20 May 2013

Baby geniuses: young guppies show number skills

I have vague memories of the first time I counted to a hundred. It felt like one of those landmarks like tying your shoes for yourself the first time, or riding the bicycle more than a few feet without the training wheels or dad holding you up.

Of course, I don't come anywhere near Adam Spencer:

Once when I was about 7, I counted to 10,000 just to check the numbers didn't run out before then #NerdConfessions

Counting large numbers is not something that comes easily for us humans. A new paper claims this little guy, a baby guppy, may be a superior number cruncher as soon as it pops out of mama’s belly:

A couple of years ago, I reported on a paper that looked at the development of “counting” ability in guppies. In that paper, they claimed that it took about 40 days for guppies to develop the sort of ability to distinguish numbers that they had as adults. Now, the same team is back, testing very young guppies again, but this time using new methods.

The team asked these tiny guppies if they recognized numbers of things by showing animals dots while they have them food. Here are the three stimuli the team used.

Both A and B differ in the number of spots, but A also differs in the average sizes of those spots (which the authors call a “continuous variable). C differs in size, but not in the number. This is try to control for the fact that when you change number of things, you also change many other factors, like amount of area reflecting light, etc.

The authors then measured the amount of time the guppies spent near each set of dots as an indication of “preference”, on the assumption that the guppies are more likely to spend time near the dots where they got food if they learned certain dots meant food. If animals don’t learn where the food is, they may well not be able to tell the stimuli apart.

The authors place these pairs of dots at the end of the tank while fish are feeding when they were four and five days old. As a control, they either feed the fish food or just in a little water without food. On day six, they placed the babies in the tank to see which set of dots they gravitate to. On day seven, they repeat this, but flip the positions of the dots.

The fish were significantly more likely to be around the set of dots that promised food when they differed by number (A and B, above), but not when the dots varied in size. That said, the guppies were not great at this. The guppies got it right only 60% of the time, which is only a slight improvement on a coin toss.

However, the authors themselves admit that this paper is hard to compare with their previous one because the stimuli are so different. The previous paper used other live fish as the stimulus, not just static dots. They also note that this test is slightly different from other training tests, which generally ask the animal to do something even more specific than “hang out at one end of an aquarium.”

It is an interesting suggestion, though, that animals so small and so young can cope with differences in number. But I still think I’ll beat them at counting to a hundred.


Piffer L, Miletto Petrazzini ME, Agrillo C. 2013. Large number discrimination in newborn fish. PLOS ONE 8(4): e62466 DOI:

Related posts

One fish, two fish... can fish count?

Picture by Shaojung on Flickr; used under a Creative Commons license.

No comments: