Patients with clinically well defined keratoconjunctivitis sicca (KCS) participated in two trials of hypo-osmolar tear substitutes. The trials were double masked, single crossover studies with computer generated random order allocation and were conducted by postal questionnaire. In the first trial a hypo-osmolar formulation using polyvinylpyrrolidone and hydroxyethyl cellulose was compared with the patients' usual treatment (BJ6 or hypromellose) and found to be inferior, especially in respect of blurring and stickiness caused by higher viscosity. In the second trial hypo-osmolar preparations of BJ6 and hypromellose were compared with their iso-osmolar equivalents and found not to be significantly different in their effects. Hypo-osmolarity alone does not seem to guarantee relief of symptoms in KCS. Other factors including viscosity and colloid osmotic pressure may be more significant.