I recently discovered the blog Slime Mold Time Mold, which is largely about the science of obesity — a matter of more than academic interest to me, and if I may say to, to Matt.

I discovered SMTM through its fascinating discussions of scurvy and citrus-fruit taxonomy. But what’s really been absorbing me recently is a series of twenty long, detailed posts under the banner “A Chemical Hunger“, in which the author contests that the principle cause of the modern obesity epidemic is chemically-induced changes to the “lipostat” that tells our bodies what level of mass to maintain.

I highly recommend that you read the first post in this series, “Mysteries“, and see what you think. If you want to read on after that, fine; but even if you stop there, you’ll still have read something fascinating, counter-intuitive, well referenced and (I think) pretty convincing.

Anyway. The post that fascinates me right now is one of the digressions: “Interlude B: The Nutrient Sludge Diet“. In this post, the author tells us about “a 1965 study in which volunteers received all their food from a ‘feeding machine’ that pumped a ‘liquid formula diet’ through a ‘dispensing syringe-type pump which delivers a predetermined volume of formula through the mouthpiece'”, but they were at liberty to choose how many hits of this neutral-tasting sludge they took.

This study had an absolutely sensational outcome: among the participants with healthy body-weight, the amount of nutrient sludge that they chose to feed themselves was almost exactly equal in caloric content to their diets before the experiment. But the grossly obese participants (weighing about 400 lb = 180 kg), chose to feed themselves a tiny proportion of their usual intake — about one tenth — and lost an astonishing amount of weight. All without feeling hunger.

Please do read the Slime Mold Time Mold write-up for the details. But I will let you in right now on the study’s very very significant flaw. The sample-size was two. That is, two obese participants, plus a control-group of two healthy-weight individuals. And clearly whatever conclusion we can draw from a study of that size is merely anecdotal, having no statistical power worth mentioning.

And now we come to the truly astonishing part of this. It seems no-one has tried to replicate this study with a decent-sized sample. The blog says:

If this works, why hasn’t someone replicated it by now? It would be pretty easy to run a RCT where you fed more than five obese people nutrient sludge ad libitum for a couple weeks, so this means either it doesn’t work as described, or it does work and for some reason no one has tried it. Given how huge the rewards for this finding would be, we’re going to go with the “it doesn’t work” explanation.

In a comment, I asked:

OK, I’ll bite. Why hasn’t anyone tried to replicate the astounding and potentially valuable findings of these studies? It beggars belief that it’s not been tried, and multiple times. Do you think it has been tried, but the results weren’t published because they were unimpressive? That would be an appalling waste.

The blog author replied:

Our guess is that it simple hasn’t been tried! Academia likes to pretend that research is one-and-done, and rarely checks things once they’re in the literature. We agree, someone should try to replicate!

I’m sort of at a loss for words here. How can it possibly be that, 58 years after a pilot study that potentially offers a silver bullet to the problem of obesity, no-one has bothered to check whether it works? I mean, the initial study is so old that Revolver hadn’t been released. Yet it seems to have just lain there, unloved, as the Beatles moved on through Sergeant Pepper, the White Album, Abbey Road et al., broke up, pursued their various solo projects, died (50% of the sample) and watched popular music devolve into whatever the heck it is now.

Why aren’t obesity researchers all over this?


Image courtesy of Emma Schachner.

Gotta say, I did not see that coming.

Today sees the publication of a new paper by Emma Schachner and colleagues in Nature, documenting for the first time that unidirectional, flow-through breathing–previously only known in birds and crocodilians–happens in freakin’ monitor lizards. The image above, which is most of Figure 1, pretty much tells the tale.

Some quick background: until the early 1970s, no-one was quite sure how birds breathed. Everyone knew that birds breathe, and that the air sacs had something to do with it, and that the bird lungs are set up as a series of tubes instead of a big array of little sacs, like ours, but the airflow patterns had not been worked out. Then in a series of nifty experiments, Knut Schmidt-Nielsen and his students and colleagues showed that birds have unidirectional airflow through their lungs on both inspiration and expiration. Amazingly, there are no anatomical valves in the lungs or air sacs, and the complex flow patterns are all generated by aerodynamic valving. For loads more information on this, including some cool animations, please see this page (the diagram below is modified from versions on that page). For a short, eminently readable summary of how undirectional airflow in birds was first discovered (among many other fascinating things), I recommend Schmidt-Nielsen’s wonderful little book, How Animals Work.

Avian breathing

After 1972, biologists had almost four decades to get used to the idea that birds had this amazing miraculous lung thingy that was unique in the animal kingdom. Then in 2010, Colleen Farmer and Kent Sanders of the University of Utah blew our collective minds by demonstrating that alligators have unidirectional flow-through lungs, too. That means that far from being a birds-only thing, unidirectional flow-through lung ventilation was probably primitive for Archosauria, and was therefore the default state for non-avian dinosaurs, pterosaurs, the other ornithodirans and the hordes of croc-line archosaurs.

Crocodilian breathing - Schachner et al 2013a fig 10

Diagrammatic and highly simplified representation of airflow through the dorsobronchi and ventrobronchi during inspiration (A) and expiration (B) in the crocodilian lung, and inspiration (A) and expiration (D) in the avian lung. The avian model is a modification of the Hazelhoff loop (Hazelhoff, 1951). Arrows denote direction of airflow, white arrows show air flowing through the parabronchi, blue arrows show air entering the trachea, the red circled “X” demonstrates the location of the aerodynamic inspiratory valve (i.e., air does not flow through this location during inspiration). Colors represent hypothesized homologous regions of the lung in both groups. Abbreviations: d, dorsobronchi; P, parabronchi; Pb, primary bronchus; v, ventrobronchi. [Figure 10 and caption from Schachner et al. 2013a.]

The birdy-ness of crocodilian lungs was further cemented earlier this year when Schachner et al. described the lung morphology and airflow patterns in Nile crocs, which have lungs that are if anything even more birdlike than those of gators. I got to review that paper and blogged about it here.

Now…well, you read the headline. Monitor lizards have unidirectional airflow through their lungs, too. This falls at about the halfway point between “whatisthisIdonteven”–I mean, dude, unidirectional airflow in friggin’ lizards!–and “yeah, that makes a weird sort of sense”. Because to sum up a lot of science unscientifically, monitors just kick a little more ass than other squamates. They have crazy high aerobic capacities for animals that aren’t birds or mammals, they’re ecologically versatile and geographically widespread, they get waaay bigger than any other extant lizards (Komodo dragons) and until recently got even bigger than that (Megalania). Is it going too far to link the success of varanids with their totally pimpin’ flow-through lungs? Maybe, maybe not. But it seems like fertile ground for further study.


Phylogeny for Diapsida showing lungs of representative taxa.
Greyscale images are modified from Milani and transected. The coloured
three-dimensional images are the bronchial tree (right lateral view). Images are
not to scale. a, Diapsida. b, Sphenodon punctatus. c, Crocodile sp. (left) and
Alligator mississippiensis (right). d, Squamata. e, Iguana iguana (left) and
Polychrus marmoratus (right). f, Gekko gecko. g, Lacerta viridis. h, Python sp.
in dorsal view . i, Varanus bengalensis (left) and V. exanthematicus (right).
The blue regions of the phylogeny reflect the hypothesis that unidirectional
airflow evolved convergently; the green arrow shows the alternative hypothesis
of an ancestral origin. [Figure 3 and caption from Schachner et al. (2013b).]

Now, obviously the gigantic question looming over all of amniote biology like one of those monoliths from 2001 is: does this mean that unidirectional flow-through lung ventilation is primitive for all diapsids? That is a super-interesting possibility, and in the new paper Schachner et al. advance some evidence both for and against. On the “for” side, well, hey, there’s uniflow in monitors, crocs, and birds, and in all three cases, air flows down the primary bronchus into a sac at the caudal end, and then back cranially through series of interconnected sacs or tubes. On the “against” side, the patterns of airflow in varanids are similar to those in archosaurs but not identical: in archosaurs, the caudal-to-cranial flow goes through dorsal, tube-shaped secondary bronchi, whereas in varanids it goes through ventrolateral, sac-like bronchi. Also, varanids and archosaurs are phylogenetically distant, so if uniflow was primitive for diapsids, it would seem to have been lost in a lot of other lineages–potentially, all the non-varanid lepidosauromorphs.

On the gripping hand, uniflow would seem to have been lost in all those other lepidosauromorphs, but maybe it wasn’t. Maybe some of them are in the same state varanids were in until this year: they’ve had uniflow lungs forever and we don’t know because no-one has looked yet. And this is one of the concluding points in the new paper: we need to go look more at how living animals actually work.

A small sample of monitor lung diversity, from Becker et al. (1989).

A small sample of monitor lung diversity, from Becker et al. (1989).

In fact, we don’t just need to look at more critters in general, we specifically need to look at more monitors. I have been casually throwing around the terms “monitors” and “varanids” as if the findings of Schachner et al. (2013b) apply to all of them. They may not–the new paper is only about airflow in the savannah monitor, Varanus exanthematicus (same species as Mike’s “sauropod” Charlie), and monitor lungs are sufficiently diverse in form to have been used as taxonomic characters (Becker et al. 1989). So monitors may actually provide multiple windows into the evolution of unidirectional, flow-through lung ventilation. This is especially tantalizing because extant monitors cover a much wider range of body sizes and ecologies than extant crocs, so–just maybe–we can find out if and how diversity in lung structure and ventilation is related to body size and mode of life. Somebody get on that, stat.

Hypothetical bird lung intermediates - Perry 1992 fig 6

Figure 6 from Perry (1992).

My favorite part of all this? Something virtually identical to how monitor lungs work was proposed just over two decades ago by Steve Perry, as a hypothetical stage between saccular lungs and bird-like lungs. See the “Euparkerian grade” lung in the above figure, with perforations between adjacent chambers? Compare that to the diagram of the monitor lung in the image at the top of the post–they’re pretty darned similar. Now, two caveats. First, Steve was suggesting this as a plausible ancestral state for archosaurs, not monitors, and as mentioned above, monitors do things a little differently than archosaurs. Second, there are some things in this figure that are now known to be incorrect, primarily the lack of unidirectional airflow in the crocodilian lung. In fact, on the page opposite this figure, Steve explicitly discounted the possibility of unidirectional airflow in croc lungs. Still, he recognized that croc lungs and bird lungs share profound structural similarities, that they are really points on a spectrum of plausible intermediate conditions, and that crocs had the potential to shuttle air around their lungs because of the complex connections between chambers. So if Steve was not completely right, neither was he completely wrong; it might be most accurate to say that he was less wrong than anyone else at the time, and for about 20 more years after. Which is pretty darned good; I’ve had to rebut myself within the space of five years (Wedel 2007: prosauropod pneumaticity is equivocal. Yates et al. 2012: oh no it’s not!).

Here are the thoughts that have been tumbling through my head since I first learned about this. Obviously structures can be simplified or lost through evolution. Birds and turtles lost their teeth, numerous tetrapods have lost one or both pairs of limbs, and, heck, the platypus lost its stomach. But I rarely see hypotheses of derived simplification entertained for organs like hearts and lungs. There seems to be an unstated but widespread assumption that complex = better when it comes to core physiological processes like breathing.

Reptilian lung morphospace - Perry 1992 fig 2

Figure 2 from Perry (1992)

But it ain’t necessarily so. Following Steve Perry’s diapsid-lung-continuum diagrams, I have often wondered if croc lungs are derived from bird lungs instead of the reverse; maybe the ancestral archosaur had a fully bird-like lung/air-sac system and the non-diverticular, not-super-aerobic lungs of crocs represent a simplification of that system to suit their more sedate lifestyle as semiaquatic ambush predators. That’s pretty much what Seymour et al. (2004) suggested for crocodilian hearts, and it seems plausible given that so many early crocodylomorphs were long-legged, terrestrial, and possibly cursorial (e.g., sphenosuchians). In other words, maybe extant crocs are secondarily ectothermic, with secondarily and possibly paedomorphically reduced air sac systems.

Heck, maybe even bird lungs are simplified compared to their ancestral condition. Most birds have nine air sacs: paired cervical, anterior thoracic, posterior thoracic, and abdominal sacs, and an unpaired clavicular air sac. Some have reduced the number further through loss or fusion of adjacent air sacs. But they all start out with 12 embryonic air sacs (the extras fuse together, IIRC almost all of them becoming part of the clavicular sac), which suggests that the ancestors of birds might have had more than the standard nine.

If we assume that there was some diversity in respiratory anatomy in Mesozoic dinosaurs–which is not much of a stretch, given the diversity we see within (let alone among) monitors, crocs, and birds–it would be an awfully big coincidence if the only dinosaur clade to survive the end Cretaceous extinction just happened to have the fanciest lungs. As far as I know, no-one has proposed that birds survived because they out-breathed everyone else. If anything, the decent-to-high survival rates of mammals, crocs, and turtles across the K-Pg boundary, and the complete extinction of air-sac-equipped pterosaurs and non-avian saurischians, suggests that lung ventilation had nothing to do with survivorship. So what are the chances that crown birds have the most complex lungs among ornithodirans? (Don’t say “flight” because enantiornithines and pterosaurs had air sacs and died out, and bats don’t have air sacs and fly just fine.)

I’m not saying these “awesomeness came first” hypotheses are currently more parsimonious than the standard view. But they’re plausible, and at least potentially testable, and if nothing else an antidote to the idea that birds sit at the top of some physiological Great Chain of Being.

Back to the homology-vs-convergence question. If flow-through lungs are primitive for diapsids, maybe they’ll turn up in a few more critters. But maybe evolving undirectional airflow just isn’t that hard, and only requires poking some holes through the walls of adjacent lung chambers–as stated above, we need to go check more critters. But either way, the form and function of the lungs in V. exanthematicus are not only fascinating in their own right, they give us a window into what the early evolution of archosaurian–and maybe even early diapsid!–breathing might have been like. And that’s phenomenal.

I have some more thoughts on this, particularly the implications for sauropods and other dinosaurs, but those will have to wait for another post.

Images and figures from Schachner et al. (2013b) appear here courtesy of Emma Schachner (website), who kindly offered to let me look under the hood before the paper came out. She also created a cool video showing the 3D lung anatomy of V. exanthematicus. Thanks, Emma, and congratulations!


I happened to be browsing Gerald L. Woods superb Guinness Book of Animal Facts and Feats (3rd edition) this morning, and happened across this fragment on page 76:

Not surprisingly, hummingbirds have the highest energy output per unit of weight of any living warm-blooded animal.

The wording struck me as strange: highest of any living warm-blooded animal? Is Wood just being redundant here, or is he implying that there are cold-blooded animals with a higher mass-specific metabolic rate? The idea seems inherently contradictory, doesn’t it?

I wondered whether he might have insect flight in mind?

Am I making a mistake in conflating “energy output” with “metabolic rate”?


Let’s look at some animals!

February 20, 2012

An important new paper is out:

R. Kent Sanders and Colleen G. Farmer.  2012.  The pulmonary anatomy of Alligator mississippiensis and Its similarity to the avian respiratory system. The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology (advance online publication). doi:10.1002/ar.22427

(It’s paywalled, naturally, but let’s just assume that everyone who reads this blog is affiliated with a big university and has access.)

First of all, congratulations to the authors on doing this properly: publishing a proper paper (sixteen pages) to follow up their big-splash Science paper of just over a year ago.  As Mickey Mortimer has shown, follow-through rates when people publish in Science and Nature are generally not at all good, and it’s always encouraging to see an exception.

Here’s the abstract:

Using gross dissections and computed tomography we studied the lungs of juvenile American alligators (Alligator mississippiensis). Our findings indicate that both the external and internal morphology of the lungs is strikingly similar to the embryonic avian respiratory system (lungs + air sacs). We identified bronchi that we propose are homologous to the avian ventrobronchi (entobronchi), laterobronchi, dorsobronchi (ectobronchi), as well as regions of the lung hypothesized to be homologous to the cervical, interclavicular, anterior thoracic, posterior thoracic, and abdominal air sacs. Furthermore, we suggest that many of the features that alligators and birds share are homologous and that some of these features are important to the aerodynamic valve mechanism and are likely plesiomorphic for Archosauria.

The main reason I want to post this (apart from the fact that it’s an important finding) is because someone had to blog David Marjanovic’s classic response on the Dinosaur Mailing List (quoted with permission, since David doesn’t have his own blog):

See, this is the kind of thing where I’m totally baffled that it wasn’t figured out a hundred years ago, or 120 or 130.

I suppose the logic that has prevented people from dissecting crocodilian lungs for so long went like this:

1) Crocodilians are reptiles.
2) So, crocodilians have reptile lungs, not mammal lungs or bird lungs.
3) What are reptile lungs like? Let’s dissect the nearest reptile and find out!
4) We’re in Europe, so let’s just take the nearest lacertid, perhaps the nearest “colubrid” and maybe the nearest viperid and cut them open.
5) <snip> <snip>
6) Hooray! We’ve figured out what reptile lungs are like!
7) Textbook describes and illustrates generic non-varanid squamate lung as “reptile lung”.
8) Everyone believes it is known what reptile lungs are like.
9) Everyone believes it is known what crocodilian lungs are like, because crocodilians are reptiles.

Ceterum censeo Reptilia esse nomen delendum.

If you must keep the name, follow Joseph Collins and restrict it to Squamata or Lepidosauria. Otherwise, destroy it. Kill it with fire.

So true.

We could draw a whole lot of conclusions from this analysis, but let’s just concentrate on one: look at animals.  See how they behave.  Then cut them open and see what’s inside.  Don’t assume.  Don’t guess.  Find out.  To quote the splendid motto of the Kirkcaldy Engineering Works, “FACTS, NOT OPINIONS”.

Seriously.  Who’d have though there was a Science paper and an Anatomical Record paper just in cutting open an alligator and having a poke around in there?  Sometimes, science doesn’t progress by paradigm shifts; sometimes it progresses just by looking at things.