The value of gastric aspirate analysis in screening for perinatal infection is disputed. We have studied the cellularity and bacterial content of 231 aspirates from 105 normal, infected and asphyxiated infants, with the following conclusions: (1) A positive culture has no pathological significance. (2). Aspirate cellularity does not correlate with bacterial culture results. (3) An excess of polymorphonuclear leucocytes (PMNL) in aspirates collected at birth is a pathological but non-specific stress response. (4) After birth there is a physiological rise and fall in the aspirate PMNL content parallel to that in the peripheral blood. Aspirate cellularity and bacterial culture are therefore of limited value in the diagnosis of neonatal pneumonia.