The differential allocation hypothesis predicts that parents should adjust their current investment in relation to perceived mate attractiveness if this affects offspring fitness. It should be selectively advantageous to risk more of their future reproductive success by investing heavily in current offspring of high reproductive value but to decrease investment if offspring value is low. If the benefits of mate attractiveness are limited to a particular offspring sex we would instead expect relative investment in male versus female offspring to vary with mate attractiveness, referred to as ‘differential sex allocation’. We present strong evidence for differential allocation of parental feeding effort in the wild and show an immediate effect on a component of offspring fitness. By experimentally reducing male UV crown coloration, a trait known to indicate attractiveness and viability in wild–breeding blue tits (Parus caeruleus), we show that females, but not males, reduce parental feeding rates and that this reduces the skeletal growth of offspring. However, differential sex allocation does not occur. We conclude that blue tit females use male UV coloration as an indicator of expected offspring fitness and adjust their investment accordingly.