Sunday, May 13, 2012

Parallel evolution in gigantic teleosts and baleen whales: filling the filter feeding niche during the Mesozoic

When people usually think of gigantic filter feeding critters in the sea, most people think of humpback whales - and the general group of marine mammals that I work on. Few marine mammal biologists could tell you much from a deep time perspective on the evolution of large filter feeding marine vertebrates - and few paleontologists in general even think about other marine organisms occupying a similar niche. In the oceans today, we also have a number of gigantic filter feeding sharks- the megamouth shark, discovered only in the late 1970's (Megachasma pelagios), the basking shark (Cetorhinus maximus), the whale shark (Rhincodon typus), and the manta ray (Manta birostris). Each of these groups (Mobulidae, Cetorhinus, Rhincodontidae, and Megachasmidae) all have relatively deep roots: early filter feeding devil rays and whale shark-like elasmobranchs are known from the Paleocene, and basking sharks from the Eocene. Megamouth sharks are known from the Pliocene and Miocene, and I am aware of some specimens in UCMP collections from the Oligocene and Miocene of Oregon. In 2007, Kenshu Shimada reidentified a tooth from the Cretaceous Greenhorn Fm. of Colorado he had originally published as Johnlongia sp. as a megamouth, and named it Megachasma comanchensis - which at the time was fascinating, because this was the first (possible) fossil record of a filter feeding shark prior to the Cenozoic. Baleen whales would not evolve until the latest Eocene.
The holotype and paratype teeth of Megachasma comanchensis, from Shimada 2007.

This has been an issue, because many researchers have noted or at least wondered about why the Mesozoic seas seemed to be devoid of gigantic filter feeders. Certainly, we know from the fossil record that there was no shortage of planktonic organisms to gulp up. How else could food webs be so drastically different between the Cretaceous and Paleogene? Nobody had a clear answer, and thus far it doesn't really look like any marine reptile we know of was really evolving towards a gigantic filter feeding ecology (there is one exception, the southern plesiosaur Aristonectes, which has a lot of tiny teeth that could have functioned as a sieve, and has generally been interpreted as a filter feeder). It was something I thought quite a bit about over the last few years - and to know avail. Aside from this Cretaceous species of Megachasma, the gigantic fish Leedsichthys was well known to be a gigantic (~10m) filter feeding teleost- but that was one genus known from a few fossils in the Jurassic.

 A depiction of Leedsichthys problematicus by Ray Troll.

In February 2010, I was fairly excited (about as excited I can get for anything older than the Paleogene...) a new article published in Science by Matt Friedman and others, publishing several new genera and records of other gigantic filter feeding pachycormid fishes, similar to Leedsichthys, and from numerous continents. These included Rhinconichthys from the Lower Chalk (Upper Cretaceous) of the southeastern UK,
an unidentifed toothless pachycormid from the Inferior Oolite (Upper Jurassic) of Dorset in the southwestern UK (for American readers - Dorset is a county along the southern coast of England, and is the home of Lyme Regis and the famed Mary Anning - along with being the setting of Jane Austen's novel Persuasion, and the setting of the book and movie The French Lieutenant's Woman starring Jeremy Irons and Meryl Streep, but I digress). They also reported the much better preserved Bonnerichthys from the Upper Cretaceous Niobrara Chalk in Kansas, and a Rhinconichthys-like fossil from the Upper Cretaceous Yezo Group of Hokkaido, Japan. Importantly, they identified fossils of this taxa from three different continents, indicating they were very geographically widespread. Secondly - and most importantly - they documented that these poorly known fishes were not just known from the Late Jurassic - but from the Middle Jurassic until the close of the Cretaceous.

The new stratigraphic range of known pachycormid fishes on the left (Figure 3 from Friedman et al. 2010) and the skull and partial skeleton of Bonnerichthys (Figure 2 from Friedman et al. 2010).

Interestingly, the fossil record now appeared to show that from the Jurassic onwards, that gigantic filter feeding vertebrates had continuously inhabited the earth's oceans until the present day. Prior to this study, the big question had been "Why did Mesozoic oceans lack abundant gigantic filter feeders?" I'll continue the dialogue by asking the opposite: Say we didn't have these, and with a hypothetically much better sampled record, had concluded the Mesozoic did not have any gigantic filter feeders. What sort of hypothetical situation could cause that? This may seem like a nonsensical and arbitrary question to ask, but I'll remind you that we don't really have any large Triassic filter feeders. I don't necessarily have an answer - but, one could easily surmise that there are a ton of morphological adaptations needed in order to even try filter feeding. Whatever morphology you start with is probably going to be something like piscivory - catching individual fish (or, crustaceans or cephalopods if you like). It's difficult to go from raptorial feeding to bulk feeding, and once a taxon is in a committed filter feeding niche - it's probably a one-way ticket; it is pretty hard to catch fast and maneuverable single prey items with a slow-closing mouth the size of pickup truck (Humpback whales do catch fish, and accidentally birds- but fish catching is simply done the same way they commit krill genocide). Some adaptations for filter feeding can probably be categorized as evolutionary 'ratchets' - development of a straining apparatus, and loss (or reduction) of teeth, both of which characterize all groups of gigantic filter feeding marine vertebrates. It is a little more clear-cut with fish, which already have a filtering apparatus needing little modification, and many fish already filter feed even at small body sizes, such as salmon; but for marine tetrapods, a really serious facelift needs to be performed, as teeth by themselves are not as efficient as, say, baleen is. This transition may necessitate some pretty weird transitional exaptations - we're still unclear on how baleen whales did it, for one. There have been and will be many more papers published on this particular topic, hopefully some from myself (regarding eomysticetids).

A painting of Bonnerichthys, an elasmosaurid, and some cephalopods by Robert Nicholls.

A more recent study by Matt Friedman (2011) expanded more on the evolution of pachycormiform fishes, and reinterpreted the Lower Jurassic fish Ohmdenia (from the Posidonia Shale of Germany) as an early pachycormiform, preserving an intermediate morphology between the ancestral condition and the derived, gigantic filter feeding condition. Derived pachycormiforms are toothless, with elongate, narrow jaws and a very large oral cavity, along with gigantic size. Ohmdenia also bore an elongate jaw, was relatively large (~2.5 m), but, like other basal pachycormiforms - retained a dentition. Its dentition had been reduced to a series of extremely tiny and stout teeth, which Friedman (2011) suggests were adapted for grasping soft bodied prey, rather than piercing the flesh of fish or other harder-bodied organisms. Two belemnites were found near the abdominal region of the skeleton of Ohmdenia, and these may represent gut contents- but the skeleton is disarticulated, so it is unclear if they arrived on the seafloor after the giant fish did. The elongate jaws of Ohmdenia also suggest comparatively weaker bite force than earlier pachycormiforms.

The holotype skeleton of the early pachycormiform fish Ohmdenia. From Friedman (2011).

To investigate the role of Ohmdenia in pachycormiform evolution, and to compare the evolution of filter feeding in these enigmatic fish and baleen whales, Friedman (2011) took a set of measurements reflecting various aspects of feeding in a number of different fossil and modern mysticetes, and pachycormiforms, and conducted a principal coordinates analysis. This allowed him to construct a "morphospace" - effectively, a field whose coordinates correspond to varying morphological characteristics (i.e. those which are measured by the researcher). A long while ago, in my Macroevolution course at MSU, we discussed the concept of "adaptive peaks" - regions of morphospace that are adaptive ideals; whether or not an organism or a clade achieves an adaptive peak, who knows - there are always circumstances that could preclude an organism from occupying some part of morphospace (i.e. anatomical constraints). Friedman's analysis showed that, interestingly - pachycormiforms and mysticetes, although originating at different regions of morphospace in the analysis - both converged onto the same adaptive peak. Furthermore, Friedman (2011) showed that in both cases, each group followed the same changes in this sequence: changes in dentition and mandibular geometry, loss of teeth, and evolution of giant body size. All in all, a rather impressive and fascinating study.
Phylogeny of pachycormiforms (A), transition of lower jaws in mysticetes and pachycormiforms (B), and morphospace analysis;  pachycormiforms in red, mysticetes in blue - solid circles = ancestral forms, triangles = Ohmdenia and transitional mysticetes, and open circles = filter feeding pachycormiforms and mysticetes, convering in the lower right hand corner (C). From Friedman (2011).


Friedman, M. 2011.Parallel evolutionary trajectories underlie the origin of giant suspension-feeding whales and bony fishes. Proceedings of the Royal Society B: 279:944-951.

Friedman, M., Shimada, K., Martin, L.D., Everhart, M.J., Liston, J., Maltese, A., and Triebold, M. 2010. 100-Million-year dynasty of giant planktivorous bony fishes in the Mesozoic seas. Science 327:990-993.

Shimada, K. 2007. Mesozoic origin for megamouth shark (Lamniformes: Megachasmidae). Journal of Vertebrate Paleontology 27:512-516.

Thursday, May 10, 2012

Elephant seal vertebra

So this image came up last week when I was searching in google images for "elephant seal vertebra". (From

Yeah, this wasn't quite helpful. Adorable, though. In recent news, I just submitted a manuscript to Palaios on barnacle encrusted sea lion bones, so hopefully that'll plug through peer review just fine and in an orderly time frame.

Saturday, May 5, 2012

Help the VMNH raise funds for an excavation

Hello, everyone - I thought I would quickly mention that my friend Dr. Alton Dooley at the Virginia Museum of Natural History is trying to raise funds for excavations at the Carmel Church Quarry in Virginia. If you can spare a few bucks, check out Dr. Dooley's webpage.

Thursday, May 3, 2012

Fossil vertebrates of New Zealand, part 1: the controversy of the "Cape Kidnappers Fur Seal"

Given my new geographic location, I've decided to start a series of blog posts spotlighting various fossil vertebrates from New Zealand. I'll tackle things like moas, Waipatia, Haast's Eagle, Mauicetus, Kaiwhekea, Kekenodon, and other extinct vertebrates. These won't be in any particular order, and they will be intermittently posted - but I'll cover those over the course of my Ph.D. program here on the south island. To start, I've decided to cover the controversial history of the "Cape Kidnappers Fur Seal".

In 1922, Mr. W. D. Southcott of Hastings, New Zealand, discovered the remains of a fossil pinniped at Cape Kidnappers on the North Island of New Zealand. The fossil included a partial lower jaw, with all premolars and molars, both lower canines (isolated), and a fragment of one of the maxillae ("snout"). The fossil was collected about 20 feet above the beach on the lower bank of a tall cliff. This partial fossil was brought to the attention of the the paleontologist J. Allan Berry. Berry apparently spent quite an effort trying to pinpoint the exact locality and stratigraphic horizon, which he concluded to have been collected from sandstones of Opoitian age, just below the "Black Reef Limestone'. At the time, Berry considered the sandstones to be of early Pleistocene age. In 1928, he published the find in the Transactions of the New Zealand Institute, and named it Arctocephalus caninus. At the time, he considered it to be morphologically closest to the New Zealand Sea Lion, Phocarctos hookeri - which at the time was known to Berry as Arctocephalus hookeri. Although the genus name Phocarctos was named in 1866 - the taxonomy of otariids (fur seals and sea lions) has been volatile in the past, and although there is finally a consensus on what names to use now (see Berta and Churchill 2011), nearly every genus name in existence for fur seals and sea lions has been used for nearly every species, regardless of the names of modern usage. Berry compared the specimen with a number of New Zealand sea lion skulls and jaws, and concluded based on the presence and lack of various cusps on the cheek teeth and the shorter and apparently more robust canines, that it was distinct from Phocarctos hookeri. Additionally, the small size and gracile nature of the jaw suggested to Berry (1928) that the fossil represented a female - but the teeth were so much larger than modern females of Phocarctos, and it must have been a separate species.

The holotype of Arctocephalus caninus. From Berry and King (1970).

Many years later, the Australian pinniped zoologist Dr. Judith E. King examined the holotype specimen of Arctocephalus caninus, after Berry had passed away. She was given access to his notes and unpublished manuscripts, and before his death had apparently come to the conclusion that Arctocephalus caninus was a synonym of Phocarctos hookeri. From her own research experience, King had come to the same conclusion, and in 1970, published an article in 'The Tuatara' where she gave Berry a posthumous first-authorship in recognition of his previous work on the subject. Berry and King (1970) synonymized Arctocephalus caninus with Phocarctos hookeri. They noted that the canines are relatively robust, and actually indicate that the holotype specimen was a male, and not a female; pinnipeds are extremely sexually dimorphic, and the females generally have skinny and shorter canines, while males have more robust canines. Canine size can typically be used to identify the sex of a modern pinniped specimen, and has been used in many cases to identify the sex of a fossil pinniped. However, to my knowledge, there has not yet really been a statistical or morphometric study of canine robustness in pinnipeds, which could certainly be useful. That being said - the fact that Arctocephalus caninus was reinterpreted as a male indicated that the teeth were of an appropriate size to be a male Phocarctos hookeri. Additionally, Berry and King (1970) observed that the cusp development in Phocarctos was so variable that the dental diagnosis for Arctocephalus caninus was unreliable as it fell within the range of variation of modern Phocarctos. This degree of dental variation is common in otariids (Boessenecker, 2011), and is probably due to the fact that otariids (along with other pinnipeds) do not have upper and lower teeth that interlock like terrestrial carnivorans, and thus have undergone a functional release (Miller et al., 2007). Berry and King (1970) also indicated that the Opoitan Stage was of early Pliocene age (5-3 Ma), and thus the south Pacific must have been an early theater for otariid evolution.

Comparison of a modern subadult male Phocarctos hookeri jaw (above) and Arctocephalus caninus (below). From Berry and King (1970).

Three years later, a study published by Weston et al. (1973) contended that the Arctocephalus caninus holotype specimen was not even a fossil. This began with the suspicion by the late Charles Repenning (one of the coauthors and expert on early otariid evolution) that Arctocephalus caninus appeared a little too derived or modern in its anatomy for an early Pliocene sea lion, and they concurred with Berry and King's (1970) reidentification of the fossil. A fully modern sea lion in the early Pliocene really does stand out - and Repenning noted (in Weston et al. 1973) that no Pliocene otariid known at the time (including unnamed fur seals with Repenning was studying, and eventually named Thalassoleon in 1977) exhibited a dentition completely composed of single-rooted teeth. Thalassoleon, and the earlier Pithanotaria (described in the 1920's by Remington Kellogg) all exhibited double rooted cheek teeth. Terrestrial carnivores like dogs, cats, and bears, which pinnipeds evolved from (and are thus a member of the group Carnivora)- exhibit double rooted premolars and triple rooted molars (most have a triple rooted upper fourth premolar as well - which is one of the carnassial teeth, but I digress), so cheek teeth with multiple tooth roots is the primitive condition for pinnipeds. Furthermore - Weston et al. (1973) note that no sea lion fossils assignable to modern genera are known until the middle Pleistocene, worldwide (an assertion that has more or less held true; there may be fossils of modern sea lion genera from the Pliocene of Japan). Repenning was a smart dude, and many of his ideas regarding fossil sea lions and walruses have stood the test of time.

The alleged type locality of Arctocephalus caninus, with the midden at the top of the cliff. 
From Weston et al. (1973).

This skepticism led Weston et al. (1973) to re-reexamine Arctocephalus caninus, and one of the first tests that they did was to do the "flame test" - an old fashioned, tried and true field test to see if a bone is modern or fossil. The idea is that modern bones still retain enough of an organic component that they will smell like shit if you hold a lit match or lighter up to them - for anyone who's ever smelled burning hair - it smells like that. Really, really, really nasty; I've only smelled it once, and that was to provide a control for the flame test - I cut off a tiny bit of my hair and lit it on fire, and damn it smelled bad (and so did the bone in question). To make a long story short, the "Cape Kidnappers Fur Seal" failed the test, and reeked of burning bone smell. Apparently the source is fumes from burning collagen; at the time (we know better now, with collagen being preserved in Cretaceous dinosaur fossils) it was thought that this test indicated an age younger than 10,000 years. It probably does, to a degree- trace amounts of collagen in Cretaceous bones likely would not be sufficient enough to produce a stink, and the test probably indicates abundant collagen remaining in the bone rather than absolute presence or absence. Weston et al. (1973) extracted collagen from bone samples of the Arctocephalus caninus holotype and conducted a thin-layer chromatographic analysis, and compared it with results for a number of other finds of known pre-Holocene age and several other finds that had been radiocarbon dated. Their results indicated that it had an intermediate amount of surviving collagen between a modern bone and a bone from a midden dated at 878 years before present, thus indicating that the Arctocephalus caninus holotype is younger than 1,000 years old and a subfossil. Weston et al. (1973) further indicated that at the discovery site, there is a late Holocene midden with Maori artifacts, and bones and shells from the midden frequently wash down the cliff. The Maori are known to have colonized New Zealand only within the last 800 years or so. Weston et al. (1973) suggested that Arctocephalus caninus is a modern Phocarctos hookeri that washed down from midden deposits and came to rest in loose Pliocene age talus on the side of the exposure, making it appear as though it had genuinely weathered out of Pliocene rocks.

Ironically, when you think about it - at the time of discovery, most holotype specimens of fossils have never been seen before by any human - except in this case, where the evidence points toward an ancient Maori hunter who not only saw the sea lion many hundreds of years ago - but probably killed the animal, long before Mr. Southcott came along and made his own discovery. Also, ironically, I chose to start off this series with something that is most likely not even a fossil. I would have said this at the beginning, but I didn't want to spoil the story.

References -

J. A. Berry. 1928. A new species of fossil Arctocephalus from Cape Kidnappers. Transactions of the New Zealand Institute 59:208-211.

J. A. Berry and J. E. King. 1970. The Identity of the Pliocene Seal from Cape Kidnappers, New Zealand, Previously Known as Arctocephalus caninus. Tuatara 18(1):15-18

Boessenecker, R.W. 2011. New records of the fur seal Callorhinus (Carnivora: Otariidae) from the Plio-Pleistocene Rio Dell Formation of Northern California and comments on otariid dental evolution. Journal of Vertebrate Paleontology 31(2):454-467.

Miller, E. H., H. Sung, V. D. Moulton, G. W. Miller, J. K. Finley, and G. B. Stenson. 2007. Variation and integration of the simple mandibular postcanine dentition in two species of phocid seal. Journal of Mammalogy 88:1325–1334.

Weston, R.J., Repenning, C.A., and Fleming, C.A. 1973. Modern age of supposed Pliocene seal, Arctocephalus caninus Berry (= Phocarctos hookeri Gray), from New Zealand. New Zealand Journal of Science, 16:591-598.