In a controlled trial 69 children with corneal xerophthalmia were given 200 000 IU oil-miscible vitamin A by mouth and a matched group of 45 children were given 100 000 IU water-miscible vitamin A intramuscularly. Both groups received an additional oral dose the next day. There was no detectable difference in the clinical response to the two regimes, even when analysis was limited to patients with concomitant diarrhoea or protein-energy malnutrition. Although serum-vitamin-A levels were significantly higher after parenteral than oral therapy, holoretinol-binding-protein levels were not. Oral administration of vitamin A is not only more practical but appears to be just as effective as parenteral administration in the treatment of severe xerophthalmia.