We evaluated the intraobserver agreement of two techniques of amniotic fluid assessment: amniotic fluid index and maximum vertical pocket. The intraobserver variation of the two forms of amniotic fluid assessment was obtained by evaluating amniotic fluid prior to fetal biometry (post-biometry value) and after fetal biometry (pre-biometry value). Between January and June 1990, this amniotic fluid assessment was by the maximum vertical pocket technique, and between July and December 1990, by the amniotic fluid index technique. In the Grace Maternity Hospital, Vancouver, 352 subjects were evaluated by the maximum vertical pocket technique, and 184 subjects by the amniotic fluid index technique.Amniotic fluid volumes were categorized into normal, oligohydramnios, and polyhydramnios, and the pre-biometry and post-biometry categories of amniotic fluid index and maximum vertical pocket techniques were compared by kappa statistics. The maximum vertical pocket technique showed poor intraobserver agreement (kappa = 0.33), but the amniotic fluid index technique showed good agreement (kappa = 0.72). With the poor reproducibility of the maximum vertical pocket technique in assessing extremes of amniotic fluid volume, it is difficult to support its use as a component of fetal assessment. Copyright 1993 International Society of Ultrasound in Obstetrics and Gynecology