We compared the relative value of the tear film breakup time, Schirmer test, lactoferrin immunologic assay and rose bengal staining in 156 patients with varying severity of symptoms characteristic of keratoconjunctivitis sicca and in 39 controls. Statistical analysis showed that in patients with minimal ocular irritation the Schirmer test in combination with the lactoferrin immunologic assay provided an optimal balance between high test sensitivity and low false-positive rates. In patients with moderate to severe ocular burning, foreign body sensation, or pain but with no systemic symptoms of dry mouth or arthritis, and in patients with both ocular and systemic complaints, the lactoferrin assay alone provided statistically significant results. Tear film breakup time and staining with rose bengal were not useful and the Schirmer test had limited value in measuring the rate of tear secretion in these patients.