The main objectives of the study were to find the incidence and risk factors associated with development of hypoglycemia in small for gestational age (SGA) babies, to compare haemoglucotest strips (Boehringer Mannheim) with the standard laboratory method (glucose-hexokinase) for estimation of blood glucose, and to measure the insulin and cortisol responses of SGA babies. This was a prospective longitudinal study. The sample included SGA babies and over a period of six months, 127 consecutively born small for gestational age babies were investigated prospectively for development of hypoglycemia in first 48 hrs of life. Plasma samples were taken during episodes of hypoglycemia for insulin and cortisol estimation and compared with non-hypoglycemic controls. The overall incidence of hypoglycemia was 25.2% in SGA babies and 98% of the episodes occurred within first 24 hrs. Compared to non-hypoglycemics, mothers of hypoglycemic babies had higher incidence of receiving i.v. fluids (5% dextrose) during labour. The hypoglycemic babies were more likely to be sick and oral feeds had been initiated by one hour of life in only 37% of them compared to 63% of non-hypoglycemic babies. Plasma insulin/glucose ratio was significantly higher in hypoglycemic than non-hypoglycemic babies, whereas the cortisol levels were similar. Small for gestational age babies are highly prone to develop hypoglycemia in first 24 hrs of life.