Cross-sectional data on birth weight and gestational age at birth are commonly used to construct a longitudinal curve of fetal growth. There is no direct evidence that such a curve actually represents intrauterine growth. Nonetheless, this hypothesis of a "fetal growth curve" is well developed and widely accepted. This paper describes certain incongruities in the "fetal growth" interpretation of these curves which have not been previously discussed. An alternative approach to these curves which makes no assumptions about fetal growth is able to resolve these incongruities. It is concluded that the "fetal growth curve" hypothesis may be an impediment to a full understanding of the relation between birth weight and gestational age at birth.