BACKGROUND: Many studies have compared the restriction of motion that immobilization collars provide to the injured victim. No previous investigation has assessed the amount of motion that is generated during the fitting and removal process. The purpose of this study was to compare the three-dimensional motion generated when one-piece and two-piece cervical collars are applied and removed from cadavers intact and with unstable cervical spine injuries. METHODS: Five fresh, lightly embalmed cadavers were tested three times each with either a one-piece or two-piece cervical collar in the supine position. Testing was performed in the intact state, following creation of a global ligamentous instability at C5-C6. The amount of angular motion resulting from the collar application and removal was measured using a Fastrak, three-dimensional, electromagnetic motion analysis device (Polhemus Inc., Colchester, VT). The measurements recorded in this investigation included maximum values for flexion/extension, axial rotation, medial/lateral flexion, anterior/posterior displacement, axial distraction, and medial/lateral displacement at the level of instability. RESULTS: There was statistically more motion observed with application or removal of either collar following the creation of a global instability. During application, there was a statistically significant difference in flexion/extension between the one-piece (1.8 degrees) and two-piece (2.6 degrees) collars, p = 0.009. There was also a statistically significant difference in anterior/posterior translation between the one-piece (3.6 mm) and two-piece (3.4 mm) collars, p = 0.015. The maximum angulation and displacement during the application of either collar was 3.4 degrees and 4.4 mm. Statistical analysis revealed no significant differences between the one-piece and two-piece collars during the removal process. The maximum angulation and displacement during removal of either collar type was 1.6 degrees and 2.9 mm. CONCLUSIONS: There were statistically significant differences in motion between the one-piece and two-piece collars during the application process, but it was only 1.2 degrees in flexion/extension and 0.2 mm in anterior/posterior translation. Overall, the greatest amount of angulation and displacement observed during collar application was 3.4 degrees and 4.4 mm. Although the exact amount of motion that could be deleterious to a cervical spine-injured patient is unknown, collars can be placed and removed with manual in-line stabilization without large displacements. Only trained practitioners should do so and with great care given that some motion in all planes does occur during the process.