The handicap hypothesis of honest signaling suggests that secondary se
xual characters reliably reflect phenotypic or genotypic quality of si
gnalers. This hypothesis is based on the assumptions that signals are
costly to produce and/or maintain and the cost of a given level of sig
naling is higher for low quality than for high quality signalers. We t
ested these assumptions in a field experiment in which the size of a s
econdary sexual character [tail length in male barn swallows (Hirundo
rustica)] was experimentally manipulated. Males were randomly assigned
to tail elongation, tail shortening, or two control treatments (tail
manipulation, or just capture, ringing, and handling). Male barn swall
ows were challenged with an injection of sheep red blood cells, and bl
ood was sampled on the day of first capture and after 3 to 4 weeks for
determination of concentrations of gammaglobulins. Tail-elongated mal
es did not increase levels of gamma-globulins while males of the other
three groups demonstrated increases. Analyses of variation in gamma-g
lobulins within treatment groups revealed a positive correlation betwe
en gammaglobulins and original tail length among males with elongated
tails. These results suggest that tail length imposes an immunocompete
nce cost on males, and that males with naturally long tails are differ
entially better able to cope with this cost.