Attenuation factors in water have been measured by a narrow-beam technique in various portions of x-ray beams with nominal energies of 6 and 25 MV, with and without a wedge in the beam. The results were analyzed in terms of an attenuation coefficient mu for small water thicknesses and a beam-hardening coefficient eta that describes the change in attenuation per unit depth. The variation of mu within the field was significant, about 0.5% per centimeter at 6 MV and 0.8% per centimeter at 25 MV for open beams. The heavy wedge used in these experiments caused significant (about 10%) beam hardening at 6 MV, softened the beam somewhat at 25 MV, and increased the variation of mu within the field to 3%-5%. These effects should be taken into account in dose calculations, and correction factors can be designed based on the variation of mu with off-axis radius for open beams and with off-axis position for wedged beams. The experimental technique, based on two measurements with the beam going through a water tank with either 26- or 50-cm path length, was simple and highly reproducible. The beam hardening with depth in water, i.e., the value of eta, was readily determined but found to be clinically insignificant.

When the collimator setting of a linear accelerator is made sufficiently small, the output factor in air, R, is greatly reduced because the collimators obstruct the periphery of the x-ray source. This has been utilized to examine the size of the source by varying the width y of a narrow field and determining how R(y) varies. The sources diameters in the two principal directions were clearly influenced by the design of the accelerators. The x-ray sources of two accelerators with bending magnets were found to be noncircular while that of a linear accelerator without a magnet showed circular symmetry. The position of the source relative to the axis of collimator rotation was determined by measuring R for offset narrow fields. For one of the accelerators, the source was initially moving and off the central axis by about 2 mm for the first five monitor units. The results correlated well with sharpness in portal-film images. The technique can serve to evaluate the major source characteristics in acceptance testing and quality control.