Human lifespan has risen dramatically over the last 150 years, leading to a significant increase in the fraction of aged people in the population. Until recently it was believed that this contrasted strongly with the situation in wild populations of animals, where the likelihood of encountering demonstrably senescent individuals was believed to be negligible. Over the recent years, however, a series of field studies has appeared that shows ageing can also be observed for many species in the wild. We discuss here the relevance of this finding for the different evolutionary theories of ageing, since it has been claimed that ageing in the wild is incompatible with the so-called non-adaptive (non-programmed) theories, i.e. those in which ageing is presumed not to offer a direct selection benefit. We show that a certain proportion of aged individuals in the population is fully compatible with the antagonistic pleiotropy and the disposable soma theories, while it is difficult to reconcile with the mutation accumulation theory. We also quantify the costs of ageing using life history data from recent field studies and a range of possible metrics. We discuss the merits and problems of the different metrics and also introduce a new metric, yearly death toll, that aims directly at quantifying the deaths caused by the ageing process.