>That was the view over 30 years ago.
Actually the research was done back in the 1930's.

This is some of what Edward O. Wilson wrote:

“The honeybee worker’s sense of visual pattern is quite bizarre to our own
way of thinking. Using the von Frisch training technique, Mathilde Hertz
(1930, 1935) discovered that the degree of dissection of a figure, rather
than its outline, is the quality perceived. For example, we easily perceive
that each of the figures in the upper row [here is shown a solid circle,
square, triangle] is radically different from the others, but the bee
cannot tell them apart. We also see the figures in the lower row [X, IIII,
Y] as differing greatly, but these too, the bee is unable to distinguish.
The bee is, however, able to tell any one of the figures in the upper row
from all of those in the lower row, and vice versa. Evidently what matters
is the number of borders, rather than their alignment. Gertrud Zerrahn
(1934) found that more precisely it is the length of the contours
surrounding a given area that is distinguished.”


Bibliography citations:

Hertz, Mathilde, 1930. Die Organisation des optischen Feldes bei der Biene,
II. Zeitschrift fur Vergleichende Physiologie, 11(1): 107-145.

--1935. Zur Phsiologie des Formen- und Bewegungsshens. III. Figurale
Unterscheidung und reziproke Dressuren bei der Bienc. Zeitschrift fur
Vergleichende Physiologie, 21(4): 604-615.

Zerrahn, Gertrud, 1934. Formdressur und Formunterscheidung bei der
Honigbiene. Zeitschrift fur Vergleichende Physiologie, 20(1-2): 117-150.

Regards,
Dick Allen