1. Direct injury to skeletal muscle results in fragmentation and necrosis of muscle fibres, though this is patchy in distribution.

2. The sarcolemmal basement membranes form the interface along which fibre regeneration takes place.

3. Phagocytosis of disorganised sarcoplasm is an essential prelude to the reconstitution of severely damaged fibres.

4. Regeneration of injured muscle begins with proliferation of basophilic cells probably originating from muscle satellite cells. After a few days typical myoblast nuclear chains are present. By a week following injury the chains of myoblasts have formed myotubes, which possess myofibrils and sarcomeres.

5. By twelve days in the monkey and by eighteen days in man the muscle fibre regenerative process shows many new fibres which have not reached a mature diameter.

6. Much collagen may be formed in the tissue space at the site of injury. It appears that as the muscle fibres increase in diameter the collagen decreases in extent.

7. In the monkey by three weeks the muscle at the fracture site appears normal. This is also true in the specimens examined at four, six and twelve weeks.

8. In the monkeys the injured limb was immediately used to run and jump. A parallel intense and early activity of muscle and joints was a cardinal point in the management of this series of fracture patients. The clinical results were satisfactory.

9. It is concluded that in both the monkey and in man, given active limb movements, permanent and functionally useful muscle regeneration occurs following soft-tissue injury associated with a bone fracture.

The experiments were performed to answer three main questions. These and our answers may be summarised as follows.

What is the precise mechanism of healing of a raw bony surface in a joint? What cells are involved? Where do they originate?—In all the implant experiments and in the control series the fundamental mechanism of healing was similar.

1. A massive proliferation of fibroblasts occurred from the cut periosteum, from the cut joint capsule, and to a lesser extent from the medullary canal.

2. Fibroblasts grew centripetally in the first few weeks after operation, attempting to form a "fibroblast cap" to the cut bone end.

3. Fibroblasts of this cap near the cut bone spicules metamorphosed to become prechondroblasts, chondroblasts laying down cartilage matrix, and hypertrophied (alkaline phosphatase-secreting) chondrocytes lying in a calcified matrix.

4. This calcified cartilage matrix was invaded by dilated capillaries probably bearing osteoblasts which laid down perivascular (endochondral) bone.

5. Some of the cells of projecting bone spicules died and their matrix was eroded in the presence of many osteoclasts.

6. In the control experiments of simple excision of the radial head new bone was produced at the periphery only by processes (3) and (4). This sealed off the underlying peripheral cortical bone from the superficially placed peripheral articular surface of fibrocartilage. At about a year from operation the central portion of the articular surface was still formed of bare bone, or of bone spicules covered by a thin layer of irregularly arranged collagen fibres. The opposite capitular articular cartilage was badly eroded.

Does the introduction of a dead cartilage implant over the raw bone end affect in any way the final constitution of the new articular surface?—In the implant experiments the new bone produced by processes (3) and (4) formed, after about a year, a complete cortical plate which entirely sealed off the cut end of the radius and left a superficially placed articular covering of smooth fibrocartilage, closely resembling a normal joint surface. The opposite capitular articular surface was normal.

What is the final fate of such an implant?—Whale cartilage implants underwent replacement by fibroblasts and collagen fibres, and took about nine months to disappear.

The cartilage of fixed autotransplants and homotransplants underwent similar gradual replacement, and took about the same time in each case. The dead bone, implanted in association with the cartilage in both cases, acted as a nidus for hyaline cartilage production by chondrocytes derived from fibroblasts. This cartilage underwent endochondral ossification. This observation suggests that induction by non-cellular osseous material is a factor in chondrification and ossification.

All the implants functioned as temporary articular menisci or in some cases as temporary radial articular surfaces. They were always replaced by a permanent fibrocartilaginous meniscus, or a fibrocartilaginous articular surface. An implant did, in fact, always act as a temporary protecting cap and mould for the subjacent growth offibroblasts which was necessary for the production of a satisfactory new joint surface.

1. Previous studies of the movements of the lumbar spine are criticised in the light of new observations from radiograph tracings. It is shown that, contrary to recent teaching, the lumbar spine is a very mobile part of the vertebral column.

2. The movement of the lumbar spine is analysed. It is shown that the lower vertebrae have the most movement, and that the range gradually becomes less in the upper lumbar spine.

3. This movement may be roughly correlated with the incidence of spurs arising from the anterior margin of the vertebral bodies.

4. These spurs are shown to arise in the anterior longitudinal ligament; they are probably caused by intermittent pressure from the intervertebral disc lying behind the ligament.