In the July 1977 issue (Pediatrics 60:124, 1977) an excellent plea for appropriate use of statistics in descriptive studies is accompanied by a glaring example of their misuse (Pediatrics 60:60, 1977). Analysis of the serum gastrin levels reported assumes a normal distribution of values, which is obviously not the case on inspection of either the raw data or the reported SEM, which is larger than the mean (22.5 ± 23.9 after an eight-hour fast). Using the data in Figure 2 and repeating the calculations, it becomes obvious that the values reported as SEM are, in fact, SDs.