- Unpublished doctoral dissertation, University of Hawaii.
Visual discrimination of relative distance by a bottlenosed dolphin (Tursiops truncatus)—Evidence for relation learning - James R. Mobley
University of Hawaii, Kewalo Basin Marine Mammal Lboratory, 1129 Ala Moana Blvd., Honolulu, HI, USA, 96814 -
- (C) 1984 University of Hawaii
A series of seven experiments examining visual discrimination capabilities of a bottlenosed dolphin (Tursiops truncatus) were performed at Sea Life Park, Makapuu, Oahu during the period January, 1982 through May, 1983. The object of all the visual discrimination experiments described here was to discriminate the nearer of two visual targets presented in air. In Experiment 1, the distance between the pair of targets (rectangular cards) was varied with each inter-target distance appearing equally and in random order, as per the method of constant stimuli. Results showed performance to be under stimulus control of the amount of separation between the targets in depth. Further a depth acuity estimate of 16.2 min of visual arc was derived—a value comparable to estimates of monocular depth acuity of other animals shown to possess steropsis. Experiment 2 revealed performance levels to remain high (i.e., above 90% correct) for targets equated for retinal size differences at their respective viwing distances, thus suggesting that the relative size of the retinal images was not the crucial cue mediating discrimination of the nearer target.
Conceptual transfer of the relational rule (i.e., “go to the nearer target”) was demonstrated by generalization from the previous two-dimensional class of targets (rectangular cards) to a novel three-dimensional class (styrofoam balls) in Experiment 3, and across a series of 24 identical pairs and 16 dissimilar pairs of two- and three-dimensional targets in Experiments 4 and 5, respectively. Such evidence of learning transfer has been wholly absent from previous experiments involving in-air discrimination using static visual targets. Experiment 4 also demonstrated some perceptual constraints in making judgements of relative distance in that objects subtending less than 3&Mac176; of visual arc as well as vertical linear targets (wooden dowels) resulted in subthreshold performance levels. In Experiment 6, the stimulus configuration was altered through a series of rotations from horizontal to vertical alignment of the visual targets. Performance remained above criterion of 80% correct for all four rotation positions used (0&Mac176;, 30&Mac176;, 60&Mac176;, and 90&Mac176;). These results contain further significance in that at 90&Mac176; rotation, the stimuli were equidistant from the response manipulanda, thereby requiring mediation of the conditional rule, “if nearer target on top, go right; if nearer target on bottom, go left”; thus further demonstrating a higher-order of learning flexibility than is normally characteristic of visual-based tasks using bottlenosed dolphins.
Based on features of the visual display apparatus as well as consideration of anatomical details of the animal in question, the cueing dimensions mediating the task (in the absence of size cues) were thought to be differential motion parallax, accomodation cues (differential “blur”), or differential brightness cues. In further efforts towards cue reduction, the relative importance of motion parallax cues was tested in Experiment 7 through comparison of stationary viewing (via bite-bar apparatus) versus free-viewing discrimination performance. Results indicated superior free-viewing performance (though statistically non-significant) which suggested the additional depth-resolving power of motion parallax. Examination of one informal videotaped session, however, revealed more than half of the fixed-viewing trials to be “contaminated” by movement of the subject. Thus, motion parallax cues may have been present throughout the fixed-viewing condition.
Overall, these results, in conjunction with reviewed literature, strongly suggest the bottlenosed dolphin to be relatively more “prepared” to extract relative distance information from features of the environment than visual form information. It was suggested that visual estimations of short-range distances may be of biological importance to this species, perhaps complementing the directional limitations of the biosonar when closing in on elusive prey.
- Mobley, J. R. (1984). Visual discrimination of relative distance by a bottlenosed dolphin (Tursiops truncatus)--Evidence for relation learning. Unpublished doctoral dissertation, University of Hawaii. Honolulu.
Back to Top Dolphin Programs | Whale Programs | Education Programs | Our Research | Resource Guide Copyright © 2002, The Dolphin Institute |
