Hypophysectomy is an accepted procedure for palliation in metastatic carcinoma of the breast. In several series reported, however, the incidence of complications due to the procedure has been high. A modified technic for transnasal yttrium-90 hypophysectomy has been developed in an effort to minimize complications and reduce irradiation exposure to the operators. Dosimetry studies indicated that precise placement of the radioactive sources is necessary, not only to destroy the anterior lobe of the pituitary but to avoid visual complications secondary to over-irradiation of the optic nerves. The film-pack dosimetry method of Freitag et al. (1) for beta emitters is employed. Two clumps containing six yttrium-90 beads each are arranged with their centers 6 mm. apart—a geometric arrangement considered to be ideal for pituitary ablation. A pack of Dupont 555 dosimetry film is prepared with paper spacers so that readings can be obtained at intervals of 1.0 mm. from zero to ten. The pack is placed on the beads for a predetermined time. After development, the films are read on a photovolt transmission densitometer and dose values are calculated. When twelve beads were used, with a total activity of 14.4 mc, films from zero to 3 mm. were too dark to read. At 4 mm., the calculated dose to decay is approximately 100,000 rep; at 7.5 mm., 10,000 rep; at 10 mm. 1,000 rep. The dark area on the films also reveals the pattern of the dose curve in anteroposterior and lateral dimensions. There is much variation in size and shape of the sella turcica and in the location of the optic chiasm. James Bull (2) has shown the normal variations in position of the optic recess of the third ventricle, which he designates as the “chiasmatic point.” Because of the many variables, he introduced the “nasion-tuberculum” plane—a line drawn from the tuberculum sella back to the posterior clinoid, parallel to the floor of the sella—as a reference for measurement. Measuring the height of the optic recess above the plane in 117 normal encephalograms, he found that the bulk of the cases fell in the 8- to 13-mm. range. In 6 cases, however, the measurement was 4 to 6 mm. In applying depth dose values to Bull's measurements, it becomes obvious that the radioactive beads must be placed well below the “nasion-tuberculum” plane. If this is done, the dose to the chiasm in most instances will be less than 1,000 rep. We consider the stereotaxic method to be the most accurate means of placing the radioactive sources. The Rand-Wells stereotaxic device was designed to afford complete control of guides, drill, and needles. At the same time it does not interfere with the endotracheal tube for anesthesia. The patient's head is immobilized, and identical anteroposterior and lateral radiographs can be obtained. Two radiographic tubes are used. Adjustable guide holders are calibrated in degrees. When one guide is aligned, the second can be placed quickly in the same relative position.