1. Our objective was to determine why territorial polygyny occurs in t
he house wren Troglodytes aedon, a small, insectivorous, cavity-nestin
g songbird. In this paper, we present evidence that choosing a mated m
ale in our Wyoming (USA) study population is inherently costly to fema
le fitness. 2. In each of three breeding seasons, females who chose ma
ted males (i.e. 'secondary' = 2-degrees females) produced significantl
y fewer fiedglings per breeding attempt than females who chose unmated
males (monogamous (M-degrees) and primary (1-degrees) females). Secon
dary females were less likely than M-degrees/1-degrees females to comp
lete nesting attempts, primarily because they lost more broods to star
vation, and had more clutches and broods destroyed by conspecifics int
ruding onto territories. Secondary females also produced fewer fledgli
ngs in completed nesting attempts, primarily because they lost more of
fspring to starvation during the first two-thirds of the nestling stag
e. Nestlings raised by 2-degrees females weighed less than nestlings r
aised by M-degrees/1-degrees females late in the nestling stage. 3. Re
duced male aid in feeding young partially explains the lowered reprodu
ctive success of 2-degrees females. Whereas males provided substantial
aid to M-degrees and 1-degrees mates, most 2-degrees females received
little or no aid. The few 2-degrees females who did receive substanti
al male aid produced more fledglings than 2-degrees females without ma
le aid suggesting that the lowered success of 2-degrees females observ
ed on a population-wide basis did not solely result from 2-degrees fem
ales being younger or less-experienced breeders. 4. These results allo
w rejection of several hypotheses for teffitorial polygyny which assum
e that mate sharing is not costly to polygynously mating females.