Preterm birth and low birthweight in Canada have shown paradoxical temporal trends, with an increase in preterm birth and a decrease in low birthweight. Mean birthweight has increased in many industrialised countries, despite a recent rise in preterm birth, suggesting a temporal increase in fetal growth (birthweight for gestational age) in Canada. We thus described temporal trends in the distribution of fetal growth from 1981 to 1997, including means and proportions of infants at both the low and high ends of the fetal growth distribution. We used data for singleton live births from Statistics Canada's Canadian Birth Data Base for the years 1981-97 (excluding Ontario and Newfoundland) and analysed temporal trends in birthweight and birthweight-for-gestational-age z-score as continuous outcomes and the derived dichotomised outcomes [i.e. low birthweight (<2500 g), very low birthweight (<1500 g), small-for-gestational-age (<10th percentile), very small-for-gestational-age (<3rd percentile), high birthweight (>4000 g), very high birthweight (>4500 g), large-for-gestational-age (>90th percentile), and very large-for-gestational-age (>97th percentile)]. The birthweight-for-gestational-age was based on a newly developed population-based Canadian reference. The results showed that in the overall sample and in a subsample of term and post-term births, mean birthweight, mean z-score, rates of high birthweight, very high birthweight, large-for-gestational-age, and very large-for-gestational-age increased whereas rates of low birthweight, very low birthweight, small-for-gestational-age, and very small-for-gestational-age decreased between 1981-83 and 1995-97. The reverse was observed in preterm births. These temporal changes were larger for more extremely distributed measures of fetal growth. For example, compared with 1981-83, the decrease in 1995-97 for very small-for-gestational-age (<3rd percentile) was 38.9%, whereas the decrease for small-for-gestational-age (<10th percentile) was only 29.7%. Corresponding temporal increases were 21.4% for very large-for-gestational-age (>97th percentile) and 15.2% for large-for-gestational-age (>90th percentile). Among infants with gestational age 34-36 weeks, all measures of fetal growth, including the rates for all dichotomous outcomes, decreased in 1995-97 as compared with 1981-83. We conclude that Canadian infants are getting bigger, but only those born at term. The temporal trends for more extremely distributed fetal growth measures are particularly marked.