The accuracy of determining marginal bone height changes around osseointegrated implants depends on the validity of comparing serial films and the reliability of the measurements. X-ray beam orientation changes can alter the validity of serial films. A human dry mandible containing a Brånemark implant was irradiated ±12 degrees in the vertical plane at 1-degree intervals to the perpendicular to the long axis of the fixture. The thread width was recorded on both sides of each fixture image using a computer. Twenty-five randomized unclassified images were remeasured and the vertical angle of the x-ray beam was estimated from the previous measurements to test for validity of comparing images. The reliability of measurements with altered image magnification and penumbra were calculated. The reliability of 24 repeated thread width measurements was a SD of 0.01 mm. Of the 25 unknown beam angulations, 32 percent matched correctly, 20 percent ± 1 degree, 16 percent ± 2 degrees, or 68 percent <m ±2 degrees. Alteration from a short to a long cone technique was estimated to produce magnification errors similar to the reliability SD of 0.01 mm. Similarly the penumbra varied from 0.057 mm to 0.032 mm with short to long cones using a 1.0-mm focal spot. With a 0.6-mm focal spot, the smallest penumbra of 0.19 mm was twice the measurement reliability. This method demonstrated x-ray beam angulation and validity for comparing serial films can be estimated for the extreme variations but not accurately for +9 to −6 degrees from a tangent to the fixture. Small focal spot, long cone techniques are needed to match the reliability of the method. Achieving valid bone loss measurements < 0.2 mm is extremely difficult. (Implant Dent 1992;1:79–83)