ABSTRACT.From two large maternity hospitals, information on the birth dimensions of liveborn infants of birthweight from 501 to 1500 g was coded for computer analysis; cohorts of inborn infants delivered in the years 1971 to 1974 and 1977 to 1978 were included in the study if there were both reliable obstetrical information on gestational age and also the clinical estimation of maturity of the neonate did not differ by more than two weeks from the obstetrical estimation.The frequency distribution of birthweights of infants indicated that only in the range from 24 to 29 weeks was the sample sufficiently complete to permit further analysis. Mean ± 2 SD and weighted centiles were calculated for birthweight, body length and head circumference; these were tabulated and displayed graphically. The 50th percentile for weight was slightly higher than that previously reported from Melbourne, but appreciably lower than some other published data. Preparation of separate charts for male and female infants was not considered to be justified