Two staining techniques for conjunctival impression cytology (CIC) were evaluated in both eyes of 106 subjects with either early clinical signs of xerophthalmia (17%) or without any obvious ocular disease (83%). One technique consisted of eleven steps of staining while the other of two steps. The slides were evaluated independently in a masked fashion by two readers using three criteria for evaluation of CIC, viz., goblet cells, mucin spots and epithelial cells. By either method, both readers had greatest agreement in evaluation of the goblet cells, the percentage difference being statistically not significant. Errors in interpretation of CIC can be minimized if goblet cells alone are considered as the major criterion of evaluation.