Forty pairs of posterior bitewing radiographs taken at the third and fourth annual examinations of three-year clinical trials were assessed twice for caries by one examiner in independent recording sessions. Three or four degrees of caries were recorded in order to determine how different scales of measurement affected approximal caries progression patterns. The use of one degree of enamel caries, instead of two reduced by about half the percentage of lesions which appeared to have progressed during one year. Most lesions seemed to remain unchanged during the 12 months of observation and those which advanced usually progressed across one demarcation line. Reversals of caries diagnosis and regression of lesions to a lower degree of caries contributed least to the pattern of change between the two examinations. It is concluded that the radiographic scoring codes with two degrees of enamel caries give a better picture of caries progression than those with only one.