Determination of offspring sex ratio is a fundamental reproductive decision that is predicted to have a major impact on an individual’s genetic contribution to future generations. We propose to test the key prediction of sex-ratio theory (Trivers-Willard hypothesis), namely that parents should adapt the sex ratio of their offspring to their local conditions; overproducing the sex that yields the highest fitness per unit of investment. Testing this prediction requires accurate quantification of how fitness costs and benefits depend on local conditions. In natural populations this is challenging, since fitness needs to be quantified in terms of reproductive values (RV), the long-term contribution of an individual to the gene pool of the population. We will undertake the first realistic quantitative test of the Trivers-Willard hypothesis in a wild population using the Seychelles warbler Acrocephalus sechellensis, a species that shows extreme variation in offspring sex ratios. To do this we will first develop new theoretical models, specifically tailored to the Seychelles warbler, that predict under what environmental and social circumstances individuals should bias the offspring sex ratio. Then, using lifetime data of all individuals from our long-term study spanning more than 30 years, we will test whether, under varying parental and environmental conditions, adjustments in sex ratios happen accordingly to predictions. Finally we will calculate the realised RVs of sons and daughters. By quantifying RVs, we will, for the first time, be able to fully understand the validity, extent and consequences of adaptive sex ratio modification in a wild vertebrate population.