Deep radiation therapy with electron beams is now being performed at a number of institutions. Physical studies were first made with a 24-Mev betatron (18, 23, 35, 37–40, 67–69), but have subsequently been reported with the electron beam of an 18-Mev betatron (3–7, 12, 13, 21, 22, 25, 46–55, 64, 75, 78–80), a 100-Mev synchrotron (14, 62), a 35-Mev betatron (57, 61, 63, 65, 66, 82–87), a 45-Mev linear accelerator (8, 26, 27, 59, 70, 71, 76, 77), and 8- and 15-Mev linear accelerators (2, 15, 16, 20, 24, 81). Detailed reviews of electron beam therapy and dosimetry have been published by Laughlin (40–42) and by Becker et al. (6). An investigative program directed toward electron-beam radiation therapy was started at Stanford University in 1954 with the Mark IV 70-Mev microwave linear accelerator (1A) originally designed and built by the Microwave Laboratory of Stanford University for microwave and accelerator design studies. The present paper describes those studies on the physical properties of the electron beam which are of radiological significance. An accompanying paper reports the clinical aspects of the program (86). The Physical Arrangement Figure 1 shows the general layout of the accelerator, treatment room, and therapy control room. (The accelerator control console was at another level, not shown.) The electrons were accelerated in two sections, after which they passed through a double deflection magnetic system. The first (analyzer) magnet and the slit served to define the electron energy and the energy spread. The second (deflection) magnet separated the electron beam from neutrons and x-rays created at the slit. The magnetic field was monitored at the first magnet by a rotating coil fluxmeter, which was calibrated in terms of electron energy by photoneutron activation thresholds. The collimation system and accessories are shown in position in the treatment room in Figure 1 and in greater detail in Figure 2. The electrons emerged from the accelerator vacuum through an aluminum window, traversed the transmission ion-chamber monitor, and then were scattered to the desired angular width. After passing through a secondary emission monitor, the beam traversed a collimator made of heavy Presdwood. The walls of the upstream end of the collimator were 5 cm. thick, while those of the downstream end were 2.5 cm. thick, as shown in Figure 2. At the downstream end of the collimator were located a double, cylindrical chamber, a block of the Presdwood sufficiently thick to stop electrons of the highest energy in use at a given time, and a transparent plastic treatment cone. Forward and backward alignment lights aided in the precise set-up of patients and measuring equipment. A photograph of the collimator set-up is shown in Figure 3, and a mock patient set-up in Figure 4.