Why do caterpillars swarm? We built a game to find out

In my last post, I broke down the science of why some caterpillars work together and form these strange, writhing formations known as rolling swarms. In a sentence: the caterpillars use their own bodies as a constantly re-assembling and dis-assembling conveyor belt, and by doing this they manage to give themselves a speed boost. If you haven’t read that post, go back and check it out.

rolling swarm caterpillars

Here’s another video of this creepy behavior:

Inspired by this notion of co-operating caterpillars, my friend Deepak decided to dig a little deeper, and try to understand why they work together in the first place.

What if, he asked, each caterpillar was just behaving selfishly, and only trying to overtake the caterpillar ahead of it? With that simple rule, would such harmonious, collective behavior emerge?

Well, we built a thing where you can play around and find out the answer for yourself. Here’s the link.

Check it out, and let me know what you think in the comments below. I welcome your thoughts, and constructive feedback. And go easy on us, this simulation was hacked together in a matter of hours, but we hope that you’ll have fun with it.


Filed under Science

  • Physicalist

    It seems like I can get them to go nearly as fast when they’re not really swarming as when they are swarming — maybe because when they’re spread out they’re always hurrying to catch up, but when they’re swarming they slow down when they reach the top or are in the pack?

    Maybe if they were only allowed limited bursts of speed to pass their neighbors the difference would be greater?

    • Hmm, once you pump up the number of caterpillars, there should be a pretty noticable speed difference.

      But you’re right – when they all chase each other, they can get pretty fast. A limited burst of speed would be a good solution. Another way we could add this is put in a speed limit that only allows you to go one notch faster than the speed for your level. so 2x for the first floor, 3x for the second floor, 4x for the third floor and so on, with an upper limit. That would keep them from busting out the unrealistic 5x speed in the ground level.

      Thanks for the feedback. That was really helpful.

  • Siddharth Nishar

    Hello. Your blog is in general extremely intriguing and fun to read. Bookmarked! 🙂

    As an engineering student, I have a fundamental doubt about the calculations made regarding the increased speed of these caterpillars.

    The limiting factor in this case is the capacity of the caterpillar’s body to do work. Work = Force X Displacement. The force is the frictional force (on a smooth surface) that the worm must overcome.

    Assuming that the power output of the caterpillar’s muscles is constant, three layers of worms means that on average the bottom worm has to go against thrice the frictional force. This in turn means he must go at 1/3rd the speed.

    However, at this point, I want to point out that due to the smooth skin of the worms, the second and third layer have a much faster relative velocity in comparison.

    If mu represents the coefficient of friction, and state 1 and state 2 represent with and without a train respectively,

    The first layer will move at a speed of [v/3]
    The second layer will move at a speed of [v*(mu1/mu2)*(1/2)] + [v/3]
    The third layer will move at a speed of [v*(mu1/mu2)] + [v*(mu1/mu2)*(1/2)] + [v/3]

    The average comes to v(2k + 1)/3 where k = (mu1/mu2)

    As is evident, when k=1, the effective speed remains constant. This agrees with the conservation of energy principle, I feel. If the extra velocity was not due to power saved in overcoming friction, you could effectively create a perpetual motion machine run on worms, yes?

    Please help me understand if my assumptions are severely limited or otherwise.