Loading…

Generalization and categorization of spectral colors in goldfish. II. Experiments with two and six training wavelengths

In part I of this study (Kitschmann and Neumeyer 2005), goldfish categorized spectral colors only in the sense that wavelengths in a range of about twice as large as the just noticeable difference were treated as similar to a given training wavelength. Now, we trained goldfish on more than one wavel...

Full description

Saved in:
Bibliographic Details
Published in:Journal of Comparative Physiology 2006-05, Vol.192 (5), p.469-479
Main Authors: Poralla, Julia, Neumeyer, Christa
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In part I of this study (Kitschmann and Neumeyer 2005), goldfish categorized spectral colors only in the sense that wavelengths in a range of about twice as large as the just noticeable difference were treated as similar to a given training wavelength. Now, we trained goldfish on more than one wavelength to prevent very accurate learning. In one experiment goldfish were trained on six adjacent wavelengths with equal numbers of rewards, and, thus, equal numbers of learning events. Generalization tests showed that some wavelengths were chosen more often than others. This indicated that certain spectral ranges are either more attractive or more easily remembered than others. As this is a characteristic of the “focal” colors or centers of color categories in human color vision, we interpret the findings in goldfish accordingly. We conclude (Figs. 5 and 6) that there are four categories in spectral ranges approximately coinciding with the maximal sensitivities of the four cone types, and three categories in-between. Experiments with two training colors indicate that there is no direct transition between categories analogous to human “green” and “red”, but that there is a color analogous to human “yellow” in-between (Figs. 2, 3; Table 1).
ISSN:0340-7594
1432-1351
DOI:10.1007/s00359-005-0082-5