PvM at the Panda’s Thumb has already written a bit about this issue in the article “Human Gland Probably Evolved From Gills”, but I’m not going to let the fact that I’m late to the party stop me from having fun with it. This is just such a darned pretty story that reveals how deeply vertebrate similarities run, using multiple lines of evidence.

Here’s the start of the situation: fish have a problem. Like most animals, they need to maintain a specific internal salt concentration, but they are immersed in a solution that is much more dilute they they are (for freshwater fish) or much more concentrated (for saltwater fish). To make matters worse, fish have large respiratory membranes, the gills, which expose a huge amount of surface area to the watery medium. Their answer to the problem is to do much of their regulation of various ions at the gill surface, using sensors and salt pumps to constantly work, extracting excess salts from one side and pumping them to the other. Gills are therefore more than just the organ fish use to breathe—it’s also where they regulate salt balance.

Us terrestrial tetrapods have a different problem. We can’t pump salts out of or into the air, so instead we maintain internal stores of salts (calcium, for instance, is packed into bone) and use hormones to regulate them, by telling cells to either sequester the ions, or to release them into the bloodstream. We don’t do this with gills, obviously—instead, we have several glands that monitor and control blood salt levels.

One of the most important sets of glands for this function are the thyroid and parathyroid glands, which regulate calcium ion balance. When the amount of calcium salts dissolved in the blood drops below a certain level, receptors called CasRs (Calcium sensing Receptors) detect the change, and trigger the parathyroid gland to release PTH (Parathyroid Hormone) into the blood, which tells cells in the bone to gnaw away and free up calcium (which can lead to osteoporosis, for instance), and it also signals the kidneys to retain calcium. The thyroid, which I’m not going to say much about, has a complementary role in managing the situation when blood calcium levels are too high.


So fish have gills to regulate calcium, and we have parathyroid/thyroid glands to regulate calcium. Do they have anything else in common? One minor thing seems to be location: gills are in the fish’s ‘neck’, and the parathyroid glands are located at roughly the same place, in your neck—and that’s interestingly coincidental, since there’s no particular need for these glands to be in that particular location. They can do their job just as well anywhere. And they happen to be located in a particularly fascinating place for those of us curious about vertebrate evo-devo; all kinds of action occurs in this pharyngeal region in early development.

For one thing, many features of the face and neck are assembled from those curious and fateful tissues, the branchial arches, well known for their homology with the gill arches of fish. Here’s a diagram of a 5-week-old embryo on the left, with the arches color-coded, and an infant on the right, showing what connective tissues develop from each arch.


The first arch contributes to the jaw, while the second through fourth are going to form various cartilages in the neck—the cartilages near where the thyroid/parathyroid will form. What an interesting coincidence! Could it be that the parathyroid is also derived from a branchial arch, and is therefore homologous with the gills of fish?

One way to find out is to use an early molecular marker. The parathyroid is distinguished by the expression of a unique transcription factor, Gcm-2. Gcm-2 is expressed only in the parathyroid, and is absolutely critical to its formation: knocking out the gene causes the parathyroid to fail to form. Another useful marker is to stain for CasR, the receptor protein, or PTH, the hormone itself. The top row of pictures (A-D) below show that these are effective markers for the thyroid gland in an older chick embryo. The lower series (E-I) are in younger embryos, at a stage when they have clear pharyngeal arches, stained with Gcm-2.

Expression of Gcm-2 in the parathyroid gland and the pharyngeal pouches in the chick. (A#D) Whole-mount in situ hybridization of chick E11 thyroid (T) and parathyroid (P) glands for the following probes: Gcm-2 (A), PTH (B), CasR (C), and TTF-1 (D). Gcm-2 can be seen to be expressed in two round masses, the parathyroid glands, which are adjacent to the thyroid, which expresses TTF-1. The parathyroid glands additionally express PTH and CasR. (E-I) Expression of Gcm-2 in chicken embryos, staged as described in ref. 16. In these micrographs, anterior is to the left and ventral is to the bottom. OV, otic vesicle; pp, pharyngeal pouch; II-IV, pharyngeal arches. Expression of Gcm-2 starts in the third pharyngeal pouch at stage 18 (E), and then, as development proceeds, expression is also evident in the fourth pharyngeal pouch and additionally weakly in the second pouch (F). At stage 24, expression in the third and fourth pouches is concentrated in a region dorsal of the pharyngeal pouches and is lost from the second pouch (G). In Vibratome sections of stage-22 embryos (H), it is clear that Gcm-2 expression is localized to the pharyngeal endoderm and that, by stage 24, the region of the pharyngeal endoderm expressing Gcm-2 has thickened and given rise to round masses that are the parathyroid gland rudiments of the third and fourth pouches (I).

What you see is that this marker appears early in the third and fourth pharyngeal arches, showing that the parathyroid is derived from the same tissues as fish gills.

This is cool enough, but Okabe and Graham take it a step further. Gcm-2 is a marker for the parathyroid in tetrapods. Fish lack a parathryoid gland, but is Gcm-2 expressed anywhere in them? I think you can guess where Gcm-2 is expressed in fish:

Phylogenetic analysis of the distribution of Gcm-2 and its expression in teleost (zebrafish) and chondrichthyan (dogfish) species. (C-F) Gcm-2 expression in zebrafish embryos. In these micrographs, anterior is to the left and ventral is to the bottom. Gcm-2 initiates expression in the second pharyngeal pouch in early 3-day-old larval fish (indicated by arrowhead in C). Subsequently, Gcm-2 is expressed sequentially in the more posterior pouches (D), and, by day 4, Gcm-2 is expressed in all of the pouches (E). It is also apparent by day 4 that Gcm-2 is expressed in the developing internal gill buds emerging from the pharyngeal pouches (F). (G and H) Gcm-2 expression in dogfish embryos. This gene is expressed in the internal gill buds protruding from the pharyngeal pouches in stage-27 dogfish embryos (17). The pharyngeal arches are numbered II-VI.

These photographs show where Gcm-2 is expressed in zebrafish and dogfish embryos: the pharyngeal arches! These are beautifully symmetrical results showing that the tetrapod parathyroid is derived from the pharyngeal arches, and that the pharyngeal arches of fish also express a parathyroid marker. In addition, the zebrafish pharyngeal arches also express PTH and CasR. And the researchers take it a step further still.

Remember that Gcm-2 is essential for parathyroid formation: mutate it, and the animal doesn’t form a parathyroid. What if we knock out Gcm-2 in a fish? It doesn’t have a parathyroid, after all…all it has are gills.

Gcm-2 is required for the elaboration of the internal gill buds from the pharyngeal pouches in zebrafish. Zebrafish embryos were injected at the one-cell stage with either control or antisense Gcm-2 MOs. The embryos were then analyzed at day 5 for the presence of internal gill buds. (A-C) Five-day-old zebrafish larva injected with control MO. (A) Nomarski viewof the pharyngeal region of a day-5 embryo injected with the control MO. The internal gill buds protruding fromthe pharyngeal pouches are clearly evident (arrowheads). (B) Embryo injected with control MO hybridized for Gcm-2. Gcm-2-expressing internal gill buds can be clearly seen protruding from the pharyngeal pouches. (C) Embryo injected with control MO, showing normal pharyngeal pouch formation as judged by Pax-9a expression. Each pharyngeal pouch is indicated by an arrowhead. (D-F) Five-day-old zebrafish larva injected with Gcm-2 antisense MO. (D) Nomarski view of the pharyngeal region of a E5 embryo injected with the antisense Gcm-2 MO. There are no internal gill buds protruding fromthe pharyngeal pouches. (E) Embryo injected with the antisense Gcm-2 MO hybridized for Gcm-2. There are no Gcm-2-expressing internal gill buds protruding from the pharyngeal pouches. (F) Embryo injected with the antisense Gcm-2 MO, showing normal pharyngeal pouch formation as judged by Pax-9a expression. Each pharyngeal pouch is indicated by an arrowhead. EY, eye; YK, yolk. Anterior is to left and ventral is to the bottom.

In zebrafish injected with a morpholino that blocks the Gcm-2 transcription factor, poof, gills fail to form.

So, what we have here are multiple lines of evidence—location, function, several molecular markers, and developmental origins and processes—that converge to show that parathyroid glands and the gills of fishes have a common evolutionary origin.

Okabe M, Graham A (2004) The origin of the parathyroid gland. PNAS 101(51).


  1. #1 John
    June 21, 2006

    This is wonderful stuff, and exactly the sort of elegant scientific account of mechanisms that you will not find at a place like Uncommon Descent.

  2. #2 Ed Darrell
    June 21, 2006

    I hope you’ll take the time to explain the vagus nerve in a similar fashion — comparing humans with giraffes, certainly, as the most egregious example of “oops” architecture, but also comparing sharks or other primitive fishes.

  3. #3 Jane Shevtsov
    June 21, 2006

    Beautiful! And thanks for keeping the terminology to a minimum. I wish my vert bio classes had been like that.

  4. #4 Greg Peterson
    June 21, 2006

    Speaking of the vagus nerve, I was listening to the dreaded Christer talk radio station that PZ referred to in any earlier post, and the host asked the “guest,” a science proponent, for what he considered the best evidence for evolution. The guest thought and replied that he considered the genetic similarities between chimps and humans to be perhaps the best evidence. Not a bad response, I thought. The host responded to this with a smarmy, ridiculous, and obviously canned bit about how clouds are mostly water and watermelons are mostly water, but that doesn’t mean that clouds become watermelons. My immediate reaction was, first, that’s a terrible analogy (as in, not even an analogy–I thought the way most of the Gospel of Mark appears in the gospels of Luke and Matthew was a closer analogy, with some “mutations,” was a closer analogy as far as that went). And second, actually, even as a bad analogy it, if anything, helps support evolutionary theory. The water could be thought of as like genetic information that flowed from the cloud (chimp) into the watermelon (human). OK, I SAID it was a flawed analogy, and maybe not amenable to rehabilitation–but it wasn’t the slam-dunk counter-evidence the host seemed to think it was, either. But the whole exchange got me thinking about what I would say if asked that, and I like the vagus nerve story. It makes perfect sense from an evolutionary standpoint, but no competent engineer would ever create a nerve intended to go from the brain to the throat loop all the way down to the aorta like that (about 14 pointless feet in a giraffe, as the previous poster alluded to). It’s hard to imagine a more obvious relic of contingency and necessity than how the vagus nerve is positioned in mammals. Of course, these creationist boobs can always come back with something about how their god did it that way for some reason of his own that’s just not obvious to us. To which the proper reply might be that a “theory” that can explain everything explains nothing.

  5. #5 Kristine
    June 21, 2006

    This is cool stuff. I’ve been enjoying The Ancestor’s Tale, but at times it is helpful to supplement that reading with information that compares other species to humans with this level of detail, just so that I can get a clearer picture of the processes and the differing concestors themselves. (I’m not always clear about what is an adaptation, and what is the result of something else, like genetic drift, etc.)

    Can that Christer radio host explain why we, rather than any of the (other) apes, are the chimps’ closest relative?

  6. #6 steve s
    June 21, 2006

    This is wonderful stuff, and exactly the sort of elegant scientific account of mechanisms that you will not find at a place like Uncommon Descent.

    That’s actually the first thing I thought, when I saw this page. It’s good to be on the pro-science side.

  7. #7 Jethro Gulner
    June 21, 2006

    Not a bad response, I thought. The host responded to this with a smarmy, ridiculous, and obviously canned bit about how clouds are mostly water and watermelons are mostly water, but that doesn’t mean that clouds become watermelons.

    Melons and Watermelons are both mostly water, but they are probably not as closely related as humans and chimpanzees.

  8. #8 Scott Hatfield
    June 22, 2006

    PZ, Scott Hatfield here. Just wanted to commend you for a tremendously helpful piece of writing. I would love to be able to reproduce parts of it in either a Power Point or a photocopied adaptation. Stuff like this really separates Pharyngula from a lot of other stuff on the Web. Well done!

    Scott Hatfield

  9. #9 Torbjörn Larsson
    June 22, 2006

    The vagus nerve connected to the tongue?! So when I order food it *is* my stomach that speaks. It’s evolution – always expect the unexpected.

    Umm, that explains why, when a buddy managed to fool me to drop 1.5 m onto my butt in my youth (he jumped of the seesaw) I was rendered momentarily breath- and speachless. I later thought it was merely that too much air was momentarily forced out of the lungs since there is a requirement on residual volume. But if the vagus nerve controls both lungs and tounge, that was probably it.

    Breathing returned after a few airless seconds. Oh yes, I kicked his butt back. 🙂

  10. #10 MNDarwinist
    June 22, 2006

    Excellent work, PZ. I am waiting impatiently for Ann Coulter’s rebuttal.
    By the way, dealing with human brain as I do, I think if god where a human engineer, he would get sued for incompetence. Would you ever comment on how all of the spinal fluid has to flow through the hair-thin aquedut of Sylvius? Very unintelligent design indeed.

  11. #11 naeem raza
    August 22, 2006

    sir im student of mbbs & i need ur help in my study

  12. #12 Haq
    November 30, 2006

    its really good stuff, after reading this article i came to know many things that i thought only humans have, other living things also have but thing is that in them it’s playing different role.

  13. #13 kgoldberg
    April 14, 2008


  14. #14 dr_karen
    June 14, 2008

    you did not include the eustasian tube in the diagram why?

New comments have been temporarily disabled. Please check back soon.