Renal weight (left-right combined), as a parameter of renal development, is required to be less than half the normal value for age for a statistically confident diagnosis of hypoplasia. "Medullary ray glomerular counting" (MRGC), counting cortical glomerular generations, has been proposed as a simple technique of possibly greater sensitivity. Recent development of the Disector method for the unbiased stereological estimation of total glomerular number has provided a, hitherto unavailable, "golden standard" with which to determine the diagnostic potential of MRCG. Both "true" (actual number of generations seen) and "assumed" (a subjective "guess" of the total number of generations) MRGC counts were determined in 11 pairs of kidneys from spontaneously aborted, normally developed, non-malformed fetuses (gestational age: 15-40 weeks). Each kidney was randomly analysed blind and on two separate occasions by two paediatric pathologists using a written protocol. Results were compared with unbiased stereological estimates of glomerular number. Intra- and inter-observer and intra- and inter-(left-right)renal reproducibility were analysed. In conclusion, MRGC, using "real" counts, is a highly reproducible parameter of renal development from 15-36 weeks' gestation. Sensitivity for detection of both hypoplasia and maturation delay increase with gestational age and generally exceeds that of renal weight.