Have you ever wondered how smart your fish really is? Can your fish learn from another "world" outside the aquarium? Can fish distinguish colors? Can they recognize shapes and symbols? Can they count? If they can, how would they let us know? Well, here is a study you can do at home that will answer these questions.
What is intelligence? Psychologists might define intelligence as the ability to make new "connections." In other words, to learn a new association between previously unrelated things. Can fish do that?
Fish intelligence is not discussed much in books. Most discussion of non-human intelligence deals with "higher" animals such as apes, dolphins, and whales. Treatment of fish behavior, though, is typically considered a study of instinct, not of intelligence. People have not tried to demonstrate much learned intelligence with fish. In fact, all of the fish specialists who were contacted never heard of a fish being trained to make decisions.
In brief, my experiment consists of showing the fish two different 8 cm discs - for example, a red disc and a brown disc (see figures). The fish is rewarded with food only if he chooses the red one. If he can be successfully trained to come to the red disc, then the fish can demonstrate to us that he can distinguish colors. Similarly, I tested to see if he could distinguish between abstract symbols by seeing if he could be trained to come to a disc with an "O" on it versus a "*" on it. Finally, I attempted to see if he could count by trying to train him to come to a disc with one circle on it versus three circles.
The Red Parrot Fish
It is known that some cichlids can recognize the world outside their tank, and may follow fingers across the glass, etc. The red parrot fish, in particular, has been reported to be trained to eat out of the owner’s hands. It has a reputation for "intelligence," "personality," and the ability to adapt. Therefore, this fish seemed a good candidate to study.
The red parrot fish is a hybrid between two species: the golden severum and the red devils. These fish belong to the family Cichlidae, which contains over 1200 species, and are a representative of bony fish with a well-developed dorsal fin. Complex instinctual behaviors by these fish have been well studied. Using its protractile mouth, it has mastered some very complicated mouth breeding techniques.
The red parrot’s eye, which is one of the qualities of fish I tested, is similar to the eye of humans. We share the same basic visual structures: a cornea, an iris, a lens and a retina. The retina contains rods (which detect black and white), and/or cones (which detect color). The presence of cones in some fish’s eyes indicates that fish should be able to see color.
I tested the fish in three different ways: discerning color, number and shapes. Here is the technique:
- First, I trained the fish to follow my finger across the outside of the glass, and dropped food into the tank near my finger. This trained the fish to associate that "finger = food."
- Then, I substituted an 8 cm red cap for my finger, and did the same thing (I used the lid of the TetraMin container). After several weeks, he was able to discern that the "red cap = food." Already, this demonstrated that the fish could make new connections of facts. He can learn!
- Next, I did a test to see whether he could discern red from brown. First, I placed two small food granules on top of the red disk. With my index finger, I covered the food on the red cap. I put the index finger of my other hand on the same place on the brown disc - but no food was hidden on that side. Standing about half a meter away from the tank with the two discs close together, I paused for five seconds. Then I moved forward slowly, spreading the caps apart (I had to look up to avoid watching the fish, so that his movements would not influence me). Whichever circle the fish went to first (within a few centimeters) was his choice. When he was right, I gave him the food; when he was wrong, I did not. When he was wrong, I took the wrong cap away and showed him the right one. I would randomly switch which hand had the red cap. Then every two minutes I would repeat the experiment, five times total each night, recording the results each time.
- After several weeks, I gave him a final test: five trials in the morning, and five trials in the evening for two days. Could he learn that "red = food" but "brown = no food?" If he could learn that, it would prove that he could distinguish colors as well.
- Next, I tested if he could distinguish abstract symbols. I put a 3 cm "O" on one 8 cm diameter white paper disc, and a 3 cm "*" on another white paper disc. He only got the food reward if he first approached the "O" disc.
- Finally, can he count? On one white disc, I placed a single 1 cm solid dot. On the other, I placed three of the dots. Could he be trained to come only to the single dot?
Overall, the fish succeeded in doing very well on the color test and the shapes test, but not the number test (see graphs). On the color test, the fish got 18 out of 20 tests correct, for an accuracy rate of 90%! Statistically, the odds of him "guessing" the correct color that well were only 1:2500 (see table). In the abstract shapes test, he got 38/50 correct, i.e., he was right 76% of the time. The odds of doing this well just by chance were only 1:5000. For the number test, he only got 9/20 trials correct, which was not any different from pure luck.
|Red disc vs. Brown Disc||20||18||90%||1:2500|
|1 Circle vs. 3 Circles||20||9||45%||1:1.5|
|Circle vs. Star||50||38||76%||1:5000|
Therefore, I proved that a fish can learn to make new associations (red gets food, not brown; and "O" gets food, not "*.") This showed that he is capable of learned intelligence. He can make his own decisions, and not just act instinctively. I also proved fish can see color and that he can see shapes through the tank. It is not clear why, but the fish was not able to tell the difference between one dot and three. This would have seemed important, because fish need to know how many fish are attacking them.
This simple procedure - testing if a fish can be trained to distinguish one target versus another - allows us to communicate with our fish. It allows the fish to demonstrate to us what he knows. If you have noticed that your fish reacts when you approach the tank, maybe you can train yours. See how smart your fish really is. Have fun with this, and maybe your fish will have fun with you!
Marc Kutscher is a sophomore at Soundview High School in Mount Kisco (NY) - USA.
His red parrot fish, "Biggy," has been a family pet for several years.
Brilliant, well done for putting in the time. My friends and family are always amazed the way my fish know me more than anyone else that comes to the tank, because I'm the one that feeds them. Really good to read an article like this to show it goes even deeper than this. Makes me feel a little guilty though, about all the poor fish that died when I was learning in the early days :-(
It has truely been ages since I did this study...I am the one who wrote this. In truth, it was my brother's fish, and I, myself, am a cat person...cats are difficult to train. Yes, some fish are intelligent. If we analyze it, however, from a neurological standpoint, fish have only what we view as the 'reptillian brain'. Complex memory and learning does not exactly exist...instead, we can attribute these patterns to neurological development of the fish, BECAUSE we trained it. If we observe learning as it occurs in mammals, certain parts (e.g., the hippocampus, prefrontal cortex) are needed to learn. The fish itself is probably using a response type memory here, I should really repeat this experiment.
I have many years of experience with pond and aquarium fish. Although I am a behavioural scientist I have not performed any tests. However, I would like to say that intelligence is important in what it accomplishes or shows itself. For the human the ability to interact with another species and be recognized by it is often an important part of intelligence. I have found most fish both recognize me and depending on their personality, learn to trust, take food and swim with me in the pond.
Got some experience to share for this page? No registration necessary to contribute! Your privacy is respected: your e-mail is published only if you wish so. All submissions are reviewed before addition. Write based on your personal experiences, with no abbreviations, no chat lingo, and using proper punctuation and capitalization. Ready? Then send your comments!