An experimental model is suggested for reproducing ocular melanoma in New Zealand white rabbits using B16 melanoma cells and protocols differing with respect to either tumour origin (subcutaneous fragments of melanoma B16 or B16-F10 tumour cell cultures) or implant site (the anterior chamber or subchoroidal). In 20 animals, 20 mg of methylprednisolone acetate was injected subconjunctivally as a local immunosuppressant. The only protocol resulting in tumour was inoculation of 4 x 10(6) B16-F10 melanocytes into the anterior chamber of the eye. Trans-scleral injections of cell suspensions produced tumour growth in 43% (13/30) of animals so treated. Thirteen animals developed non-neoplastic pigmented lesions formed of numerous melanophages. Another 19 animals showed non-pigmented lesions caused by reaction to the surgical procedures. Subconjunctivally injected methylprednisolone acetate did not increase the incidence of tumour growth.