The results are described of prolonged corticosteroid treatment in 52 patients with pulmonary sarcoidosis observed for a mean duration of 10-5 years from diagnosis. Twenty of the patients -were treated for less than one year, the remainder for a mean of five years. The mean length of observation after discontinuing treatment was 5*5 years. Selection of patients for treatment was based on (1) evidence from radiographic observation, even for two years or more, that the disease was progressive so that spontaneous remission was improbable; and (2) evidence of disturbed lung function, especially the onset of dyspnoea or an impaired diffusing capacity. Treatment, conveniently with prednisone at a maintenance dose of about 15 mg. daily, should be continued for at least a year and often for much longer until withdrawal is not followed by radiographic relapse. Such treatmnent does not increase the chance of a complete remission even when given at an early stage of the disease; and relief of dyspnoea is unpredictable and seldom considerable. But prolonged treatment will halt progressive lung destruction and appears to curb the eventual mortality.