Prospective analysis in the first 24 hours of life of 116 consecutively admitted infants with respiratory distress allowed early identification of nine cases of earlyonset neonatal sepsis. The following significant differences (P < .05) between septic and nonseptic groups were observed: (1) leukopenia (WBC count ±SEM in septic infants, 6,370 ± 1,140/cu mm vs. 16,020 ± 710/cu mm in nonseptic infants; (2) absolute neutropenia (band forms plus polymorphonuclear cells, 2,840 ± 1,220/cu mm in septic infants vs. 9,100 ± 620/cu mm in nonseptic infants); (3) elevated band form/total neutrophil ratio (0.30 ± 0.09 in septic infants vs. 0.14 ± 0.01 in nonseptic infants); (4) increased nitroblue tetrazolium incorporation or absence of neutrophils on nitroblue tetrazolium smear in septic infants; (5) positive gastric aspirate Gram stain (5 of 7 septic infants vs. 23 of 83 nonseptic infants for polymorphonuclear cells, 5 of 7 vs. 18 of 83 for presence of bacteria); (6) an examination for buffy coats revealing organisms (4 of 4 in the septic group vs. 0 of 105 in the nonseptic group). Using leukopenia (WBC count < 10,000/cu mm) or neutropenia (polymorphonuclear cells plus band forms < 4,500/cu mm) as criteria, eight of nine septic infants and two suspect septic infants would be identified, 15 of 105 nonseptic infants would be so labelled. Hence, absolute leukopenia or neutropenia is a simple and reliable indicator for the identification of early-onset sepsis in infants who have respiratory distress in the first few hours of life.