Human chorionic gonadotropin (hCG) stimulation test is a reliable dynamic test for the evaluation of testicular function during childhood. Several protocols have been recommended but their reliability is controversial. In order to decide the best timing to measure stimulated testosterone levels in short- and long-term hCG protocols, we evaluated 83 prepubertal patients in two group. In group A, 34 patients with isolated micropenis and in group B, 49 inguinal cryptorchidic patients were enrolled. In group A short-term hCG protocol (3000 IU/m2/im/3 days) and in group B long-term hCG protocol (1500 IU/m2/im; thrice a week for 3 weeks) was administered. Blood samples were drawn at the initiation of the test and then at the 1st and 4th days after the last hCG injection. Each case's peak stimulated testosterone (Tmax) and the incremenet in plasma testosterone (deltaT) were calculated and compared with the 1st and 4th day responses within the group. In the short-term protocol the 4th day responses were higher than the 1st day responses (p<0.01). Interestingly, while four patients had insufficient responses at the 1st day, three had sufficient Leydig cell response at the 4th day. In the long-term protocol group, in contrast to the short-term group, the 1st day responses were higher than the 4th day (p<0.01). According to our results, while performing hCG test in a patient, if a short-term protocol is planned, it is more convenient to check the 4th day testosterone response. On the other hand, in a long-term protocol it is best to check the 1st day response. We suggest that even if a patient's 1st day response is inadequate, the 4th day response should be checked in order to avoid misdiagnosis.