1 We examined the validity of classifying tree species as early, mid-, or late-successional based on age and height-growth rates, by comparing the age and height-growth rates of trees in the boreal forest. 2 Age was first examined using the traditional method of coring 30 cm above the root collar; then dendrochronology was used to locate the root collar and missing annual growth rings. Traditional ageing differentially underestimates tree age; species classified as early successional (Populus tremuloides, Betula papyrifera, and Pinus banksiana) are less severely underestimated than those classified as mid- and late-successional (Picea glauca, Picea mariana, and Abies balsamea) (0—11 vs. 0—43 years), and also have relatively fewer locally missing growth rings. Ageing at the root collar shows that all tree species recruit within 5—10 years after fire and age cannot therefore be used to determine successional status. 3 Mean time taken to grow to each 1-m increment from the root collar was estimated for each species. Species classified as early successional have relatively higher growth rates between the root collar and the first metre; they are therefore less severely underestimated when aged above the root collar, explaining why they often appear older than species classified as mid- and late-successional. The lack of species differences above 1 m means that height-growth rates cannot be used to classify these tree species as early, mid-, or late-successional. 4 In the boreal forest of Saskatchewan, the rapid recruitment of all tree species after fire, and the short fire cycle mean that the forest dynamics between catastrophic wildfires are driven primarily by the mortality rates of each species. © Blackwell Scientific. Abstract reproduced by permission.