Experimental set-up to test distance estimation in the goldfish. Black and white striped panels (width 2 cm) covered the tank walls and floor providing constant optic flow information to the individual (only floor and left-side wall are represented for clarity). The fish was placed in a movable start area, with a sliding door, for acclimation. Once the door opened, the fish was trained to swim until the experimenter waved above the tank at the 70 cm target distance and then to return to the start area to receive a food reward. Two white partitions (gray on picture for clarity) prevented the fish to obtain external visual cues from both ends of the tunnel, one was placed at the end of the experimental tank and the other one was placed 20 cm behind the starting door. An overhead camera (gray cylinder) connected to the laboratory computer displayed fish movement in real time and recorded distance estimates during the testing phase. The linear tunnel was constructed inside a flow-through tank, with water flow in behind the start area (left blue pipe) and passive water flows out at the opposite end of the tank (right blue pipe). Water flowed outside of training and testing sessions. T1, T2 and T3 indicate the three start positions used during training. P1, P2 and P3 indicate the three start positions used in test. Credit: Proceedings of the Royal Society B: Biological Sciences (2022). DOI: 10.1098/rspb.2022.1220
× close
Experimental set-up to test distance estimation in the goldfish. Black and white striped panels (width 2 cm) covered the tank walls and floor providing constant optic flow information to the individual (only floor and left-side wall are represented for clarity). The fish was placed in a movable start area, with a sliding door, for acclimation. Once the door opened, the fish was trained to swim until the experimenter waved above the tank at the 70 cm target distance and then to return to the start area to receive a food reward. Two white partitions (gray on picture for clarity) prevented the fish to obtain external visual cues from both ends of the tunnel, one was placed at the end of the experimental tank and the other one was placed 20 cm behind the starting door. An overhead camera (gray cylinder) connected to the laboratory computer displayed fish movement in real time and recorded distance estimates during the testing phase. The linear tunnel was constructed inside a flow-through tank, with water flow in behind the start area (left blue pipe) and passive water flows out at the opposite end of the tank (right blue pipe). Water flowed outside of training and testing sessions. T1, T2 and T3 indicate the three start positions used during training. P1, P2 and P3 indicate the three start positions used in test. Credit: Proceedings of the Royal Society B: Biological Sciences (2022). DOI: 10.1098/rspb.2022.1220
A team of researchers at the University of Oxford has found via experimentation that goldfish use markings on the floor below them to measure how far they have traveled. The findings are published in Proceedings of the Royal Society B.
Prior research has shown that humans are able to navigate by using certain unique features around them as they move to serve as reminders of where they have been. This allows for returning to a starting point. In this new effort, the group in England has found that goldfish do the same thing to a certain degree.
To learn more about how goldfish navigate, the researchers painted stripes on the bottom of a fish tank and then trained several fish to travel a certain distance on demand. In the beginning of the experiment, the fish were trained to stop swimming when a researcher waved a hand over the top of the tank, which resulted in a food reward. Next, they tested how far the fish would swim without that signal and found the fish swam nearly the same distance before stopping.
The researchers then took things further by training the fish to travel a certain distance before stopping without using a waving hand鈥攁gain, to receive a food reward. The fish were trained instead by rewarding them when they had swum a certain distance. To make sure that the fish were not simply stopping at a certain point in the tank, the researchers varied the starting position.
After working with all of the test fish, the researchers found that all but one could be taught to swim a distance of 70 centimeters very accurately. The team assumed their accuracy was due to tracking the stripes that had been painted on the bottom of the tank. To confirm their suspicions, the researchers repeated the experiment with the same group of fish in a tank that had stripes painted at shorter intervals鈥攖he fish swam a shorter distance in response. The researchers also note that time was not a factor in their navigation skills, traversal times were random.
They acknowledge that they were not able to explain why one of the fish, named Bubble, was unable to be trained like the other fish鈥攖hey suggest it was analogous to people who tend to get lost.
More information:
Adelaide Sibeaux et al, Distance estimation in the goldfish ( Carassius auratus ), Proceedings of the Royal Society B: Biological Sciences (2022).