Eosinophilia is a common finding in premature babies during the neonatal period. Serial eosinophil counts were determined in 94 hospitalized, appropriately grown premature in fants whose gestational ages ranged form 28 to 36 weeks. The incidence, severity and etiologic factors of eosinophilia were retrospectively studied in premature infants, who were divided into three groups according to their gestational age. The results were as follows: 1) Absolute eosinophilia (>700/mm3)was documented in 46.8%(44/94). 2) The duration of TRN and antibiotics was cignificantly higher in infants with eosinophilia than withour eosinophilia (p<0.05). The gestational age of the infants with eosinophilia was significantly shorter than that of the infants without eosinophilia (p<0.05). The infants with eosinophilia started with bottle feeding significantly earlier than the infants without eosinophilia (p<0.05). 3) The infants younger than 30 weeks of gestational age have greater incidence of eosinophilia (75%) than the infant with the gestational age between 34 and 36 weeks (34.3%)(p<0.05). 4) The incidence of mild eosinophilia was higher in the group with gewtational age 30 weeks or below(37.5%)than in the group with gestational age between 34 and 36 weeks(17.1%)(p<0.05).Also, the incidence of severe eosinophilia was significantly higher in the group with gestational age 30 weeks or below(37.5%)than in the group with gestational age between 34 and 36 weeks(17.1%)(p<0.05). 5) Eosinophilia was more prevalent in the infants who received parenteral nutrition composed of glucose, amino acid and lipid(77.8%) than the infants who received only glucose(34.1%)(p<0.05). 6) Gestational age and birth weight were significantly lower in infants with severe eosinopilia than those of mild or moderate eosinophilia. The duration of TRN and antibiotics were significantly longer in infants with severe eosinophilia than those of mild or moderate eosinophilia (p<0.05). 7) The peak eosinophil count was observed significantly later in severe eosinophilic group (26.5 13.1 days)than in non-eosinophilic group(14.4 9.0 days)(p<0.05). The above results suggest that eosinophilia in premature infants may be the effecto of immature immunologic responses to the intravenous administration of extemal antigens like amino acid, lipid and antibiotics.