The article reports a new hypothesis of whiplash injury based on a series of experimental studies using isolated human cadaveric specimens. Although the clinical symptoms of whiplash are widely known, the understanding of the underlying injury mechanism is poor. The prevailing view of neck-hyperextension as the essential injury mechanisms was not supported by recent experiments. In a series of experiments using eight human cadaveric specimens which underwent experimental stepwise whiplash acceleration from 2.5 to 10.5 g functional radiographs and flexibility tests were performed at the end of each acceleration step. Ligament strains, vertebral alignment and elongation of the vertebral artery were monitored during the whiplash trauma by highspeed cinematography and specially designed transducers. After the trauma CT- and MRI-scans were taken and specimens were sectioned using Cryomicrotomy. We found a distinct biphasic kinematic response of the cervical spine to whiplash trauma. In the first phase the spine formed an S-shaped curve with flexion at the upper levels and hyper-extension at the lower levels. This phase was found to be the vulnerable phase of whiplash trauma. The largest dynamic elongation of the capsular ligaments was observed at the C6-C7 level during this initial S-shaped phase of whiplash. The maximum elongation of the vertebral artery could be observed synchronously in the first S-shaped curve of the cervical spine. In the second phase of whiplash all levels of the cervical spine were extented, so that the head reached is maximum extension. No injuries were observed in the second phase. We propose, based on our experimental findings, that with low accelerations the anterior structures of the lower cervical spine are injuried during the first phase of whiplash, when the cervical spine forms an S-shaped curve and before the neck is fully extended. At higher trauma accelerations there is also a tendency for the injuries to occur at upper levels of the cervical spine. Based on our findings the traditional view of whiplash as hyper-extension injury can be modified by a differentiated, time dependent, biphasic biomechanical model of the injury, thus allowing better and more effective injury prevention, diagnosis and therapy.