Can fish count?听An experiment involving archerfish spitting at computer monitors has provided the best evidence yet that they really can distinguish between different numbers.
Various studies over the past decade or so have suggested that fish have an innate sense of numbers like many birds and mammals do and are surprisingly good at mathematics. But these typically involved tests such as individual fish choosing to join the largest of two shoals.
Critics say these experiments , rather than of number. For instance, if fish have to choose between two sets of dots, they might be picking the higher number based on more dots covering a larger area, rather than the actual number of them.
Advertisement
鈥淭here is a debate regarding the existence of a number sense, based on the fact that it is empirically impossible to separate numerical information from all other continuous properties at once,鈥� says at the University of Trento in Italy. 鈥淪everal experiments have tried to address this issue, but usually not in a complete way.鈥�
Now, Potrich and his colleagues have developed software that controls for other factors that might influence choices by randomly changing the size of the dots, their arrangement and other details. For instance, when fish had to pick a higher number of dots, these would sometimes take up a smaller area than the lower number of dots.
For the study, the team decided to use banded archerfish (Toxotes jaculatrix) that spit jets of water at prey both above and below the surface. They can be trained to hit targets on a computer monitor above the water.
The researchers displayed one set of three dots and another set of six dots simultaneously 鈥� varying the area and other details each time 鈥� and taught different groups of archerfish to pick one of these numbers. They then showed them various other pairs of dot sets, such as arrangements of six and nine or five and eight. The results show that the fish can be taught to distinguish sets based on number alone.
The fish also generalised this to 鈥減ick the largest or smallest number鈥� when offered new choices. For instance, fish taught to pick six dots over three would choose nine dots over six and eight dots over five around 75 per cent of the time. Fish taught to pick three dots over six, however, would choose six over nine and five over eight around 75 per cent of the time.
鈥淲hat is unique in our study is that we controlled for non-numerical variables in the best way possible,鈥� says Potrich.
In nature numerical and non-numerical factors tend to go together, says at University College London. The size of prey matters, for instance, so it is hard to separate it from other factors, but he thinks Potrich and his colleagues have done it.
鈥淚t seems to me that the team have done a very good job in trying to separate out these non-numerical cues from the numerical ones,鈥� says Butterworth. 鈥淎rcherfish can at least make relative numerosity judgements.鈥�
Reference:
Sign up to Wild Wild Life, a free monthly newsletter celebrating the diversity and science of animals, plants and Earth鈥檚 other weird and wonderful inhabitants
