In most workplaces, spitting at computer terminals is regarded as a bit of a faux pas — but not so for certain banded archerfish in a laboratory in Italy.
Experts from the University of Trento trained the fish — which use jets of water to capture prey — to spit at screens to show they can distinguish between numbers.
Numerous studies in the past have shown that fish, just like other birds and mammals can count and possess an inherent sense of numbers.
However, these tests have demonstrated, for instance, that fish will join one of two shoals. Critics argue this is merely an indication of their size.
A fish might decide to pick one of two prey-like dots, and choose it simply because it has more dots.
These confounding factors were controlled by the team using dots of different sizes and relative locations, which proved that fish could count.
Scroll down to see the video
In most workplaces, spitting at computer terminals is regarded as a bit of a faux pas — but not so for certain banded archerfish (pictured) in a laboratory in Italy, who have done so as part of an experiment to show that fish can count and distinguish between different numbers
Experts from the University of Trento trained archerfish — which use jets of water to capture prey, as depicted — to spit at screens to show they can distinguish between numbers
“There’s a lot of debate about the existence and meaning of many senses.” [in fish], based on the fact that it is empirically impossible to separate numerical information from all other continuous properties at once,’ paper author Davide Potrich told the New Scientist.
‘Several experiments have tried to address this issue, but usually not in a complete way,’ the University of Trento animal cognition expert added.
“What makes our study unique is the fact that we managed to control for all non-numerical variables as best possible.”
In their experiments, Dr Potrich and colleagues placed banded archerfish in tanks beneath monitor screens that displayed two groups of dots. The fish were recorded by a camera that was placed underneath the tank. It also showed which dots they chose.
Repeatedly showing the fish two sets of dots — one with three, the other with six — the team succeeded in training archerfish to consistently pick one of the numbers.
Software was created by the team that changed randomly the dot size, arrangement, and details of each test to ensure other factors, such as area, did not influence fishes’ decision making.
For example, six dots might take up more space than three, while the next time, they may cover a greater area.
The team found that the fish were still able to distinguish between numbers when presented with new alternatives, such as six and eight, rather than six and three.
However, the fish tended — around three-quarters of the time — to default to picking the largest or smallest number (depending on their training), rather than their usual number, when presented with these different choices.
So, for example, a fish taught to pick six over three would pick nine over six and eight over four,— whereas a fish trained to pick three over six would choose the lower of the two numbers when given a different combination.
In their experiments, Dr Potrich and colleagues placed archerfish in tanks beneath monitor screens that displayed two groups of dots (left). An under-the tank camera captured a view (right), of the fish, and the number of dots that they chose.
Brian Butterworth, a neuroscientist from University College London who was not involved in the present study, told New Scientist that, in nature, numerical factors tend to go with non-numerical ones, making studies like this difficult.
He said, “It seems that the team has done a very great job trying to distinguish these non-numerical clues from the numeral ones,” he continued.
The ‘Archerfish’ can make at least relative numerosity judgements.
You can view a preprint from the article by the researchers on bioRxiv. It hasn’t been peer-reviewed yet.