Abstracts Suppl p 97

loss of crystals occurring either during tissue preparation or in vivo. Crystals were situated either randomly in the matrix, many of them within a few micrometers of the articular surface, or around the cells. Although the number of crystals varied, this did not seem to be related to the time interval between the fracture and the sampling of the tissue. X-ray emission microanalyses of crystals (n = 10) in formalin fixed sections showed calcium and phosphorus. The ratio of calcium to phosphorus, uncorrected for absorption, fluorescence, and atomic weight was 1-95 (SD 0 30), suggestive of apatite rather than pyrophosphate deposits. Ultrastructurally articular surfaces were often roughened, though smooth to the naked eye, and showed infiltration with electron dense material. They were only rarely fissured. In many specimens the superficial collagen fibrils retained the normal density of packing and orientation tangential to the surface. More deeply, the fibres could show some disorganisation. Flattened chrondrocytes were often present in the superficial zone. Chrondrocytes in deeper tissue contained irregular lobed nuclei and several lipid droplets in addition to active Golgi and endoplasmic reticulum. Crystal deposition in the hip joints of these women may be related to osteoporosis, and this requires further investigation. Localisation of the crystals within the cartilage (superficial zone and pericellular regions) corresponds to sites of extracellular lipid in aging articular cartilage generally. Perhaps the most noteworthy features are the occurrence of crystals in non-arthritic hip joints, and their occurrence in all the elderly women in this group