An intrauterine growth-retarded infant is commonly defined as one weighing less than the 10th percentile in birth weight for its gestational age. However, because there is no standard population from which to derive these percentiles, the birth weights that serve as the cutoff point in various published studies may differ by several hundred grams at any gestational age. For this reason, we examined the studies from which the currently used 10th-percentile standards are derived to determine which factors may be responsible for the variation. In addition to obvious differences in the populations and geographic areas on which they were based, studies differed in how gestational age was determined, whether the gestational age was "rounded" or given in completed weeks, which types of infants were excluded, whether the studies were hospital or population based, and whether they were controlled for sex of the infant and race and parity of the mother. These differences in study methodology may be as or more important than the population differences in defining the 10th percentile cutoffs. A single national standard for intrauterine growth retardation would allow comparison between studies of risk factors, diagnostic tests, management, and long-term follow-up status of fetuses and infants with intrauterine growth retardation.