Bamboo shark walks the walk

by Matthew Cobb

This video of a bamboo shark (Hemiscyllium halmahera) strolling across some coral was posted on Twitter by Kyle Hill (@sci_phile):

This oviparous species, which is about 70 cm long, has just been described by Dr Gerald Allen of the Western Australian Museum. For the shark-heads out there, the abstract of the paper says:

Hemiscyllium halmahera new species is described from two specimens, 656-681 mm TL, collected at Ternate, Halmahera, Indonesia. The new species is clearly differentiated on the basis of colour pattern. Its features include a general brown colouration with numerous clusters of mainly 2-3 dark polygonal spots, widely scattered white spots in the matrix between dark clusters, relatively few (< 10), large dark spots on the interorbital/snout region, a pair of large dark marks on the ventral surface of the head, and a fragmented post-cephalic mark consisting of a large U-shaped dark spot with a more or less continuous white margin on the lower half, followed by a vertical row of three, smaller clusters of 2-3 polygonal dark marks. The new species is most similar in general appearance to H. galei from Cenderawasih Bay, West Papua, which differs in having 7-8 large, horizontally elongate dark spots on the lower side between the abdomen and caudal-fin base, a cluster of solid dark post-cephalic spots, and usually about 25 dark spots on the upper surface of the head.

But the really interesting thing to me is the gait the fish is using – this is apparently typical of Hemiscyllium sharks, which prowl across coral looking for food. It is using the classic alternate movement of a tetrapod.

We’ve previously talked about the evolution of the tetrapod gait in lungfish. Now you are more closely related to Tiktaalik, the famous fish/tetrapod intermediate form, than Tiktaalik was to a shark (because you and I and Tiktaalik all have bones; sharks do not have bones, only cartilage).

So this suggests that the neuronal control of the way that you run (your right arm moves with your left leg, and your left arm moves with your right leg – try it) goes waaaayyyy back even beyond our fishy ancestors, to the time before the evolution of bone.

Another alternative is that this is convergent evolution – if you are going to ‘walk’, the alternate gait is the best way of doing it. Today’s question: How could we test between these two hypotheses?

I may be exaggerating the importance of this gait in sharks – are there any locomotion experts out there who can comment?

The film was made near Ternate, on the Malaku islands in Indonesia, by Mark Erdmann of wedaresort.

Allen GR et al. 2013. Hemiscyllium halmahera, a new species of Bamboo Shark (Hemiscylliidae) from Indonesia. aqua, International Journal of Ichthyology, 19 (3): 123-136


  1. Griff
    Posted August 27, 2013 at 2:05 am | Permalink

    The walk is so reminiscent of a crocs.

    • Posted August 27, 2013 at 4:07 am | Permalink


      It appears to me to be the same kind of side-to-side body movement that “fish” swim with.

      Don’t we see that same kind of body movement in crocs and lizards … but not in mammals? (And why not?)


      • Diana MacPherson
        Posted August 27, 2013 at 5:09 am | Permalink

        Crocs & other lizards have their legs more on the sides instead of directly under them like a lot of mammals (and dinosaurs). Crocs can put their legs under them a bit more when they really motor (but you don’t want to be in front of the croc when that happens :).

        Reptiles are so awesome.

        • Posted August 27, 2013 at 5:29 am | Permalink

          That ties in with Michael’s comment (#5) that was supposed to go here.

          So, dinos branched off from lizards (276 Mya; sparrow v. Gekko) later than mammals (296 Mya; Homo v. Gekko), which suggests that legs under the body/vertical articulation of the spine evolved twice… 


          PS. “Crocs & other [sic] lizards… ”: Crocs aren’t lizards! 

          • Diana MacPherson
            Posted August 27, 2013 at 5:43 am | Permalink

            Oh noes, I’ve been sic’d 😀 Oops, I meant reptiles. I do that all the time and conflate the two because I have a certain animal in my head and apply the wrong word. In my defence, I have tummy troubles today & hadn’t finished my morning coffee (which probably doesn’t help the tummy troubles) 😀

            • Posted August 27, 2013 at 1:21 pm | Permalink

              I have a certain animal in my head

              It could be the reptilian brain acting up again…

              • Diana MacPherson
                Posted August 27, 2013 at 1:54 pm | Permalink

                Yes, that would explain the headaches!

  2. Posted August 27, 2013 at 3:31 am | Permalink

    There are two notable exceptions, as far as I know – the camels (including dromedaries) and the giraffe who move the two legs of each side in unison (pacing gait) and not in cross-crawling like all other mammals and walking vertebrates… and this shark.

    • Posted August 27, 2013 at 4:20 am | Permalink

      Oh, and there is also the cheetah:

      watch?v=XYYU1XNupj0 (YouTube)

      As well as the hyena:

      watch?v=Mvsv4OdcT3g (YouTube)

      • Garnetstar
        Posted August 27, 2013 at 7:25 am | Permalink

        Also, cats! They walk with a pacing gait, but when they speed up to a trot, they switch to an alternate gait.

        So, they use the best of both: wouldn’t you just know they would?

    • Posted August 27, 2013 at 7:47 am | Permalink

      Many breeds of horse (typically called gaited horses) do the same, producing a much smoother ride. In the era when horse riding was a major means of individual transport, riding a non-gaited horse was sometimes seen as somewhat infra dig.

    • Notagod
      Posted August 27, 2013 at 7:55 am | Permalink

      Bunnies, roos?

    • lkr
      Posted August 27, 2013 at 10:57 am | Permalink

      “Pacing” in giraffes and camels looks to be a matter of the ratio of leg/stride length to the distance between shoulder and hip. Try to trot/gallop and they would be apt to trip over their own feet.

      My Manx cat is near the tipping point for this relationship: she [and Manxes in general] has a short torso and relatively long legs, particularly the hindlimbs. She moves to a pacing gait at a fast walk, maintains that up to a transition to a bounding gallop that is almost rabbit-hop-like. [My wife, seeing her wobble side-to-side, initially thought there was “something wrong”, but it’s pretty clear that what she does works for her.]

      Presumably in the cat, and in many other mammals, there are several possible quadripedal gaits, and the optimal one is learned by individuals.

  3. Posted August 27, 2013 at 4:22 am | Permalink

    So this suggests that the neuronal control of the way that you run (your right arm moves with your left leg, and your left arm moves with your right leg – try it) goes waaaayyyy back even beyond our fishy ancestors, to the time before the evolution of bone.

    Count me as skeptical. For us, this is more likely learned behavior.

    It is not hard to walk the other way — moving the left arm with the left leg and the right arm with the right leg. But the balance is poor due to the changing center of gravity. Trial and error learning would quickly get us to the way we normally walk. And balance is more important for running.

  4. Robert Secatore
    Posted August 27, 2013 at 4:28 am | Permalink

    To determine whether or not it’s a possible instance of convergent evolution, see if more primitive taxa or taxa thought to be ancestral to this species also have alternate movement . If not, then clearly this type of movement arose independently in both sharks and bony fish and it’s an example of convergent evolution.

    • Matt G
      Posted August 27, 2013 at 6:08 am | Permalink

      People should check out Salamandra robotica. The side-to-side gait is the same as the movement of the spine in swimming.

  5. Michael Fisher
    Posted August 27, 2013 at 4:42 am | Permalink

    My best guess re running on land with four legs…

    Feet tucked underneath the centre of gravity wastes less energy in side-to-side motions ~ leading to faster speed & increased manoeuvrability. But animals using this format are obliged to articulate the spine in the vertical plane so that the front & back pair of paws/hooves/feet can move past each other without interference.

    • Michael Fisher
      Posted August 27, 2013 at 4:43 am | Permalink

      The above was responding to Ant @ comment 1. What happened there? 🙂

    • Posted August 27, 2013 at 5:14 am | Permalink

      Ah, that question (mammalian articulation of the spine in the vertical plane) was at the back of my mind. Which explains why cetaceans swim with up-and-down tail motions, unlike “fish”.

      Does this also explain camels’ and giraffes’ walking gait (#2), that they don’t have enough vertical articulation? (Because of (c) the hump(s) [are llamas “normal”?] and (g) the short length of the body relative to the length of the legs.)


      • Michael Fisher
        Posted August 27, 2013 at 5:30 am | Permalink

        Camel/Giraffe pacing walk allows fore & hind feet to be kept well out of each other’s way. Some long-legged dogs pace too.

        BUT… Camels/Giraffes do gallop when necessary ~ I assume at the risk of colliding front & back feet. I must watch some videos & see if the gallop is different in such animals…

      • Michael Fisher
        Posted August 27, 2013 at 5:42 am | Permalink

        THIS GALLOPING GIRAFFE is interesting.

        The rear pair of feet hardly impinge on the front pair at all & the rear legs are slightly knock-kneed [to my eye] to sweep the rear feet well out of line with the front pair.

        Rear legs are kept stiff as if in plaster casts 🙂
        It reminds me of someone trying to move rapidly on crutches, with the crutches being the rear legs/feet

  6. marksolock
    Posted August 27, 2013 at 4:48 am | Permalink

    Reblogged this on Mark Solock Blog.

  7. teacupoftheapocalypse
    Posted August 27, 2013 at 5:07 am | Permalink


  8. Diana MacPherson
    Posted August 27, 2013 at 5:10 am | Permalink

    It probably makes sense with the shark’s body type. Snakes move like that too (they just don’t have legs).

  9. Kevin
    Posted August 27, 2013 at 6:01 am | Permalink

    I would suggest that walking might have originated with just the backbone wiggle as in swimming which would cause an alternating gait though with the limb moving passively. Possibly this would later lead to more refined neural control of limbs. The limb control of this type would then evolve into the crossed-extensor reflex that we are familiar with. Preumably the limbs in any case would have the normal neural control that they would have had for normal swimming

  10. Desnes Diev
    Posted August 27, 2013 at 8:42 am | Permalink

    Notwistanding the absence of gravity pressSetting aside the absence of gravity pressures, the pattern of fin movements displayed by this shark reminds me of crawling in salamander (ex.,
    The basal locomotor pattern in Vertebrates is trunk undulation by lateral bending. Numerous studies in lampreys, fish and amphibians have shown that this is done by activating sequentially (with a small lag) neuronal networks located in every spinal segment. Lateral gait is probably the most efficient way to superpose paired appendage (fin or limb) movements on an undulatory pattern, supporting the convergence hypothesis. However, as all these classes show species that swim, walk or crawl using a kind of lateral gait, I think that phylogeny – the fundamental organization of the spinal cord – may be the best explanation to your question.

    If I remember well, it has been reported that coelacanths coordinate their paired fins in a lateral gait manner when swimming (at low speeds at least). Other fish also crawl on the sea floor, some seem to use both pectoral and pelvic fins (frogfish) but more often they use only their pelvic fin. My favorite example is sea robins (or gunards) that crawl in a more arthropod-like fashion on six “legs” ( These “legs” are in fact specialized rays of the pectoral fins covered with chemoreceptors (“taste buds”) that are used to detect food. In these fish, the spinal segments innervating the rays are much differentiated, in good correlation with their increased sensorimotor capacities. Note that, in this case, the motor pattern is not lateral gait as the two rays from a pair are activated synchronously.

    The locomotor networks in mammals (and probably birds) are far more differentiated – it is not definitely proven but it is generally accepted that the networks are restricted to a few spinal segments (caudal cervical for forelimb control, and rostral lumbar for hindlimbs). Moreover, trunk undulation in mammals is dorsoventral and not lateral. So the comparison with a shark is more difficult.

    Desnes Diev

  11. moarscienceplz
    Posted August 27, 2013 at 9:23 am | Permalink

    I’m surprised nobody is questioning WHY the bamboo shark is walking rather than swimming. I’d definitely choose swimming as opposed to scraping my belly on that sharp coral. Maybe walking reduces the shark’s exposure to predators?

    • Michael Fisher
      Posted August 27, 2013 at 10:32 am | Permalink

      They feed on bottom-dwelling invertebrates & smaller fish

    • teacupoftheapocalypse
      Posted August 27, 2013 at 12:05 pm | Permalink

      Many species of shark have sets of electroreceptors in their heads, and particularly in their snouts, which they use to detect prey that live either in murky conditions, just underneath the gravel and sand on the sea floor, or in small nooks and crannies. Bamboo sharks also have barbels, which contain taste buds, and are also used in prey detection.

      Unsurprisingly, Bamboo sharks main prey are bottom-dwelling animals, so it is possible that they have evolved this means of locomotion as the most efficient means staying as close as possible to the sea floor and catching prey.

      • microraptor
        Posted August 27, 2013 at 2:46 pm | Permalink

        That makes me wonder if any of the hammerhead sharks every display such behavior.

  12. Gregory Kusnick
    Posted August 27, 2013 at 11:25 am | Permalink

    As others have pointed out, many tetrapods have several distinct gaits available to them (walk, trot, gallop) depending on the needs of the moment. This suggests to me that biomechanics dominates in determining gait, and to the extent that phylogeny plays a role, it’s due to shared biomechanical features of the vertebrate body plan rather than deep behavioral programs.

    One way to test this might be to induce deformities of the body plan and see how that affects gait. If the ancestral gait persists despite deformities, that would tend to indicated that it’s separately coded in the genes. If the gait adapts to the deformity, that would tend to indicate that it’s learned rather than genetically programmed.

  13. John
    Posted August 27, 2013 at 12:49 pm | Permalink

    It would be extremely difficult to test the hypothesis that neural control for walking evolved before terrestriality. Bamboo sharks are within a different class of vertebrates (chondrichthyes) than lungfish and other tetrapods. Thus, sharks, lungfish, and tetrapod species have all been independently evolving for over 400 million years. With no clear indication of walking being present in extinct shark species it is hard to determine when walking evolved in sharks and how many times. Additionally, the lungfish in your previous post have two muscles spanning their joint between the pelvic girdle and the femur, and other sharks in the Hemiscyllium genus have three muscles spanning this joint. I end with the same question you’ve raised: how could we test between these two hypotheses.
    As far as the importance of this gait in sharks: it is has been hypothesized that walking in sharks is used to sneak up on prey because it causes less displacement of the water around them. The smaller displacement of water may enable them not to be sensed by their prey as easy. Other critiques of gait would include stability. However, the buoyancy of a benthic walking shark would remove constraints to have a stable gait. For instance, one of the gaits the epaulette shark uses would be very unstable on land. However, it works well for them underwater.

  14. Harold
    Posted August 30, 2013 at 11:46 am | Permalink

    Its an abomination before the Lord!

    Just kidding… this is really cool and definitely shows some pretty ancient wiring present in animals that have been around for millions of years for a walking type gate.

  15. scott
    Posted September 25, 2013 at 2:21 pm | Permalink

    It must be a jesus shark

2 Trackbacks/Pingbacks

  1. […] Bamboo shark walks the walk ( […]

  2. […] for 2:15 p.m. ET Aug. 30: On the “Why Evolution Is True” blog, Matthew Cobb says the shark’s walking style looks a lot like the gait of a typical […]

%d bloggers like this: