Two-hundred-seventy-eight retinoblastoma tumors in 169 patients were photocoagulated by one of two doctors with the xenon arc photocoagulator. More than 70% of tumors treated were cured by photocoagulation. The following features correlated with success: 1) size of the tumor, 2) location of the tumor and 3) elevation of the tumor. Seventy of 72 tumors (97%) up to 1 dd in size were cured while only 9 of 22 (41%) tumors larger than 5 dd were cured. The mean size of tumors cured was 2 dd, while the mean size of those that failed was 3.9 dd. Tumors located anterior to the equator were more often successfully treated (67/81 = 83%) than those posterior to the equator (44/74 = 60%). Tumors with the height equal to half the base were classified as 'low elevation', while those with heights larger than half the base diameter were classified as high elevation. Tumors with low elevation did significantly better (162/201 = 81%) than those with high elevation (33/76 = 43%). No correlation was found between success and the following features: age at diagnosis, sex, age at first photocoagulation, eye involved (right or left), time from radiation to first light coagulation, nasal versus temporal tumors or the clock hour of the tumor or category of tumor independent of size of location. Forty-four percent of tumors that failed photocoagulation went on to develop vitreous seeding and 55% required enucleation. Fifty percent of the eyes that were treated initially with photocoagulation went on to develop new tumor foci elsewhere in the eye. In all cases the new tumors appeared anterior to the equator. The children who developed additional tumors in the eye were younger (5.5 months) when photocoagulated than those who did not develop additional tumors (47.75 months). Photocoagulated retinoblastomas must be followed for at least three years before a cure is certain.