About the SA Blog Network

Tetrapod Zoology

Tetrapod Zoology

Amphibians, reptiles, birds and mammals - living and extinct
Tetrapod Zoology Home

Everybody loves glassfrogs

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Glassfrogs (or centrolenids) are a really interesting but comparatively little known group of anurans, you might have heard of them. Ha ha, just kidding – you know them well already since they were recently covered at reasonable length here at Tet Zoo. Since that article went live, I’ve been talking with glassfrog expert Juan Manuel Guayasamin, author of Guayasamin et al. (2008, 2009) and so many other published articles. Juan was kind enough to share the several excellent glassfrog photos you see here. They give me an excuse to talk more about these amazing frogs, and also to better illustrate some of the aspects of glassfrog biology and diversity that I mentioned last time.

We start with a brilliant close-up shot of the ventral surface of Hyalinobatrachium aureoguttatum (photo by Martin Bustamante). The Hyalinobatrachium species exhibit transparency across much of their ventral surface (note the transparent/translucent tissues of the limbs), whereas some other glassfrogs are only transparent across part of their undersides (their hands, feet, arms and legs have properly pigmented skin). The peritoneal tissue covering the liver and guts can clearly be seen as bright white in this animal.

Want to learnt more about the basics of anuran anatomy? There are some excellent models (designed for children) out there: this one is made by National Geographic. My son has one. It's large and sturdy and I really recommend it.

If you’re wondering which organs we’re looking at here, the heart is obvious while the white mass located further posteriorly is the intestines. That round yellow organ on the animal’s left side is the gall bladder. The white area between the heart and gall bladder should be the liver, but it’s hard to be sure about this since – in this photo – it’s essentially indistinguishable from the intestines. We tend to think of anurans as simple animals with a very basic anatomy; that’s partly true, but remember that part of the reason for this is that they are heavily and weirdly modified relative to the ancestral tetrapod condition. As I’ve said before on Tet Zoo, a case can be made that they are, in fact, among the most profoundly modified of all tetrapods.

Left hand of Hyalinobatrachium aureoguttatum; photo by Martín Bustamante.

Something I also wanted to mention last time concerns the detailed anatomy of glassfrog hands. Anurans have four fingers. The homology of these has been much-debated. It seems that the true digit I is absent, with the first digit actually being digit II (Fabrezi & Alberch 1996). There’s webbing between all four digits in Hyalinobatrachium, but other glassfrogs only have webbing between the third and fourth digits.

Some glassfrogs have the additional structure present in various anuran hands known as the prepollex, sometimes called the pseudothumb.

The hand skeletons of several glassfrogs, from Guayasamin et al. (2009). Note the variation in prepollex form, as well as the T-shaped terminal phalanges and the distinctive flange on metacarpal III.

There’s been substantial interest in the idea that the prepollex might represent the true digit I – if so, this would mean that at least some anurans still have five digits. Recent embryological studies seem to show that this isn’t the case, and that the prepollex is a wholly novel structure. The anatomy of the prepollex is tremendously variable across anurans (Fabrezi 2001). It’s typically sexually dimorphic, sometimes being enlarged in males, sometimes being spade-like, and sometimes being spiny. We know that prepollices are used in male-male combat across anurans, and some frog species might use them in fending off snakes. You know, there’s a lot to say about anuran hands and digits… maybe I should come back to this subject some time.

Moving on… you’ll recall, I’m sure, the amazing humeral spines present in many glassfrog species. Here’s a calling male Centrolene heloderma, its relatively small humeral spine being visible. He’s calling while clinging to the upper surface of a leaf, a habit typical for Centrolene species. Should he succeed in attracting and mating with a female, she’ll lay her eggs on the leaf underside. The photo is by Jaime Garcia.

Now check out the enormous spines visible in this front view of a male Centrolene geckoideum (below). The photo is by William E. Duellman. I’d really love to know more about the way in which these spines are used in combat, and also about the sorts of injuries they cause: this area of discussion will be familiar to you if you’ve listened to ep 1 of the Tet Zoo podcast. So far as I know, injuries caused by glassfrogs to other glassfrogs have yet to be reported or described (please say if you know otherwise). C. geckoideum was named by the brilliant Spanish biologist Marcos Jiménez de la Espada in 1872 and is the very first glassfrog species to be recognised by science. Jiménez de la Espada probably created the generic name Centrolene by combining the Greek words ‘kentron’ (meaning spike) and ‘olene’ (meaning elbow). The glassfrog Espadarana is named after him (Guayasamin et al. 2009), but it’s also fitting that espada is Spanish for sword.

Those humeral spines are remarkable enough, but there’s also the fact that members of several glassfrog lineages fight by hanging upside-down from the edges of leaves or branches. So far as we know, this behaviour is unique to members of the large glassfrog clade Centroleninae (in Guayasamin et al.’s (2009) phylogeny, this is the group that includes Centrolene, Cochranella and all descendants of their most recent common ancestor: in other words, most glassfrogs except Celsiella and Hyalinobatrachium). Here, we see two individuals of Centrolene lynchi engaging in this behaviour. The photo is by Henry Imba.

As per usual, I ended up saying far more here than I intended to. However, there’s tons more that I could have said. More anuran articles are due to appear here in the near future. My sincere thanks to Juan Manuel Guayasamin for sharing the photos used here, and to the people who took the photos in the first place.

For previous Tet Zoo articles on anurans, see…

Refs – -

Fabrezi, M. 2001. A survey of prepollex and prehallux variation in anuran limbs. Zoological Journal of the Linnean Society 131, 227-248.

- . & Alberch, P. 1996. The carpal elements of anurans. Herpetologica, 52, 188-204.

Guayasamin, J. M., S. Castroviejo-Fisher, J. Ayarzaguena, L. Trueb y C. Vilá. 2008. Phylogenetic relationships of glass frogs (Centrolenidae) based on mitochondrial and nuclear genes. Molecular Phylogenetics and Evolution 48, 574-595.

- ., S. Castroviejo-Fisher, L. Trueb, J. Ayarzagüena, M. Rada, C. Vilá. 2009. Phylogenetic systematics of glassfrogs (Amphibia: Centrolenidae) and their sister taxon Allophryne ruthveniZootaxa 2100, 1-97.

Darren Naish About the Author: Darren Naish is a science writer, technical editor and palaeozoologist (affiliated with the University of Southampton, UK). He mostly works on Cretaceous dinosaurs and pterosaurs but has an avid interest in all things tetrapod. His publications can be downloaded at He has been blogging at Tetrapod Zoology since 2006. Check out the Tet Zoo podcast at! Follow on Twitter @TetZoo.

The views expressed are those of the author and are not necessarily those of Scientific American.

Rights & Permissions

Comments 35 Comments

Add Comment
  1. 1. Heteromeles 5:18 pm 02/16/2013

    They are clearly cool, as is the article.

    One observation: it strikes me as weird that the humerus is so heavily modified, but the femur apparently is not. Is that normal? I’m thinking of the Panda’s thumb mutation being copied in the Panda’s foot bones, and wondering about developmental pathways and such.

    A humerus is a weird place for a spine, too. I’m trying to figure out how one grapples with it. Hmmmm.

    Link to this
  2. 2. David Marjanović 12:10 pm 02/17/2013

    It seems that the true digit I is absent, with the first digit actually being digit II (Fabrezi & Alberch 1996).

    Yep, that’s also what development genetics say. I’ll dig the references up later, and have provided a brief discussion of this in a paper that’s supposed to come out next month in Geodiversitas.

    The fascinating part, you see, is that salamanders have fingers I through IV by the same evidence. Did they have a five-fingered last common ancestor???

    Prepollex/prehallux are continuations of the line radius/tibia–radiale/tibiale–centrale 1. They are indeed not digits (Johanson et al. 2007, ref later).

    A humerus is a weird place for a spine, too. I’m trying to figure out how one grapples with it. Hmmmm.


    Link to this
  3. 3. Dartian 2:11 am 02/18/2013

    this would mean that at least some anurans still have five digits

    For the record, Kermit the Frog has five fingers* (most other Muppets, incidentally, have four fingers). Of course, he’s a felt frog, not a glass frog…

    * At least it looks like he has five fingers. But maybe the fifth digit is actually Kermit’s prepollex?

    Link to this
  4. 4. Jerzy v. 3.0. 4:10 am 02/18/2013

    These “spines” actually look blunt. I wonder if they are useful to grip female during amplexus?

    Link to this
  5. 5. naishd 5:05 am 02/18/2013

    Huh, I always thought Kermit had four fingers, but – you’re right (Dartian, comment 3).. he does indeed have a pentadactyl manus. Maybe members of his lineage modified the prepollex such that it now resembles a ‘true’, semi-opposable pollex.

    Humeral spines (comment 4): yes, they look blunt-tipped in the animals shown here. They might still be sharp-edged and blade-like in those particular taxa, and they really are sharp-tipped in others.


    Link to this
  6. 6. naishd 5:33 am 02/18/2013

    Of relevance to Kermit’s hand is the fact that there are indeed some frogs where the prepollex really does look like a thumb (e.g., the Babina dagger frogs and many others).

    And we should mention the term frame-shift here, just for completeness. I don’t think gene expression has yet been studied in the anuran hand – it would be interesting to know if topographical digit I (embryologic digit II) has or lacks the hox gene expression associated with a true pollex. It does in birds, indicating that their digit I really is a digit I (contra Feduccia).


    Link to this
  7. 7. David Marjanović 9:47 am 02/18/2013

    I don’t think gene expression has yet been studied in the anuran hand –

    It has been. I don’t know if there’s yet a publication other than an SVP meeting poster; I’ll dig the ref for that up.

    It does in birds

    It also does in salamanders. It does not in frogs.

    *spooky music*

    I have no idea what to make of this.

    Link to this
  8. 8. naishd 9:53 am 02/18/2013

    Hmm, thanks, David. Hey, maybe this means that anurans and salamanders are not at all closely related… [smiley]


    Link to this
  9. 9. Tayo Bethel 11:27 pm 02/18/2013

    Are anurans and salamanders closely related? It doesn’t seem very likely …

    Link to this
  10. 10. Jerzy v. 3.0. 3:56 am 02/19/2013

    I think it is related to pteroid and an evidence of surviving descendants of flightless pterosaurs. Or something. ;)

    Actually, I just point to the totally off-topic and cool pictures of Ospreys bitten by fish:

    Link to this
  11. 11. naishd 4:54 am 02/19/2013

    Tayo (comment 9): yes, anurans and salamanders are most likely very close relatives, but there are alternative hypotheses where salamanders form a clade (termed Procera) with caecilians. A world authority on this issue is a regular commenter here at Tet Zoo; he may add more detail.

    Jerzy (comment 10): interesting photos! What I really want to see: one of those mythical giant carp or salmon specimens that has the remains of a dead osprey attached to the skin of its back…


    Link to this
  12. 12. Dartian 6:20 am 02/19/2013

    What I really want to see: one of those mythical giant carp or salmon specimens that has the remains of a dead osprey attached to the skin of its back…

    I think the fish involved in such supposed incidents is most commonly a pike. In some languages, there are even specific words for such pike specimens which mean something like ‘pike with a crown’. ‘Crown’, obviously, refers to the remains of the claws and the feet of an osprey (or, depending on the storyteller, a sea-eagle) stuck in the pike’s back. But yeah, the key word here is ‘mythical’. AFAIK no authentic specimen (involving any fish or bird species) has ever been documented from anywhere in the world – though I’d love to be proven wrong!

    Link to this
  13. 13. naishd 6:26 am 02/19/2013

    Dartian (comment 12) – yes, pikes would make ‘more sense’. I’ve definitely heard the story related for carp and salmon though. Time permitting, will see if I can find them in the literature.


    Link to this
  14. 14. SRPlant 8:59 am 02/19/2013

    Yes, the osprey versus monster fish is perfect for myth, the hubristic apex predator and its nemesis, Captain Ahab being dragged down into the watery depths!

    Link to this
  15. 15. Jerzy v. 3.0. 9:17 am 02/19/2013


    Ospreys drowned by fish are well known and cited in serious sources like “The Birds of the Western Palearctic”.

    I have an old popular animal book with a photo of a carp in a fishing net, with a whole partially rotten Osprey attached to it. The photo is b/w and sharp with no indication of faking. It is certainly a reprint, but I was not able to trace the source. It must have been made in the first half of 20.century and I guessed it might be from Brehm’s Tierleben. I actually once had to scan this photo and send it across the pond to the person making a book on animal remains who wanted to see that such a thing exists!

    The book is in my parent’s home, so for now you must believe it. :)

    Link to this
  16. 16. greg_t_laden 9:56 am 02/19/2013

    So, one idea is that the osprey talons get jammed in with the fish scales (or other bits) and they can’t let go. See this:;f=24;t=001418;p=0

    As unlikely as that seems, the possibility that there really is a photograph of a fish in a net with an osprey clinging to it really can’t be assumed to be zero. Stranger things have happened.

    I’m still waiting to see the photograph, though. It would need to be accompanied by a written report of a forensic analysis by a recognized expert.

    Link to this
  17. 17. Halbred 5:54 pm 02/19/2013

    Darren, I’d love an article about tadpoles–specifically the morphological journey from tadpole to adult and how the tadpole stage evolved and what advantages it confers. AFAIK, other modern amphibians don’t go through a tadpole stage. Did any fossil groups?

    Link to this
  18. 18. souhjiro 9:37 pm 02/19/2013

    Seconded, the anuran tadpole is such a weirdosity,the closest thing to a caterpillar among vertebrates

    Link to this
  19. 19. naishd 6:09 am 02/20/2013

    I’ve been meaning to write a review of tadpole diversity and morphology for a while – tadpoles are incredible, but their developmental plasticity (and what effect this has on the metamorphosed stage) is too. I’m also really taken with the idea that selective pressures on tadpoles have shaped the anatomy of adult anurans – that the weird, tailless, short-bodied anatomy of frogs and toads is a ‘default’, and not necessarily adaptive. Very controversial though! Ok, more on all of this some other time…


    Link to this
  20. 20. David Marjanović 11:07 am 02/20/2013

    Caudata (salamander total group) and Salientia (frog total group) are clearly sister-groups, judging from morphological as well as molecular evidence. There is support for the alternative (Caudata and Gymnophionomorpha as sister-groups), mostly from molecular evidence, but it’s quite a bit weaker.

    What we need are more albanerpetontids, and I don’t mean the usual isolated fused frontals. They’re almost certainly not caudates as originally thought (also Anderson 2007, Anderson et al. 2008, Maddin et al. 2012 – different versions of the same data matrix with the usual amount of mistakes in it); and nobody has ever proposed that they’re salientians; but everything else is an option. This includes them being the sister-group of Lissamphibia as a whole.

    Also, Rainer Schoch is said to be redescribing *Triassurus*, which may be the only known Triassic salamander or just a temnospondyl larva. Can’t wait.

    AFAIK, other modern amphibians don’t go through a tadpole stage. Did any fossil groups?

    Nope. And what’s worse, stem-group frog larvae are entirely unknown so far, so there’s no direct evidence on how the tadpole stage evolved. Salamanders and caecilians (and, to a lesser extent, amphibamid temnospondyls) do have a distinct metamorphosis, but the larval stage is much less modified than in frogs.

    the closest thing to a caterpillar among vertebrates

    Exactly! A very early ontogenetic stage elaborated into a feeding stage, with all stages between it and the adult shape squeezed into metamorphosis!

    that the weird, tailless, short-bodied anatomy of frogs and toads is a ‘default’, and not necessarily adaptive

    Exhibit A: any aquatic frog ever. Not one of them retains the tail or the gills or even just a single pair of gill slits. They’re all hindlimb-propelled swimmers. Somewhere in the frog stem-group, sexual maturity became part of the later stages of metamorphosis, then it ended up being triggered by them, and thus frogs who don’t completely metamorphose can’t reproduce and win a Darwin Award. Evolution: the exploration of every cul-de-sac there is.

    Link to this
  21. 21. David Marjanović 11:08 am 02/20/2013

    …Really interesting that I didn’t put Triassurus into italics. ~:-|

    Link to this
  22. 22. naishd 11:29 am 02/20/2013

    Aquatic frogs are clearly a demonstration of intelligent design. After all, we’re dealing with aquatic animals that do away entirely with a superb swimming organ and end up with a body plan better suited for terrestrial locomotion. Brilliant.


    Link to this
  23. 23. David Marjanović 1:51 pm 02/20/2013

    Oh, and, there are megophryid tadpoles that ossify a few vertebrae in the tail that they then proceed to lose. Otherwise, tadpole tails have only the notochord as their skeleton, AFAIK.

    I’m correcting the page proofs for my next paper* right now. This reminds me of the fact that the Procera hypothesis has never been found by an analysis that included albanerpetontids.

    * With Michel Laurin. Geodiversitas 35(1), due to appear next month.

    Link to this
  24. 24. Andreas Johansson 3:41 pm 02/20/2013

    The “Procera hypothesis” appears to refer to the idea that caecilians and salamanders are sister groups. Googling this was less than entirely straightforward because “Procera” appears also the be the name of some sort of vitamin supplement.

    Link to this
  25. 25. naishd 4:44 pm 02/20/2013

    I’ve found exactly the same thing when googling Procera – there’s also a California-based communications company that has pretty strong web presence. It also didn’t help that, for a while, I mis-remembered the group name as Procoela…


    Link to this
  26. 26. David Marjanović 5:55 am 02/21/2013

    The “Procera hypothesis” appears to refer to the idea that caecilians and salamanders are sister groups.

    Yes, as mentioned in comment 11.

    Link to this
  27. 27. vdinets 8:11 am 02/21/2013

    I wonder how this obligatory metamorphosis restriction could have evolved. It’s tempting to see it as an adaptation to ephemeral environments such as desert ponds or vernal pools, where you have to secure an escape option (by growing legs) before you even think about breeding.

    If you consider Leiopelmatides to be the most primitive frogs, this theory doesn’t look promising, because they live in permanently moist environments (and show a bizarre diversity of breeding strategies within just two genera). But if you consider Rhinophrynus to be the earliest split among extant taxa, the idea seems more likely.

    I wonder how many frogs would go neotenic if this restriction was not in place.

    Link to this
  28. 28. David Marjanović 9:20 am 02/21/2013

    If you consider Leiopelmatides to be the most primitive frogs

    And why would you do that? They’re the sister-group to all other extant frogs, but it’s not like no stem-frogs were known. :-)

    I don’t know much about the environments of Triado- or Czatkobatrachus. But Prosalirus lived in desert oases…

    Rhinophrynus is a pipoid, nested well within the crown-group of frogs – though the earliest known rhinophrynid, the Late Jurassic Rhadinosteus (from the Morrison Fm), is the earliest known definite crown-frog.

    I wonder how many frogs would go neotenic if this restriction was not in place.

    Basically all the pipimorphs ever?

    Link to this
  29. 29. David Marjanović 11:04 am 02/21/2013

    Oopsie. Enneabatrachus from Quarry Nine in the Morrison Fm is also apparently a crown-group frog (thought to be a costatan/bombinatoroid/alytoid/discoglossoid).

    Link to this
  30. 30. vdinets 11:45 am 02/21/2013

    And why would you do that?
    Because, like most zoologists, I am pathetically ignorant about much of paleontology, so I was talking about extant taxa.

    As for Rhinophrynus, EDGE website claims that it is “the most evolutionary distinct” extant Anuran:

    But why would you call pipimorphs neotenic? They don’t look like tadpoles to me, except maybe their facial expressions…

    Link to this
  31. 31. vdinets 11:47 am 02/21/2013

    Oh, sorry, you mean they would go neotenic? Yes, that makes sense. Would be fun if they eventually evolved a second metamorphosis and could go back to tadpole-like stage.

    Link to this
  32. 32. David Marjanović 1:19 pm 02/21/2013

    As for Rhinophrynus, EDGE website claims that it is “the most evolutionary distinct” extant Anuran:

    With this I agree. Of course this doesn’t say whether it’s the least or the most autapomorphic one (or anywhere in between), just that it separated from its closest extant relatives (Pipidae) in the Late Jurassic or perhaps earlier.

    (The leiopelmatids sensu lato/amphicoelans have almost no fossil record. The oldest known fossil is an isolated ilium from the end-Cretaceous Lance Formation in Wyoming, said to be indistinguishable from Leiopelma. As far as biogeography is concerned, Ascaphus and Leiopelma should have diverged earlier than that, but not necessarily much earlier; the middle of the Late Cretaceous is when New Zealand broke off and floated away.)

    you mean they would go neotenic?


    Link to this
  33. 33. vdinets 3:06 pm 02/21/2013

    Are there any Mesozoic reptiles that could be specialized frog- or tadpole-eaters? Is there anything in dentition or neck anatomy that would indicate that, or would it look the same as in fish-eaters? Of modern birds, Ciconia storks are probably the most frog-specialized ones, unless I am forgetting something. But there’s plenty of frog-eating snakes.

    Link to this
  34. 34. naishd 3:45 pm 02/21/2013

    With reference to comment 33… to my knowledge, there are no Mesozoic animals with dental adaptations that look at all like frog-eating specialisations. Not yet, anyway. No goo-eaters or mollusc-eaters, either.


    Link to this
  35. 35. David Marjanović 7:54 pm 02/22/2013

    Is there anything in dentition or neck anatomy that would indicate that

    Would surprise me. And storks are hardly specialized.


    Well, placodonts (and pycnodontiform actinopterygians), but nothing terrestrial. There’s the alligatoroid Brachychampsa near the end of the Cretaceous, but that may have been more of a turtle specialist than a mollusc specialist.

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article