AimWe studied dynamics of four populations of New Zealand forest birds for 5–9 years after reintroduction to islands. We primarily aimed to predict whether these populations were viable, and what, if any, management was needed to maintain them. However, the small scale of these islands also provided an opportunity to study density‐dependent population growth over a short time frame.LocationWe studied New Zealand robin (toutouwai,Petroica australis) and stitchbird (hihi,Notiomystis cincta) populations reintroduced to Tiritiri Matangi, a 220‐ha offshore island near Auckland, and saddleback (tieke,Philesturnus carunculatus) and stitchbird populations reintroduced to Mokoia, a 135‐ha island in Lake Rotorua. These islands are free of mammalian predators, but have highly modified habitat following clearing and regeneration.MethodsWe closely monitored each population, individually marking all or most of the birds and in some cases experimentally manipulated population density or food supply. We used model selection procedures to understand factors affecting survival, fecundity and dispersal, and developed stochastic simulation models.ResultsThe Tiritiri Matangi robin and Mokoia saddleback populations grew without management and appear to be viable. Both showed strong evidence of density‐dependent growth, with fecundity (saddlebacks) and juvenile survival (both populations) declining with increasing density. Neither stitchbird population appears viable without management and supplementation experiments showed reproduction and/or survival to be limited by food supply. The Tiritiri Matangi population appears viable as long as supplementary feeding continues. However, the Mokoia population has a high mortality rate regardless of supplementary feeding, resulting in tenuous viability even with intensive management. Mokoia stitchbirds suffer from infection byAspergillus fumigatus, a pathogenic fungus that is prevalent in highly modified habitats and more abundant on Mokoia than Tiritiri Matangi.Main conclusionsSome forest birds can thrive in regenerating forest on islands and strong evidence of density dependence can be detected in such populations in as little as 5 years. This allows density‐dependent models to be developed, providing guidance when island populations are harvested for further translocations. Other species are limited by food supply in regenerating environments, a problem potentially overcome by management. However, prevalence ofA. fumigatusmay render highly modified environments uninhabitable by some species regardless of management.