1. The effect of removal of mechanical loads from bone. Lanyon and various co-workers studied functionally isolated avian bone preparations to which external loads could be applied in vivo through external fixation devices. They showed that the application of a rigid external fixator unloaded the bone, and that this stress shielding resulted in a substantial remodelling of the bone on three fronts: endosteal, cortical and, to a lesser extent, periosteal. The balance of remodelling was negative, resulting in a net loss of bone mass. Similar results with rigid external fixation have been reported in other animals. These findings are consistent with what we know about disuse osteoporosis resulting from muscular inactivity and reduction in weight bearing. Clinically such bone atrophy commonly occurs: after a fracture necessitating various degrees of immobilisation; after muscle inactivity due to diseases of joints and muscle, or bed rest; after long-standing systemic debilitating disease; after muscle paralysis; and after periods of weightlessness in space. The results are also consistent with what we know about bone that is unloaded by various fixation devices. Woo and his colleagues have shown that in intact bone, fixed with a stainless-steel plate, there is significant stress shielding and that this results in loss of bone mass. Similar results have been reported by other investigators. Likewise, in fractures fixed by rigid plate fixation there is similar stress shielding, which again results in loss of bone substance, together with persistence of woven bone at the fracture site. Bone remodelling is very sensitive to small changes in cyclic bone stresses and changes representing less than 1% of ultimate strength can cause measurable differences in bone atrophy after a period of months. Experimental studies have shown that greater bone remodelling and bone loss is observed when the rigidity of fracture fixation is increased. Progressive bone loss may occur after fixation of fractures with metal plates. This leads to an ubiquitous clinical dilemma: if the plate is removed too early, fracture may occur because of insufficient union, whereas if the plate is removed too late, re-fracture may occur because of structural weakening and loss of bone mass. In summary, removal of mechanical loads from bone, whether it be physiological, by rigid plate fixation or by rigid external fixation, results in negative remodelling and a net loss of bone mass. 2. Effect of cyclic mechanical loads on intact bone. Rubin and Lanyon, again using isolated avian bone preparations, found that the application of a cyclic load of only four consecutive cycles a day prevented negative bone remodelling and resulted in no change in bone mass. This suggested that a suitable strain regimen prevented remodelling. Furthermore, they found that 36 consecutive cycles per day not only prevented cortical resorption, but also resulted in substantial periosteal and endosteal new bone formation over a six week period. An increase in the number of strain cycles to 360, or 1800 provided no increased benefit. That mechanical loading of intact bone results in cortical thickening and increased bone deposition has been confirmed by other studies. Physiological loading of intact bone produces the same increased bone deposition in laboratory animals. Similar effects have been shown in humans, for example, in tennis players, baseball pitchers and cross country runners, as well as in other sportsmen. Resection of the radius or ulna, thereby increasing the load of weight bearing in the remaining bone, has been shown to result in hypertrophy of that bone in dogs and in various animals. Fixation of fractures with less-rigid fixation results in healing with external callus formation, and earlier weight bearing. In summary, these studies have shown that, in animals or humans, the application of physiological levels of strain to bone, either physiologically or mechanically, causes remodelling which results in a net gain of bone mass. 3. Effect of static mechanical loads on intact bone and fractures. Using the same avian model, Lanyon and Rubin showed that static loads of similar physiological magnitudes of strain did not have a positive influence on the remodelling process. Hart, Wu, Chao and Kelly obtained similar results using external fixators. They concluded that static compression increased the rigidity of fixation but, of itself, provided no direct benefit for bone healing. Anderson studied compression plate fixation and the effect of different types of internal fixation and reported no evidence of stimulation of osteogenesis by compression. Other researchers have reported similar findings. The effects of static compression produced at the fracture site by plate fixation have been reviewed extensively. Some investigators have claimed that compression promotes fracture healing, but there is no evidence of this from paired comparisons in the literature. In summary, static compression does not directly stimulate fracture healing. 4. Effect of cyclic mechanical loads on fractures. Yamagishi and Yoshimura showed in 1955 that intermittent compression forces applied to healing fractures in rabbits caused proliferation of cartilaginous callus. In 1981 Wolf and co-workers reported that when long bone fractures were treated with cyclic loading, bone strength increased more rapidly than when fractures were treated by constant compression. In 1985 Goodship and Kenwright published their work on the influence of induced cyclic micromotion on the healing of experimental tibial fractures, using an Oxford External Fixator. When 500 cycles were applied per day, they found that the micromotion produced external callus sooner, namely at one week, compared with static external fixation where callus was just commencing at three weeks. The micromotion resulted in more callus formation, which extended over a wider portion of the diaphysis. Consequently, they found that fracture stiffness increased at a greater rate in the stimulated group than in the rigid group. When the animals were sacrificed at twelve weeks they found that there was increased torsional stiffness in the stimulated group, ie. 83% of the intact control stiffness, compared with 54% in the rigidly-fixed group. These findings have been replicated by others. Yamagishi and Yoshimura, as well as Woo and co-workers, have shown that those models which allowed some fracture movement produced proliferative external callus formation. This callus was inhibited proportionally as the rigidity of the fixator was increased. Similar studies have been performed in humans. Kenwright, Goodship and co-workers showed that controlled axial cyclic micromotion decreased the time to full weight bearing, compared with rigid tibial fixation33, and further studies showed the same findings. In summary, both animal and human studies have shown that the application of controlled cyclic micromotion to fractures promotes bone healing. 5. Summary and application. An understanding of the manner by which various loading regimes affect bone formation and fracture healing allows the treating physician to plan effective treatment of fractures. It forms a rationale for total perioperative management of patients, in terms of the choice of treatment, the choice of implant, the weight-bearing status and the timing of physical activity. It has also lead to the concept of ‘dynamisation’ of fractures and the development of second and third generation external fixators