The Apgar score is widely used for several purposes: to discriminate between infants who require resuscitation at birth and those who do not; to predict outcome; and to evaluate change in the condition of the newly born over the first minutes of life. Using published evidence of its clinical reliability and validity, this article explores whether the Apgar score serves all three measurement purposes equally well. Methodologic guidelines for assessing health indices are applied to examine the structure of the Apgar score as well as its function where performance data are lacking or inadequate. Despite the advent of modern technology, the Apgar score remains the best tool for the identification of newly born infants in need for cardiopulmonary resuscitation. For predicting later death or handicap, the Apgar score is insensitive but fairly specific. The ability of the Apgar score to measure change over time has not been studied systematically; however, available data suggest that serial Apgar ratings in infants with early low scores detect clinically important recovery of lack thereof.