Female-limited polymorphism is often attributed to selection to avoid excessive male mating attempts. It is encountered in various taxonomic groups, but is particularly common in damselflies, where one female morph (andromorph) typically resembles the conspecific male in colour pattern, while the other(s) (gynomorph(s)) do not. Two sets of theories have been proposed to explain the phenomenon in damselflies, which can be classified as the learned mate recognition (LMR) and male mimicry (MM) hypotheses. To test predictions of these hypotheses, we evaluated the rate of male sexual response towards female morphs and conspecific males in the damselfly Nehalennia irene. The LMR hypothesis predicts that males should respond sexually to andromorphs at greater rates in populations containing a higher relative frequency of andromorphs. The MM hypothesis predicts that males respond more often sexually to both andromorphs and males as the ratio of andromorphs to males increases. While LMR predicts that the rate of mating attempts towards gynomorphs should vary, the MM predicts that it should be relatively fixed. On experimentally presenting live specimens to focal males in five different populations with extreme variation in female morph frequencies, we observed that as the andromorph frequency and ratio of andromorphs to males increased, the proportion of male mating attempts increased on both andromorphs and males, whereas it decreased on gynomorphs. While the simplest form of the MM hypothesis is rejected, the results support specific predictions of both hypotheses and suggest that future studies should not treat these hypotheses as mutually exclusive.