It was proposed almost 50 years ago that ageing is non-adaptive and is the consequence of a decline in the force of natural selection with age1. This led to the theory that ageing results from detrimental effects late in life of genes that act beneficially in early life1,2, so any genetic alteration that increases lifespan might be expected to reduce fitness, for example. We show here that a mutation that greatly increases the lifespan of the nematode Caenorhabditis elegans does indeed exhibit a fitness cost, as demonstrated during starvation cycles that may mimic field conditions, thereby validating the pleiotropy theory of ageing2.