Zyra's website //// Fish //// Goldfish //// Concepts //// Site Index

Fish Memory

How long can fishes remember?


For a long time now, knowledge of the memory capabilities of fish has been clouded by a false belief that "Goldfish only have a memory of 7 seconds". Some folklore now has it as a memory of 3 seconds. This has been believed by many people, and yet there isn't the evidence to back it up. What's really odd about it is people not questioning the "wisdom" of such a thing.

If creationism had any credibility, it is conceivable that some perverse deity could indeed create fish with preposterously short memory-spans. However, Darwin's evolution does not support such things. Survival tends to favour those with the better survival features, and that includes having a reasonably good memory. Fish with better memories would survive better and get to produce more fish eggs that would hatch into new fish.

It's all very well to reason something out like this, but what's really significant in discovering facts which become knowledge is... experiment, first-hand experience. Now that I've moved to Panama where it's tropically hot, I have had a chance to encounter fish more closely. Tropical fish, of the type you might seen in tropical fishtanks with little water heater elements and other life-support systems, live in the wild. So, when I go along to the small river at the bottom of the garden/jungle, the clear water has tropical fish in it. As the water is warm, I have been swimming with tropical fish. When I first started doing this, the fish kept their distance, and fish are quite wary of getting too close to any creature much larger than they are. But then I started feeding them.

Fish are not known for being fussy eaters, and they were quite keen to eat the scraps of dinner I had brought them. Curiously, they were especially fond of cheese. Rice, chocolate, and chicken were also avidly consumed.

Where this is relevant becomes clear when a few days later I went to the river to go swimming with tropical fish again. This time, the fish were very interested as soon as I got in the river. I thought, if they are a school of fish then they are fast learners! On subsequent occasions, the fishes were quickly there to accept food and it's clear they associated the presence of mad person in the river as a likely sign they were going to get fed. This must mean that the fish have memory at least good enough to remember what happened last time. Fish memories of many days, obviously.

I'm not the only person to have experienced fish behaviour which shows memory. There's a notable set of experiments someone has done with goldfish to find out about their memories. See Goldfish-memory-myth-busted

This also raises the issue of why goldfish have in particular been previously believed to have a very short memory. It has been suggested that the rumour was made up so people didn't feel so bad about keeping pet fish in small bowls. If this were true, it would be a scandal, and would suggest that maybe a few of the other beliefs going about in human society are similarly corruptly based.

A while ago, when I heard about "goldfish only have a memory of 7 seconds", I considered it was probably not true, and I tried to do an experiment that would help to determine if it were likely. I had no fish, but at the time I was living in a temperate climate and there were mosquito larvae in a barrel in the garden. Mosquito larvae are aquatic creatures which breathe air, and they spend some time underwater and some time on the surface. They dive when they perceive a threat. I found that they didn't dive when a hand was waved over the water, but they would dive when the water surface was disturbed. So, I waved my hand over the water and then disturbed the surface, and soon the mosquito larvae were conditioned into the association between the events. As a result, they would now dive when the hand waved over the water. They had been conditioned into associating the sight of the hand going over the surface with a likely disturbance of the water. They had learnt, but would they remember? I went away and came back half an hour later. The mosquito larvae responded to a hand waving over the water by diving immediately. Therefore mosquito larvae have memory of at least half an hour. This is an actual result! I concluded that as mosquito larvae are young insects and therefore much less advanced creatures than fish, then it was likely that fish had memory at least as good as that.

Whilst this last step is speculative, it's the best that I could do as I had no fish at the time.

My later experiments of swimming with tropical fish have results which show fish have memories of many days at least. Admittedly these wild freshwater tropical fish are not "goldfish" as such, but it is likely the memory-spans have something in common.

I am unsurprised that fish have reasonably good memory. It would have been absurd if fish forgot everything after a few seconds.


Another thing about this: It shows how it's a good idea to do experiments wherever possible. It's no good just believing water goes down plugholes clockwise just because someone has told you so. Experimentalism is triumphal over assumption. Believing "experts" are always right is folly. A few other experiments I have done here and written about are... cats can see in the dark? , we dream in colour , and measuring diabetic blood sugar by urine tests. You can see this idea carried to extremes in the proposed experiment "glugtime" where a watch which is said to be "waterproof down to 100 metres" is put to the test!