Different spinal ranges of motion (ROM) were measured and the results of 17 repeated tests correlated with spinal radiological changes in 52 male patients with ankylosing spondylitis (AS). Both Schober tests and measurements of lumbar and cervical rotations (TRi, TR, CR, CRt) and lateral flexions (LFLf, LFLx, CLFLt, CLFLm), together with thoracolumbar flexion (ThFL), cervical flexion-extension measurements (CFL, CExt), and tragus - wall and occiput - wall distances (OWD,TWD), showed significant correlations with detailed radiological spinal changes. Cervical rotation (CRm, CRt) and flexion (CFLm) correlated only with cervical changes, and thoracolumbar rotation as assessed by instrument (TRi) correlated only weakly with lumbar changes, while chin-chest distance (CCD) and chest expansion (CE) showed no correlation. Inter- and intratester reliability was good in all tests (the intraclass correlation coefficients ranged from 0.84 to 0.98). Three new tape methods for measuring thoracolumbar and cervical rotations and cervical lateral flexion also proved to be valid and reliable, as did the Schober-S1 modification. We conclude that the thoracolumbar segment (Schober), whole (ThFL) and lateral (LFL) flexions and rotation (TR), and chest expansion (CE) (after careful standardisation) together with cervical rotation (CR), extension (CExt) and/or lateral flexion (CLFL) comprise the set of mobility tests for the follow-up and assessment of disease progression in AS. On the other hand, cervical (forward) flexion (CFL), chin-chest distance (CCD) and an instrument method for thoracolumbar rotation (TRi) are not approaches to be recommended.