| Abstract: | The causal relationships between shoot growth variables were analysed on the basis of a study in the Phytotron where six full‐sib families of Picea abies L. (Karst.) were grown for six growth periods. Phenotypic correlations and path analysis were used for the analysis of causes. The phenotypic correlations were estimated in three different ways: 1) separate analyses within each growth period based on individual tree data, 2) based on family means at different growth periods (0, and 3) using differences in family means between two consecutive growth periods (dt). The two methods based on t or dt‐values gave different results in a number of cases. We argue that, because of the rapid concurrent changes that occurred between growth periods, we could not conclude that the correlations based on family means (t) reflected causal relations and, also, that the method based on dt‐values gives a better indication of causal relationships between different variables. Using dt‐values, the variation in shoot elongation period was mainly explained by the variation in free growth (82%) while the variation in shoot growth initiation only explained a minor fraction. Variation in leader length was mainly explained by growth rate and only to a minor extent by shoot elongation period, meaning that the observed decrease in shoot elongation period did not cause a corresponding changé in leader length. One conclusion from this finding is that we do not necessarily lose growth capacity if we breed for a better adapted growth rhythm (i.e. shorter shoot elongation period) of the seedlings. |
