During the development of normal bone in the young child, osteoblasts and then haematopoietic stem cells progressively colonise the cartilaginous matrix, resulting in its ossification to bone matrix. 1 It is at this time that red bone marrow appears in the skeleton, starting distally in the foot and hand at around one year of age. In normal children, the marrow in all bone cavities and the spongiosa at the epiphysis is red, because the amount of medullary space in young children is limited due to an abundance of cartilage and the thickness of spongy bone, while the haematopoietic requirements are high due to the growth-related expansion of the blood volume. At the end of the growth period, in late adolescence, the need for haematopoietically active marrow stabilises and adipose involution occurs in many areas of the medulla and spongiosa. Haematopoietic cells are replaced by fat, giving rise to yellow, inactive marrow, but red marrow still persists in certain areas such as the iliac crests. The adult skeleton possesses two types of bone marrow, one red and haematopoietically active, the other yellow due to adipose involution. The haematopoietic stem cell in bone marrow has been studied extensively because of its clinical relevance in therapeutic transplantation of bone marrow. Haematopoietic stem cells are pluripotent and capable of producing progeny that can differentiate into any and all of the cells of circulating blood and the immune system through a well-defined series of steps leading to differentiation into mature blood cells. In addition to the haematopoietic element, red marrow also contains a stroma where the osteogenic precursor cells are found. The osteogenic capacity of bone marrow was first demonstrated in rabbits by Goujon2 in 1869. This capacity has been exploited by several authors to reinforce the osteogenic properties of allografts, 3 xenografts4 and composite grafts5 by mixing bone marrow removed during the operation with the bone graft.

Burwell3 showed that primitive osteogenic cells in bone marrow are responsible for much of the biological efficacy of cancellous bone grafts. Friedenstein et al6 showed that new bone was formed by proliferative fibroblastlike marrow cells which persisted in vitro after haematopoietic cells had died and that the number of cells able to proliferate rapidly could be assayed by counting the number of fibroblastic colony-forming units (CFU-Fs) in bone marrow. Several studies7-13 have shown that at least some of these cells are pluripotent and can differentiate into osteoblasts, chondrocytes, adipocytes, myoblasts and haematopoiesis-supporting cells. Many names have been used to describe the fibroblastic colonyforming cells that can be grown from bone marrow, periosteum, bone fragments, or callus digested from trabecular bone. In this review the term ‘tissue progenitors’ is used to describe the connective proliferative cell population that can be harvested from bone marrow and which is capable of differentiating into one or more connective tissue phenotypes. Because bone marrow is known to contain osteogenic progenitors, its implantation was perceived to have the potential to lead to effective bone regeneration. Various preclinical investigations, 14, 15 and clinical studies, 15-19 have confirmed this. In clinical practice, autologous marrow is harvested from the iliac crest and immediately transplanted to the site in need of skeletal repair. This type of marrow transfer, or grafting, is a relatively simple procedure which is inexpensive and can be done on an outpatient basis. This technique of transplanting autogenous connective tissue progenitors is the first of the four major cell-based strategies of tissue …