The issue of neutron leakage has recently been raised in connection with dedicated electron-only linear accelerators used for intraoperative radiation therapy (IORT). In particular, concern has been expressed about the degree of neutron production at energies of 10 MeV and higher due to the need for additional, perhaps permanent, shielding in the room in which the device is operated. In particular, three mobile linear accelerators available commercially offer electron energies at or above the neutron threshold, one at 9 MeV, one at 10 MeV, and the third at 12 MeV. To investigate this problem, neutron leakage has been measured around the head of two types of electron accelerators at a distance of 1 m from the target at azimuthal angles of 0 degrees, 45 degrees, 90 degrees, 135 degrees, and 180 degrees. The first is a dedicated electron-only (nonmobile) machine with electron energies of 6 (not used here), 9, 12, 15, and 18 MeV and the second a conventional machine with electron energies of 6 (also not used here), 9, 12, 16, and 20 MeV. Measurements were made using neutron bubble detectors and track-etch detectors. For electron beams from a conventional accelerator, the neutron leakage in the forward direction in Sv/Gy is 2.1 x 10(-5) at 12 MeV, 1.3 x 10(-4) at 16 MeV, and 4.2 x 10(-4) at 20 MeV, assuming a quality factor (RBE) of 10. For azimuthal angles > 0 degrees, the leakage is almost angle independent [2 x 10(-6) at 12 MeV; (0.7-1.6) x 10(-5) at 16 MeV, and (1.6-2.9) x 10(-5) at 20 MeV]. For the electron-only machine, the neutron leakage was lower than for the conventional linac, but also independent of azimuthal angle for angles > 0 degrees: {[0 degrees: 7.7 x 10(-6) at 12 MeV; 3.0 x 10(-5) at 15 MeV; 1.0 x 10(-4) at 18 MeV]; [other angles: (2.6-5.9) x 10(-7) at 12 MeV; (1.4-2.2) x 10(-6) at 15 MeV; (2.7-4.7) x 10(-6) at 18 MeV]}. Using the upper limit of 6 x 10(-7) Sv/Gy at 12 MeV for the IORT machine for azimuthal angles > 0 degrees and assuming a workload of 200 Gy/wk and an inverse square factor of 10, the neutron dose equivalent is calculated to be 0.012 mSv/wk. For the primary beam at 12 MeV (0 degrees), the 10 x higher dose would be compensated by the attenuation of a primary beam stopper in a mobile linear accelerator. These neutron radiation levels are below regulatory values (National Council on Radiation Protection and Measurements, "Limitation of exposure to ionizing radiation," NCRP Report No. 116, NCRP Bethesda, MD, 1993).