Who researched:
Graziano Fiorito and Pietro Scotto
When researched:
Published on April 24, 1992 in Science volume 256, number 5056, pages 545-547
How the research was conducted:
The research conducted went through three phases: training the demonstrators (Octopus that were conditioned to “attack” one ball over another), observing the actions of the untrained Octopus when given the same task, and testing the observers.
Phase one: the demonstrators were rewarded when they attacked the right ball (given fish) and punished when they attacked the wrong ball (shocked). They were considered fully trained when they attacked the right ball five times consecutively. Phase two: Once fully trained, an observer Octopus was placed in a glass tank adjacent to the demonstrator Octopus tank to witness four trials in which the correct ball was attacked. Researchers witnessed eye/head movement of the observers, proving the observer Octopus was paying attention to the actions of the demonstrator.
Phase three: The observer Octopus was given the same task of choosing between two balls (red or white) in five trials. No reward was given for choosing the correct ball.
What they found:
Octopus vulgaris demonstrated an ability to learn through observation of others. The failure rate of the observer Octopus (when given the same task) was significantly less than that of the demonstrator Octopus when it began training. Furthermore, the demonstrator Octopus learned to choose the correct ball much faster than their demonstrator counterparts when they were put through the classical training methods that were used on the demonstrator Octopus. When tested again five days later, the observer Octopus seemed to retain its knowledge of which ball to choose proving the Octopus vulgaris’ ability to retain knowledge.