Angelfish: it’s all about numbers


From parrots that can imitate human speech to typewriting chimpanzees, the complex ins and outs of animal intelligence are only just beginning to be revealed, and as one renowned scientist at UTM attests, much more remains to be learned.

“Our lab works with zebrafish, one of the simplest vertebrate laboratory model organisms with which one can study complex biological phenomena, including brain function and behaviour,” explained professor Robert Gerlai, a psychology instructor on campus who specializes in the study of animal cognition and behavioural genetics. “The ultimate goal of our work is to understand mechanisms of alcoholism, alcohol abuse, and fetal alcohol syndrome. We literally make fish drunk.”

But giving new meaning to the phrase “drank like a fish” was only the beginning of Gerlai’s experiments with the creatures. While trying to better understand how the brain in fish (and by extension all vertebrates, including humans) works, Luis Gomez-Laplaza, a Spanish professor who visited the lab, brought a whole new dynamic to his research: the discovery that angelfish appear to possess the ability to count to three.

“Angelfish is a species that belongs to Cichlidae, or perch-like fishes,” Gerlai continued. “Members of this family are ‘known’ anecdotally to be highly intelligent, an anthropomorphic notion that is based upon observation of how many different and seemingly complex behavioural responses (movement patterns) this fish has. And indeed this is the reason we chose angelfish: they have a complex behavioural repertoire, and also importantly, they move slowly, and thus can be observed easily, and they shoal [form groups] when they are young.”

Shoaling behaviour exhibited in fish species, such as angelfish, are the result of an adaptation similar to that of herding animals. Sheltering in a large population of similar-looking animals acts as a defence against predation by individual fish, as hunter species are spoiled for choice of prey. As well it offers more eyes scanning an area for food sources. The ability to discern a larger school of fish from a smaller therefore contributes to survival for these species.

But Gerlai makes it clear that the behaviour demonstrated by the angelfish was not “counting” in any human conception of the term.

“What we did was ask angelfish which group they prefer, the small or the larger one, in a simple choice task where one shoal was presented on one side of the test tank and another on the other side,” says Gerlai. “What we varied was the ratio of the shoal sizes left versus right side, and also we varied the absolute difference between the shoals.”

“In this series, what we found was that the relative preference for the larger shoal depended on the size of the ratio and not on the absolute numerical difference between the shoals.”

In other words: the larger the ratio the stronger the preference. In essence, the fish were not counting individual fish in the shoals and were not subtracting one shoal number from another. Instead, the individuals chose which group to join according to the ratio of fish in each. Other factors, including location of a shoal colony (e.g. open water or close to a reef bed) and the amount of light being reflected off a school of fish, can also determine a fish’s preference.

Gerlai’s findings show that although angelfish are incapable of counting individual fish in a shoal, when tested in smaller ratios with very few individuals, angelfish seem to possess the ability to select shoals based on counting.

“It turns out that with smaller shoal members in a shoal, angelfish can go below 1.8:1. In fact, we found they can distinguish three from two fish, a ratio of 1.5:1. It is thus likely they use a method to judge shoal sizes that is different from what they do when they look at large shoals. One possibility is counting the individual members of  the shoal.”

But why are fish only limited to low-number ratios when counting? The answer is again applicable to our own species. Most infants of a certain age can distinguish between three items or less; similarly, non-human species have a similar limitation. Gerlai notes that counting above three may not have had any adaptive advantage to individuals in the past, so this ability never had to  develop.

The big question that this relates to is where our more sophisticated mathematical abilities have come from. Could they originate from an ancient fish ancestor? At this point we can’t answer this question, but Gerlai suggests that analyzing genes may hold the key. If researchers can find genes underlying counting abilities in multiple species similar to ours, it would corroborate the common evolutionary origin.