Recent work suggests that life-history strategies lie along a precise line of equilibrium between mortality, fecundity and growth. However, it has proved remarkably difficult to find a convincing ecological explanation for why different organisms are at different positions along this line. This is surprising because life-history traits are generally considered to be closely connected to fitness. Here, we test four candidate variables which represent competing ecological explanations for the origin of avian life-history diversity: food type; foraging range; developmental mode; and nesting habit. First, we find that over 90% of the variation in key life-history traits occurs lineages corresponding to the phylogenetic level of between families or above. This suggests that variation among living species is almost entirely due to events which occurred in the ancient evolutionary history of birds. Hence we use minor-axis regression models to quantify the direction and magnitude of ancient changes in `reproductive effort'. We then test for ecological correlates of these ancient changes. Of the four ecological variables tested, only changes in nesting habit are significantly correlated with ancient changes in `reproductive effort'. The adoption of safe nest sites among archaic birds is consistently associated with reduced `reproductive effort'. We suggest that life-history variation among living birds is largely due to diversification in nesting habit which occurred over 40 Ma BP, between the Cretaceous and Eocene.