Life history theory predicts that parents should desert a reproductive attempt if the costs of rearing this brood exceed the expected benefits. Thus, if the value of the current breeding attempt is reduced, for example through an unexpected reduction in size, parents are expected to reconsider whether it is worth continuing investing in their brood. With regard to nest desertion behaviour two predictions can be made: individuals are (1) more likely to desert if the reduction in clutch size is large and (2) less likely to desert if the reduction is at a late stage of breeding. We investigated the threshold at which nest desertion takes place by experimentally reducing great tit, Parus major, clutches to different sizes and at different stages of the incubation period. The results were in accordance with the predictions: clutch desertion rates were negatively related to the number of eggs that remained in the nest, and nest desertion was less likely nearer the end of the incubation period. In addition, we estimated the fitness consequences of nest desertion behaviour. For this purpose we made one group of birds desert in favour of a replacement clutch and another group rear a reduced brood. The latter were more likely to produce a second clutch after the first-brood fledglings had left the nest. As a consequence, the number of fledglings produced over the entire breeding season did not differ between the two experimental groups. We also counted the number of recruits and breeding adults in the following breeding season and found that the experimental groups did not differ in local recruitment and adult survival. Therefore, the results did not indicate that parents improved their fitness by deserting a reduced clutch.