N-terminal propeptide of type III procollagen (PIIINP) concentration was measured in cord serum, amniotic fluid, and maternal serum from high-risk pregnancies. The fetal PIIINP variability was shown to be independent of the maternal serum PIIINP values. Although a highly significant negative correlation was found between the fetal propeptide level and gestational age in both appropriate-for-gestational-age neonates (n = 504) and small-for-gestational-age infants (n = 98), the PIIINP concentration in cord serum or amniotic fluid of small-for-gestational-age infants was significantly lower compared with that of appropriate-for-gestational-age infants matched for postconceptional age. PIIINP assay may thus serve as a dynamic biochemical indicator of deviant fetal growth. The PIIINP results were also related to the severity or duration of intrauterine growth retardation, as indicated by significantly lower propeptide cord serum values in nonmalformed small-for-gestational-age infants with small head circumference, known as an index for the chronicity of fetal nutritional deprivation. Preeclampsia, maternal diabetes or smoking, and congenital anomalies appeared not to be associated with any alteration of fetal propeptide concentration, provided they did not cause fetal growth deceleration. The finding of extremely high cord serum PIIINP values in six newborn infants with the Potter malformation sequence led to the speculation that large amounts of propeptides or their fragments usually are excreted by the fetal kidneys into the amniotic fluid. We suggest that determination of the PIIINP level in amniotic fluid or cord serum, obtained by amniocentesis and percutaneous umbilical sampling, may be a helpful adjunctive biochemical parameter in future research protocols assessing fetuses at risk for intrauterine growth retardation.