Proprioception and Visual Discrimination of Orientation in Octopus

1. Twenty-two small octopuses were trained to make visual discriminations before and after removal of the statocysts. 2. Removal of one statocyst (left or right) did not affect performance in a visual discrimination of rectangles shown horizontally and vertically (8 experiments), nor did it determine which eye was used to guide attacks upon things seen (a record of this was kept in four of the experiments). 3. Removal of both statocysts did not affect performance in a discrimination between black and white circles (7 animals) but did destroy the capacity to discriminate successfully between rectangles shown horizontally and vertically (11 animals). 4. After removal of both statocysts the orientation of the retina (as indicated by the slit-like pupils) is no longer constant over the normal wide range of bodily positions, but depends upon how the animal is sitting on the side or bottom of its tank. 5. An analysis of the postoperational responses of animals trained to make the orientation discrimination showed that nearly all of them were biased in favour of one or other of the test figures, not always that used as ‘positive’ figure during preoperational training. The postoperational scores of these trained animals depended mainly upon the positions in which they habitually sat in their tanks; ndividuals in positions such that the retina of the eye used to observe the test figures lay at right angles to the position held during pretraining tended to behave perversely. 6. These results indicate that orientation discrimination is dependent upon correct orientation of the retina, rather than upon central integration of proprioceptive and visual information. This is discussed in relation to a possible mechanism of visual discrimination proposed by Sutherland (1957), and in relation to what is already known about the role of proprioceptive information in learning.