Friday, June 12, 2009

Aquatic Mesozoic Mammals

I have been working on a Mesozoic mammal with some colleagues in Kansas (Michael Engel) and China (Dong Ren) and cannot help but comment on the strangeness that is the world of Mesozoic mammals, in particular those that are deemed "semiaquatic".

The poster-child for aquatic Mesozoic mammal is Castorocauda, the beaver-like docodontan from the Middle Jurassic Daohugou Beds of Liaoning Province in northern China. This is the same locality that has had the earliest gliding mammal, Volaticotherium, as well as some weird feathered dinosaurs like Epidendripteryx, and the mammal I am studying with Engel and Ren (I promise to share more about this critter when the paper is published). Castorocauda is an interesting animal, if anything because it shows few specializations for an aquatic life other than a beaver-like tail, unusual limb proportions, and larger than normal body size for a Jurassic mammal.

Back in 1994, Fred Szalay proposed that stagodontid marsupials, found in the Cretaceous of North America, were aquatic, based on the morphology of the bones in the ankle. In 2005 Nick Longrich presented an abstract at the meeting, Evolution of Aquatic Tetrapods in Akron, OH (hosted by Hans Thewissen at NEOUCOM), detailing how stagodontids might have been durophagous and semiaquatic, based on aspects of their dentition and postcrania, especially a caudal vertebra that resembled those dirsoventrally compressed caudal vertebrae of beavers which are also found in Castorocauda. Recent studies of stagodontids (Fox & Naylor 2006), however, have discredited these claims of being stagodontids as aquatic, although it would be interesting to see if some of Longrich's ideas can be further explored.

So, if semiaquatic mammals are rare in the Mesozoic, why? It has been fairly well documented that being semiaquatic (whatever that means - I'll rant on this some more in the future to be sure) is energetically more costly than being either fully terrestrial or fully aquatic (Williams 1999), so that may have been a hurdle impassible for them, but then why would so many other mammal groups manage it in the Cenozoic even strictly in freshwater, from a variety of body sizes such as desmans to beavers? Mesozoic mammals had been pegged as limited to smaller body sizes in the past, but it is increasingly evident that this was not the case.

I would suspect it is something altogether much less exciting, and much more mundane, expected, and depressing - the fossil record. The "pull of the recent" strikes again, and this time I wouldn't be surprised that because the fossil record of Mesozoic mammals is limited by exposures and the longer periods of time in which fossils may have been destroyed, we simply have fewer of them.

Plus, it is really hard to recognize some of the subtler aspects of adaptations for being semiaquatic in gorups which are still fairly rarely known from anything more than fragmentary teeth and jaws. Hell, if you had a river otter jaw in your hand, would you know it was semiaquatic? No, at least not until we start getting to understand the finer relationships of the skeletal and dental adaptations of aquatic and semiaquatic mammal mammals in a broader context.

Hmmm.... that is a tempting distraction from Mesozoic mammals, isn't it?


  1. All kinds of things to consider there. How aquatic do you have to be to be semiaquatic? Do polar bears count? (I would say yes.) But they don't have any skeletal indicators of their lifestyle that I'm aware of.

    How about elephants? I would probably say they are not semiaquatic, but they are good and willing swimmers. But again, nothing in the skeleton. If an ancient elephant spent say, 25% of it's time in the water, how would we know from the skeleton?

    IIRC, gomphotheres were early arrivers in South America, leading the rest of the North American fauna, and ground sloths were early arrivers in North America. I'd bet that both groups were swimming across Panama before the isthmus formed.

  2. Yes, you make some very good points. I tried to start a discussion about this topic in the facebook group, Aquaic Amniote Paleobiology, but it didn't get off to a great start.

    I think the simplest way to put it is that being "semiaquatic" is really a set of behaviors and anatomy that forms a continuum that is different for every group trying it. I think there are some patterns that can be seen that differ between semiaquatic marine taxa and freshwater taxa (Josh Samuels and I have been talking about that for a while), but the subtlety of those differences might overwhelm the differences specific for individual lineages. Who knows - that's why I keep asking and trying it out.

    Speaking of that, I have a student at NYCOM involved in one of a couple of projects going on about otters that aim at answering some of these questions.

  3. What if only the skull of Castorocauda was known, would it have still be considered semiaquatic? Would the morphology of the molars still be enough?

  4. to Jorge -
    No, I don't think that anyone looking at Castorocauda's skull or teeth alone would ever have guessed it was semiaquatic. And that gets at a major issue with the whole notion of recognizing semiaquatic mammals. Obviously exceptional preservation, like that at Daohugou, allow for better recognition about paleobiology, but what about the remaining 99% of the fossil record?

    In some cases people have been finding ways to get at questions of matters that relate to paleobiology, though rarely do they come to any positive conclusions because this is such a tenuous issue. The best example I can think of is Matt Mihlbachler's studies of population structure in fossil rhinos in comparison with modern rhinos, hippos, and other megaherbivores (Paleobiology 29(3):412-428

    But aside from Matt's study using population biology as a means of recognizing demographics and distributions of fossil animal populations in comparison with modern semiaquatic and terrestrial mammals, very little exists.

    It would be ideal to find an osteological correlate of being semiaquatic, but the only one that (very tenuously) seems to make sense is that of orbit position. I have some morphometric data (using a Microscribe) that I hope to complete in the coming year about desmostylians as compared with hippos and some others, but that's all I can think of. Ironically, despite the cover image on Thewissen and Nummela's book "Sensory Evolution on the Threshold", no chapters detail, or even comment on the notion of orbit position in semiaquatic "transitional" groups. I reviewed the book for Journal of Mammalian Evolution, and plan to post the full version of my review (the unedited, ~12 pages version) on this blog in the near future. It is an excellent book, but there are some curious interpretations and lacking references that did not make it into the final published version for space reasons that I would like to share for the sake of completeness.

  5. I recently came accross your blog and have been reading along. I thought I would leave my first comment. I dont know what to say except that I have enjoyed reading. Nice blog. I will keep visiting this blog very often.