DAVID INGLE
74
learned by pretraining subjects on an easier version of the problem in
order to reward them for “switching in the appropriate analyzer.” Indeed,
fish pretrained on a circle-triangle discrimination were able to solve an
otherwise impossible circle-square discrimination. A second pretraining
series on circle vs. larger square, emphasizing the irrelevant dimension
of size, did not facilitate learning of the difficult circle-square problem.
However, a study by Sutherlands group (Mackintosh et al., 1966)
failed to show evidence of selective attention: Overtrained goldfish
reversed a shape discrimination more slowly than controls, unlike rats
who reverse more quickly under similar conditions ( Mackintosh, 1965b).
Other tests will be required to decide whether fish have evolved this
interesting facet of behavior. A final answer will be useful to the physiologist who wonders how radically the visual coding processes of retina
and tectum may be altered by centrifugal influences.
VI. TOWARD A UNIFIED OUTLOOK ON VISUAL BEHAVIOR
Aspects of fish behavior which have not yet been analyzed in a sufficiently rigorous manner have been deliberately excluded, although it is
clear that vision has been designed for ecological and social functions
and not for playing games with psychologists. Analysis of visual mechanisms might be pursued by experiments that describe optimal stimuli
for guiding natural movements or eliciting consummatory responses.
Briefly summarized below are some hypotheses designed to link observed behavior toward artificial stimuli with the uses of vision in the
real world. A more complete discussion of this material has been
presented elsewhere ( Ingle, 196813, 1971).
The study of motion detection by goldfish (Ingle, 1967, 1968b) reveals
two or three independent processes that have been hypothetically
identified with aspects of natural visual behavior. The fast-moving single
spot is taken as a prototype of predator motion or the motion of a rival’s
markings used to direct aggressive attack (an eye or a body spot). The
higher probability of catching forward rather than backward moving
prey might explain the specialization for nasalward directions with a
fast-moving spot. Furthermore, it has been noticed that aggressive mouthbreeders, Tilupiu, seldom, if ever, attack a smaller fish moving in the
opposite direction, but they readily chase a rival who moves past in a
head-to-head orientation.
The slow-moving spot, seen best while moving backward, might correspond to a prey object being pursued; the ternporalward velocity would
inform the pursuer how rapidly the gap was being closed and help set
74
learned by pretraining subjects on an easier version of the problem in
order to reward them for “switching in the appropriate analyzer.” Indeed,
fish pretrained on a circle-triangle discrimination were able to solve an
otherwise impossible circle-square discrimination. A second pretraining
series on circle vs. larger square, emphasizing the irrelevant dimension
of size, did not facilitate learning of the difficult circle-square problem.
However, a study by Sutherlands group (Mackintosh et al., 1966)
failed to show evidence of selective attention: Overtrained goldfish
reversed a shape discrimination more slowly than controls, unlike rats
who reverse more quickly under similar conditions ( Mackintosh, 1965b).
Other tests will be required to decide whether fish have evolved this
interesting facet of behavior. A final answer will be useful to the physiologist who wonders how radically the visual coding processes of retina
and tectum may be altered by centrifugal influences.
VI. TOWARD A UNIFIED OUTLOOK ON VISUAL BEHAVIOR
Aspects of fish behavior which have not yet been analyzed in a sufficiently rigorous manner have been deliberately excluded, although it is
clear that vision has been designed for ecological and social functions
and not for playing games with psychologists. Analysis of visual mechanisms might be pursued by experiments that describe optimal stimuli
for guiding natural movements or eliciting consummatory responses.
Briefly summarized below are some hypotheses designed to link observed behavior toward artificial stimuli with the uses of vision in the
real world. A more complete discussion of this material has been
presented elsewhere ( Ingle, 196813, 1971).
The study of motion detection by goldfish (Ingle, 1967, 1968b) reveals
two or three independent processes that have been hypothetically
identified with aspects of natural visual behavior. The fast-moving single
spot is taken as a prototype of predator motion or the motion of a rival’s
markings used to direct aggressive attack (an eye or a body spot). The
higher probability of catching forward rather than backward moving
prey might explain the specialization for nasalward directions with a
fast-moving spot. Furthermore, it has been noticed that aggressive mouthbreeders, Tilupiu, seldom, if ever, attack a smaller fish moving in the
opposite direction, but they readily chase a rival who moves past in a
head-to-head orientation.
The slow-moving spot, seen best while moving backward, might correspond to a prey object being pursued; the ternporalward velocity would
inform the pursuer how rapidly the gap was being closed and help set