A predator that's cool will pressure fish to school

Sep 12 2012 Published by under Evolution, Uncategorized

Sings: *but that's why birds do it, bees do it, little fishies in the seeeeeas do it, let's do it...*

Let's come together. Into groups, schools, herds, flocks. Prey animals all over the world do it.

We know why prey animals tend to clump together. More animals means more eyes to spot predators. More animals means if the predator gets someone, it's less likely to be you. More animals means that the predator could get confused by so many bodies and miss entirely.

Those are the general reasons why prey animals hang together. But when they do hang together, such as when fish school, they also tend to MOVE together. It's not just a group of fish milling around, the fish group into clumps and move in coordinated patterns. Since so many species do it, we assume this has an evolutionary advantage, but we had no direct proof of this.

Until the authors of this paper decided to frustrate the heck out of some bluegill sunfish.


C. C. Ioannou, V. Guttal, I. D. Couzin. "Predatory Fish Select for Coordinated Collective Motion in Virtual Prey" Science, 2012.

The question at hand here is how the individual responses of prey animals to their neighboring prey animals influences predation risk. You'd think that if you were going to investigate this, you would need a whole bunch of prey animals to look at, but actually, you don't. What you need is a predator that is easily fooled by virtual reality.

So the authors of this study looked at the bluegill sunfish, a freshwater fish (I didn't know what it was until I saw it and went 'hey, that's a bream!') which is pretty common throughout North America. Bluegills will eat a lot of stuff, but they like insects and small fish. When they hunt, they "hover" close to the prey, and then lunge, suctioning water rapidly into their mouths and pulling the prey along with it.

We've got the predator, but what about the prey? Well, for this study, you don't want to actually use REAL prey. The prey would react to the predator, and the reaction is dependent on a lot of factors (like group size, approach of the predator, etc). The authors wanted to look at how the behavior of the PREY affects how likely they are to get eaten, and this means you need prey you can control. You need virtual prey.

So the scientists projected an image onto a screen on the side of a large tank. The image was of about 16 dots, representing prey. The 'prey' was carefully programmed to behave in a combination of three specific ways: (1) it could be attracted to its neighbors and stick close to them, (2) it could orient in a specific direction of travel, or (3) it could ignore its neighbors entirely. Combinations of these three scenarios gave them a large range of behaviors, from full schooling (sticking together and moving together) to dots moving randomly.

Then they released the bluegill.

(You can see here a layout of the chamber, with additional figures of the bluegill and the 'prey')

They looked at how likely the bluegill was to make an attack upon its 'prey'. What they showed was that the bluegill exerted selection pressure on the prey, showing a strong preference for attacking the randomly moving dots. The lowest risk for attack was for 'prey' that not only stuck together, they moved in the same direction together.

The bluegill tended to target prey in small groups OR large groups (as you can see up in the top box), targeting fish particularly on the edge. But it also tended to target prey that were moving with either very low or very high "tortuousity" (the bottom box), that is, in less or more complicated paths, rather than wandering about. The higher tortuousity, the more fish you have moving in complicated patterns, the less the prey can stay together, resulting in a random swarm that is much easier for a bluegill to predate. Conversely, if you have one fish moving with extremely low tortuousity (in a straight line), it's more likely to be outside the group and more subject to predation.

In other words, this paper shows that predation risk is lowest when prey of a feather groups together, not only collecting in groups, but also moving together. There's selective pressure for this by predators, which are more likely to target other modes of movement, making the highly grouped and oriented school the best formation. So the next time you see that large school of fish, or that flock of birds, or herd of horses, racing together in long big arcs, you know what's up, and that they are probably giving a predator a very hard time.

Ioannou CC, Guttal V, & Couzin ID (2012). Predatory fish select for coordinated collective motion in virtual prey. Science (New York, N.Y.), 337 (6099), 1212-5 PMID: 22903520

2 responses so far

  • Thank you for this clear (and visually appealing) post on a great topic! I'm always curious about what animal behaviors say about human behavior. If schooling, flocking and herding occur across fish, birds, and mammals, might we also carry the urge to perform coordinated group behaviors...and reap their protective benefits? My personal theory is that our inner flocking mechanism is what makes public dancing excruciating for so many people...one step out of place...all eyes are on you...and BAM! you're lunch.

  • David says:

    Very cool post, thanks !!!

    The inverse question is also interesting: what is the selection pressure on the predator to go after the individuals and not the schooling group?

    Another question about fish evolution: do groupers experience group selection? (sorry).

Leave a Reply