1. After removal of the median inferior frontal lobe, blinded octopuses already trained to discriminate by touch between rough and smooth spheres continued to do so, but at a lower level of accuracy.
2. Animals without pre-training showed a strong tendency to take rough objects after this operation and learned to discriminate well only when trained to take rough and reject smooth.
3. When animals with intact inferior frontal lobes were given food in the presence of a smooth sphere they learned to take the smooth; in subsequent extinction tests they continued to take the smooth but soon ceased to take rough objects.
4. Animals without median inferior frontal lobes also increased their tendency to take a smooth object associated with food. But they did not behave in the same way as controls in extinction tests; they continued to take the rough objects even if they had not been rewarded for doing so.
5. Operated animals thoroughly pre-trained to take smooth objects showed some capacity to discriminate these from rough objects in subsequent successive training with food and shock, though continuing to take the rough far more than control animals.
6. Animals without brain damage could be taught to take smooth rather than rough objects on one side, and continued to do so when trained in the reverse direction on the other. There was, however, some lateral interference; performance on the unreversed side was worse after the introduction of reversed training.
7. Animals with lesions to the median inferior frontal lobe failed to learn on the reversal (rough+/smooth-) side, responses to both objects declining progressively as training continued. At the same time as this discrimination by the non-reversal (smooth+/rough-) side continued to develop. There was thus no evidence of lateral transfer in these animals.
8. It was confirmed that tactile learning is still possible after removal of the vertical and basal lobes, but with some decrease in the normal preference for smooth objects.
9. The median inferior frontal is thus not essential for tactile learning, but greatly facilitates it, making some contribution to the acquisition of both positive and negative responses, perhaps by spreading information through both sides of the touch-learning system. The effect of its removal in touch learning can be compared with the effect of vertical lobe removal on visual learning. It is concluded that one function of these parts is to compensate for the intensity of stimulation so that animals do not pay undue attention to brightly reflective or texturally rough objects.