Appropriate use of the "anode heel effect" of the output beam from an X-ray tube can reduce the effective dose to patients in some common radiological examinations. We investigated the variation in radiation intensity across the X-ray beam caused by the anode heel effect, and quantified the difference in absorbed dose to critical organs resulting from lumbar spine X-ray projections carried out with the two possible orientations of the patient along the tube axis (cathode to anode). A Rando phantom and some high sensitivity thermoluminescent dosemeters (TLDs) (LiF:Mg,Cu,P) were used. With the tube axis horizontal, radiation intensity profiles, parallel and perpendicular to the axis, were measured. Lumbar spine radiographs were recorded using the Rando phantom in the standard anteroposterior (AP) and lateral projections. TLD pellets were used to measure the absorbed radiation dose at various sites corresponding to critical organ tissues (ovaries, testes, breasts, thyroid and lens). Each set of projections was recorded in two phantom orientations, first with the phantom head placed towards the cathode end of the X-ray tube, and then in the reverse direction. From the radiation intensity profile of the incident X-ray beam, the "cathode end" to "anode end" air dose ratio was found to be 1.8. In lumbar spine radiography, with the phantom head placed towards the anode end of the X-ray tube, the ovaries and testes received an average dose 17% and 12% higher, respectively, in the lateral projection, and 16% and 27% higher, respectively, in the AP projection, than those obtained in the reverse "patient" orientation. These results indicate that patients (particularly females) should always be positioned with the head placed towards the cathode end of the X-ray tube for lumbar spine radiography to achieve significant dose reductions.