Although sexual selection theory has proved successful in explaining a wide
array of male ornaments, the function of ornaments occurring in females is
largely unknown. Traditionally, female ornaments have been considered nonf
unctional, being merely a genetically correlated response to selection for
male ornamentation. However, this hypothesis is only relevant to species in
which the ornament is basically the same in the two sexes. Alternatively,
female ornaments may be influenced by selection acting directly on the fema
les, either through female-female competition or male choice. We tested the
latter hypothesis in mate-choice experiments with two-spotted gobies (Gobi
usculus flavescens). In this small marine fish, females have bright yellow-
orange bellies during the breeding season, a conspicuous trait that is not
present in males. We conducted two aquarium experiments to test whether mal
es preferred to mate with more colorful females. In the first experiment, m
ales had a choice between two females that varied in natural coloration (an
d belly roundness). In the second experiment, we manipulated belly colorati
on and kept roundness constant. Males spent more time with colorful than wi
th drab females in both experiments and also performed far more courtship d
isplays toward colorful females. our study provides experimental evidence t
hat males prefer ornamented females in a fish that is not sex-role reversed
, supporting the hypothesis that female ornamentation is sexually selected.