This study was designed to plot the motor development of infants with cerebral palsy (CP) from four to 12 months of age in terms of scores on the Gross Motor Function Measure (GMFM). Twelve infants at risk of CP and one infant with no abnormal history were tested weekly using the GMFM. An analysis of the scores of all children suggested very variable patterns of development and in some cases, a large fluctuation in scores from one week to the next. This fluctuation was also evident in the infant who was neurologically intact. Infants seemed to fall into one of three groups. The most severely affected group started with a score of less than 10 on the GMFM at the initiation of the study and gained scores at a mean rate of 0.21 per week. The next group started with scores between 19 and 30 and had a mean 0.34-point improvement per week. A characteristic of this group was that scores peaked between 30 and 40 weeks and were not equalled or exceeded by 52 weeks. The infants who demonstrated the most improvement started with scores higher than 23 and developed skills at the rate of one point per week. These patterns are in contrast with the infant who was neurologically intact and who started at 50 points at 19 weeks and whose score improved at the rate of 2.9 points per week throughout the year. It was not possible to identify variables that would predict (at the initiation of the study) the different patterns of development. Scores on the GMFM do not increase in a linear fashion, even in children who could be expected to develop “normally.” Because so little improvement was noted in some children (0.84 points per month in one group, 1.36 points per month in another group) counseling and maintenance might be legitimate aims of physical therapy treatment in such cases. Repeated measures of the dependent variable should be incorporated into monitoring of treatment effects of infants with CP as single pre- and postmeasures may be misleading.