Two-sex populations are usually studied through frequency-dependent models that describe how sex ratio affects mating, recruitment and population growth. However, in two-sex populations, mating and recruitment should also be affected by density and by its interactions with the sex ratio. Density may have positive effects on mating (Allee effects) but negative effects on other demographic processes. In this study, we quantified how positive and negative inter-sexual interactions balance in two-sex populations. Using a dioecious grass (Poa arachnifera), we established experimental field populations that varied in density and sex ratio. We then quantified mating success (seed fertilization) and non-mating demographic performance, and integrated these responses to project population-level recruitment. Female mating success was positively density-dependent, especially at female-biased sex ratios. Other demographic processes were negatively density-dependent and, in some cases, frequency-dependent. Integrating our experimental results showed that mate-finding Allee effects dominated other types of density-dependence, giving rise to recruitment that increased with increasing density and peaked at intermediate sex ratios, reflecting tension between seed initiation (greater with more females) and seed viability (greater with more males). Our results reveal, for the first time, the balance of positive and negative inter-sexual interactions in sex-structured populations. Models that account for both density- and sex ratio dependence, particularly in mating, may be necessary for understanding and predicting two-sex population dynamics.