For some 45 years, it has been known that marrow contains mesenchymal cells which are capable of modulating to the osteoblast class when they are grafted beterotopically [1, 2]. Friedenstein [3] defined those cells as determined osteogenic precursor cells (DOPCs), and his innovative ex vivo studies provided an experimental model that has greatly increased our knowledge about the cellular biology of osteoblast differentiation and bone formation [4-8]. Such information is particularly valuable in studies of the mechanisms of fracture healing [9, 10], iliac crest bone graft incorporation [11], and the success of composite grafts of osteoinductive [12] and osteoconductive biomaterials [13, 14] supplemented with marrow. These are emergency situations in which the mechanisms of healing involve autocrine and paracrine factors. The question is whether marrow-borne DOPCs can be said to play an important osteogenic role in situ in the absence of a requirement for repair. Based on the anatomy of marrow, it may be argued that marrow-borne stromal cells do not regularly participate as osteoprogenitor cells in situ. Scanning EM studies have shown that, in the rat femur and tibia, marrow is contained within a loose sac of electron-dense, epithelioid-like pericytic cells which are usually too thin (< 1.0 gin) to be clearly visualized at the light microscopic level--except in infrequent tangential cuts (Fig. 1)[15]. We are not certain whether marrow sacs as such are present in all species, but it seems likely that sac cells and underlying stromal elements are always interconnected and perhaps immobilized within a mesh of reticular fibers, and are separable from the osteoblast or bone lining cell populations (Figs. 1, 2). Unlike lymphocytic cells, there is no direct evidence that stromal cells migrate through the confines of the sac. Thus, there may be structural reasons for questioning the dictum from the 1960s that the marrow houses the" translocatable" fibroblast-like stromal cells (FLSCs) that radiothymidine autoradiographers identified as the endosteal osteoprogenitor cells [5]. The operating philosophy in experiments involving the parenteral administration of radiothymidine was that osteogenic stimuli would first increase the fraction of the osteoprogenitor FLSCs that could be isotopically pulse labeled, and this would be translated within 24 hours to an increase in thymidine-labeled osteoblasts. Even when treatments augmented the pool size of putative precursors, the flash-labeled osteoprogenitor FLSC numbers remained small and percentagewise, entirely equal to control values (estrogen in 30-day-old mice= 5%; parathyroid extract in 5-6-day-old rats= 15%)[16-21]. There is now evidence that osteogenicity is the property of perhaps 2.5% of the fibroblastic stromal cells that exhibit clonal growth in vitro [24]. Perhaps the take-home message is that radiothymidine pulse-labeling experiments mark too few of these cells to trace their movements adequately by autoradiography.