'arrival time' of sound waves between your two ears - interpreted by your brain as a direction of the sound wave. Underwater, this is not possible - so cetaceans have evolved a series of adaptations to acoustically isolate each ear from the rest of the body, funneling left/right sound waves into the fat pads, channeled through fat channels directly to each middle ear - thereby restoring directional hearing. Fahlke et al. (2011) proposed that the asymmetry in the skull and jaws was a sort of analog to the hearing system in owls, which famously have one ear placed higher than the other - this lets owls 'zero in' on the direction of even very quiet sounds in the dark of night. In Fahlke's hypothesis, basilosaurid whales did something similar after acoustically isolating their inner/middle ears.
*Ted Cranford and others (2008) demonstrated, however, that sound likely travels through the "gular" pathway - through the throat and into the medial side of the fat pad where it is exposed by the cavernous mandibular foramen (e.g. not roofed over by bone medially) and not laterally through the pan bone as originally proposed. They indicated, as others had, that sound waves probably cannot travel through the pan bone, no matter how thin it is, owing to the somewhat higher density of the bone - this is called acoustic impedence, and would cause much of the sound signal to bounce off the bone rather than travel through it. This would explain why the mandibular foramen is so large in cetaceans - keeping a bony lid off of the fat pad. In Basilosaurus isis, the thinnest part of the pan bone is in completely in front of the mandibular foramen - suggesting that this part would not really receive sounds, and because the mandibular foramina in the left and right mandibles look nearly identical, this mode of asymmetry may not necessarily indicate asymmetrical hearing.
We largely agree with this hypothesis, though Fahlke et al. (2011) did not really explain why the snout was bent or twisted as a consequence. We found that the "pan bones" in Xenorophus are probably asymmetrical in orientation, with the left posterior mandible being somewhat more anteroposterior facing and the right being posteromedially facing. In other words, the mandibular fat pads have been rotated counter-clockwise in dorsal view. Evidence from the vertebral column of CCNHM 168 suggests it's therefore not the rostrum that's twisted - it's the braincase! In sum, reorientation of the posterior mandibles drove asymmetry of the posterior skull, and the asymmetry of the rostrum resulted in stronger hydrodynamic forces (e.g. drag) that drove asymmetrical muscles, which resulted in asymmetrical muscle attachments in the vertebral column.
Phylogenetic Placement of Xenorophus
As in previous studies, Xenorophidae are recovered in our analysis as one of the earliest diverging clades of Odontoceti. Only two peculiar early toothed whales show up more basally - Mirocetus riabinini, an unusual toothed whale from the Oligocene of Azerbaijan, and Ashleycetus planicapitis from the Ashley Formation. Mirocetus is positioned as the basalmost odontocete, and Ashleycetus is positioned as sister to the Xenorophidae. While the skull of Ashleycetus is well-preserved, it is missing teeth and earbones, and the skull of Mirocetus is poorly preserved with uncertain sutures and no earbone morphology that can really be scored - meaning that interpretation of Mirocetus is challenging. Limited codings for these fossils mean that each tends to float around a bit on different variations of this same matrix (e.g. Sanders and Geisler, 2015; Velez-Juarbe, 2017, 2023).
Regardless, these are singletons missing lots of character data, and the Xenorophidae are markedly different in several ways such as 1) their diversity, 2) consistent placement as a major clade at the base of Odontoceti (e.g. existing as a bit of an 'anchor point' for these other Oligocene dolphins that float around a bit on the tree), and 3) their anatomical distinctiveness that leads to well-resolved relationships within the clade Xenorophidae. Furthermore, this is a bit of a preservational issue, but for several genera of xenorophids - Albertocetus, Echovenator, and Xenorophus - there are multiple specimens known for each taxon (a massive understatement) and these taxa are known from nearly all cranial, earbone, and dental characters (and even most of the "singleton" taxa like Cotylocara are as well).
New Odontocete Clades
There have been a proliferation of fossil toothed whale genera and families, especially so on the stem group, over the past 20 years or so. However, we haven't had any 'suprafamilial' clades to help break this up - so we thought of two possibilities, and defined these clades in our paper.
The first is the Amblyoccipita. This can be thought of as the most inclusive group of toothed whales that excludes the Xenorophidae - the most basally positioned group of odontocetes. Amblyoccipita means "round braincase" in Greek, referring to the rounded supraoccipital shield in these dolphins and contrasting with the primitive triangular vertex seen in Xenorophidae. Amblyoccipita includes most simocetid dolphins, agorophiid-grade dolphins, along with waipatiid-grade and squalodontid-grade dolphins - and of course, Crown Odontoceti. This clade had been begging to be named for 20 years, and saying "the most inclusive odontocete clade excluding xenorophids" has gotten a bit tiring.
The second clade is the Stegoceti, meaning "roofed cetaceans" in Greek. This refers to the intertemporal region of the skull being transversely expanded so that the temporal fossa is partially or completely 'roofed over' in dorsal view. This clade excludes all odontocetes with a narrow intertemporal region that is round in cross-section - so the Xenorophidae as well as the Simocetidae and a few others like Ashleycetus and Mirocetus - and includes the "Agorophiidae" (e.g. Ankylorhiza, Agorophius) and all later stem odontocetes.
The basic definitions of the new clades. The Amblyoccipita are all odontocetes with a rounded vertex (blue line) - in most xenorophids, the vertex is pointed (though Cotylocara and Echovenator have secondarily converged on the rounded morphology). In simocetids like Olympicetus and Simocetus, agorophiids like Agorophius and the possible agorophiid Ankylorhiza, waipatiids, squalodontids, and nearly all Miocene to modern odontocetes in the crown clade, the anterior edge of the occipital shield is rounded. The Stegoceti include agorophiid-grade odontocetes like Ankylorhiza which have the intertemporal region roofed over by the supraorbital process - the posterior end of which is shown in red. Waipatiid and squalodontid-grade odontocetes, and all Crown Odontoceti have a wide intertemporal region that is 'roofed' over.
Other odontocete clades need to be named, but the later-diverging parts of the odontocete stem are much less clearly resolved. However, further exclusive clades defined by the exclusion of "agorophiids", "waipatiids", and "squalodontids" are worth naming in the future. Further, it seems possible from our phylogeny that single-rooted teeth may have evolved just outside the crown group, and so a "Homoceti" or "Homodontoceti" might be nameable at some point.*
*One problem is that different matrices give different results. Smaller matrices with fewer characters and fewer taxa often place waipatiids, squalodontids, and other oddballs with clearly double rooted teeth into the poorly defined "Platanistoidea" and well into the crown group, though I don't have much confidence in these hypotheses. I personally think single rooted teeth is likely to be ancestral for the entire crown group - but this is the essence of differences in phylogenetic philosophy and parallelisms/convergence.
It's worth stressing here that these two clades are very commonly recovered in nearly all phylogenetic analyses of stem odontocetes. The only analyses that tend not to recover these clades have only coded for a few stem odontocetes and therefore simply don't have enough taxa to reflect these relationships. Given that we've seen these clades again and again and again in 20+ years of phylogenetic analyses of early odontocetes, we felt it was time to name them.
Evolution of Body Size in Odontoceti
I was prompted to investigate the evolution of body size across the archaeocete-neocete transition because many early mysticetes and odontocetes are much smaller than your average basilosaurid.* We did find some support for this - the mean archaeocete body size, represented by skull width as our proxy for body size - is somewhat larger (~40 cm skull width) than in odontocetes, which start off around 20 cm and the mean rarely exceeds 25 cm. This is a bit of a modest change, but it's also the average of each - the minimum body size of odontocetes is considerably smaller, around 10-15 cm, than the equivalent minimum in basilosaurids - around 20 cm or so (and even these are outliers among basilosaurids).
*Sizes are more congruent with protocetids - however, protocetids are semiaquatic and therefore likely have a maximum body size imposed upon them by virtue of hauling out on land - just like your average pinniped is on the whole much smaller than your average cetacean.
The story with toothed mysticetes is a bit more complicated - complicated by none other than Llanocetus denticrenatus, the largest - and earliest - toothed mysticete. It's got a skull much larger than any basilosaurid (skull width almost a meter), though I imagine it was probably considerably outsized in body length by Basilosaurus cetoides, Basilosaurus isis, and Perucetus colossus (not included in our study). However, if you ignore Llanocetus as a bit of an outlier, toothed mysticete mean body size is otherwise much smaller than basilosaurids - see for example the relatively small species of Fucaia, Chonecetus, Mammalodon, and even every single species of Aetiocetus is considerably smaller than your average basilosaurid.
Evolution of Polydonty in Cetacea
Xenorophus is quite interesting owing to the fact that it is already polydont despite being one of the earliest diverging odontocetes. Polydonty refers to the evolution of additional tooth positions beyond the primitive number for placental mammals - 11 per quadrant, for a total of 44. Humans and other old world primates have a total of 32 (we've lost a pair of incisors, and two different premolar positions). Dolphins are well-known for their toothy grins, with many species possessing 20-40 teeth per quadrant. The La Plata river dolphin, Pontoporia blainvillei, possesses between 50 and 60 teeth per quadrant - the highest tooth count of any modern mammal. The extinct dolphin Parapontoporia from the Miocene-Pliocene of California has the highest tooth count of any cetacean ever, with approximately 90 teeth per quadrant - for a blistering total of about 360 teeth! (I've been told Pomatodelphis might exceed this, but am unaware of any specimens conclusively beating this number). Modern baleen whales of course have no adult teeth, but incredibly have polydont fetal dentitions that are resorbed. So, there's abundant evidence of polydonty in modern odontocetes and extinct crown odontocetes - but when did this start? And did polydonty evolve once or multiple times within Neoceti? Let's take a closer look at whale ancestors.
Most archaeocete whales like protocetids have 11 teeth in each quadrant, and basilosaurid whales interestingly have lost one upper molar (for a total of 10). The toothed mysticete Coronodon has 11 upper teeth, but a total of 12 mandibular teeth - indicating that early mysticetes developed incipient polydonty very, very early. This seems to indicate that the ancestral condition for the clade Pelagiceti (Basilosauridae + Neoceti) is to have one additional tooth in the mandible - and evidently, losing an upper tooth is a basilosaurid feature. So, Coronodon seems to have evolved one additional tooth position in each quadrant versus basilosaurids. Other toothed mysticetes have only modest increases in tooth count, with 1-3 additional teeth present in species of Aetiocetus and basilosaurid-like tooth counts in Mammalodon and Janjucetus.
On the other hand, Xenorophus has 13 teeth per quadrant, and one specimen even has 14 teeth - CCNHM 168, for a total of 10 postcanine teeth instead of the typical 7 present in placentals. Other xenorophids have tooth counts all over the place, with zero in Inermorostrum and over 23 in one unnamed xenorophid (ChM PV 4746). Remarkably, within this family, not only is there polydonty but it seems as though there's no constraint on the number of teeth and quite a bit of variation already - in stark contrast to the cookie-cutter dentition of their basilosaurid ancestors.
No comments:
Post a Comment