Animals assess the quality and quantity of food and choose among diffe
rent foods based on these assessments. We explored whether there was g
enetic variation for assessment of pollen quality by foraging honey be
es, Apis mellifera. Honey bees derived from two genotypic strains fora
ged for pollen of varying quality from a petri dish placed inside an o
utdoor flight cage. The strains were the result of a colony-level, two
-way selection on amount of stored pollen. We used the forager's round
dance to quantify the assessments of pollen quality by individually m
arked worker bees. The dance rate (number of 180 degrees turns per min
ute) and the probability of dancing were each greater when bees forage
d for pure pollen compared with a lower-quality mixture of pollen and
alpha-cellulose (1:1 by volume). Bees from the high-pollen genotypic s
train had a higher dance rate than those from the low-pollen strain, s
uggesting different assessments. Bees from the low-pollen strain, howe
ver, had a higher probability of dancing than did bees from the high-p
ollen strain. Dance duration was not affected by a bee's strain or by
the quality of pollen. We conclude that the dance rate may be used to
quantify a forager's subjective evaluation of pollen quality and that
this evaluation has a genetic component. Our results also suggest that
the dance may function at the colony level to recruit bees to more pr
ofitable pollen sources. (C) 1998 The Association for the Study of Ani
mal Behaviour.