PROBLEMS IN THE MEASUREMENT OF RAINFALL — AN EXPERIMENT AT ILORIN, NIGERIA