According to the concept of canalisation, infants and children stay within one or two growth channels, and therefore, any crossing of height centiles always warrants further evaluation. In view of evidence against this concept we re-investigated the variability of individual growth in the First Zürich Longitudinal Growth Study. The investigation is based on height measurements of 232 children (112 females, 120 males) measured at annual intervals during childhood and half-yearly during adolescence. Height data were transformed into height standard deviation scores (SDS) and canalisation defined by the width of an individual's growth channel, i.e., by the differences between maximum and minimum height SDS, in the individual series of measurements. Many subjects of the First Zürich Longitudinal Growth Study crossed numerous centiles with patterns that often seemed to show characteristic features. For approximately two thirds of the subjects, the SDS channel during the whole growth process covers more than one SDS. In childhood, between the age of two and age of minimal height velocity, only about one fourth of the subjects have an SDS channel below 0.5, indicating acceptable canalisation. During childhood, growth in boys appeared slightly more canalised than in girls (P = 0.02). Conclusion. The present investigation does not support the concept of strict canalisation of individual growth. We suggest to consider crossing of centiles a normal event in child development, though in a clinical setting crossing centiles should still be taken seriously, at least at first until a medical cause for this has been excluded.