Mathematical models are often used to quantify the overall forces and moments acting on the lumbar spine. However, if the purpose of the research is to explain how spinal tissues can be injured, it is necessary to distribute the overall forces and moments between land within) different spinal structures, because it is the concentration of force which causes injury, and elicits pain. This paper reviews recent experimental evidence concerning the distribution of forces and moments acting on the lumbar spine. Lordotic postures increase loading of the posterior annulus and apophyseal joints, whereas moderately flexed postures tend to equalise compressive stress across the disc, and unload the apophyseal joints. Sustained compression reduces the volume and pressure of the nucleus pulposus, while increasing compressive stresses in the annulus and neural arch. Sustained compression also reduces disc height, giving some slack to collagen fibres in the intervertebral disc and ligaments, and causing them to resist bending less. Disc degeneration has a similar effect on disc height, and stress distributions. On the other hand, discs and ligaments can be subjected to greater bending moments following a period of sustained or repetitive bending, because sustained bending impairs the normal protective reflex from the back muscles, and repetitive bending fatigues the back muscles, reducing their ability to protect the spine. Incorporating this information into mathematical models will make them better able to identify which activities are most likely to injure the lumbar spine in life. (C) 2001 Elsevier Science Ltd. All rights reserved.