Generalization of convex shapes by bees: what are shapes made of?

J Exp Biol

Department of Neurobiology, Institute of Zoology, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland.

Published: September 2005

For about 70 years, bees were assumed not to possess the capacity to discriminate among convex shapes, such as a disc, a square or a triangle, based on results of early studies conducted by presenting shapes on horizontal planes. Using shapes presented on a vertical plane, we recently demonstrated that bees do discriminate among a variety of convex shapes. Several findings, summarized here, provide indirect evidence that discrimination is based on a cue located at the shapes' boundaries. In the present study, we test this hypothesis directly in two different ways. (1) Three groups of bees are each trained with a different pair of convex shapes, one positive (rewarding), the other not (negative), producing colour contrast, luminance contrast or motion contrast against the background. The trained bees are then offered a choice between pairs of stimuli whose shapes are identical to those of the training shapes, but whose contrast against the background is varied by changing the pattern, the colour or the luminance of the areas. The results show that bees discriminate between the pairs of novel shapes, i.e. they generalize the shapes among the different types of contrast, revealing that they use a particular cue extracted from the positive shape. The bees' choices between a stimulus that produces the correct contrast but has the wrong shape and one that possesses the correct shape but the wrong contrast show, in addition, that the relevant cue is not located within the area of the shape. (2) Bees trained with pairs of convex shapes are tested with the same pairs of shapes, but which lack the inner area, i.e. only the contours or fragments of the contours are presented in the tests. Bees are found to prefer the stimulus whose contours (or fragments of contours) agree with those of the positive training shape. Taken together, the results suggest that convex shapes are not represented by the form of their areas but rather by some cue located at their boundaries.

Download full-text PDF

Source
http://dx.doi.org/10.1242/jeb.01790DOI Listing

Publication Analysis

Top Keywords

convex shapes
24
shapes
14
cue located
12
bees
8
bees discriminate
8
bees trained
8
contrast background
8
contours fragments
8
fragments contours
8
contrast
7

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!