Rat tail tendon was stained with a cationic phthalocyanin dye, Cupromeronic Blue, in a 'critical-electrolyte-concentration' method [Scott (1980) Biochem. J. 187, 887-891] specifically to demonstrate proteoglycan by electron microscopy. Hyaluronidase digestion in the presence of proteinase inhibitors corroborated the results. Collagen was stained with uranyl acetate and/or phosphotungstic acid to demonstrate the banding pattern a-e in the D period. Proteoglycan was distributed about the collagen fibrils in an orthogonal array, the transverse elements of which were located almost exclusively at the d band, in the gap zone. The proteoglycan may inhibit (1) fibril radial growth by accretion of collagen molecules or fibril fusion, through interference with cross-linking, and (2) calcification by occupying the holes in the gap region later to be filled with hydroxyapatite.

1. Developing tail tendons from rats (19-day foetal to 126 days post partum) were examined by electron microscopy after staining for proteoglycan with a cationic copper phthalocyanin dye. Cuprolinic Blue, in a "critical electrolyte concentration" method. Hydroxyproline was measured on papain digests of tendons, from which glycosaminoglycuronans were isolated, characterized and quantified. 2. Mean collagen fibril diameters increased more than 10-fold with age according to a sigmoid curve, the rapid growth phase 2 being during 30-90 days after conception. Fibril periodicities were considerably smaller (50-55 nm) in phases 1 and 2 than in phase 3 (greater than 62 nm). 3. Dermatan sulphate is the main glycosaminoglycuronan in mature tendon. Chondroitin sulphate and hyaluronate preponderate in foetal tissue. 4. Proteoglycan was seen around but not inside collagen fibrils. Proteoglycan and collagen were quantified from electron micrographs. Their ratios behaved similarly to uronic acid/hydroxyproline and hyaluronate/hydroxyproline ratios, which decreased rapidly around birth, and then levelled off to a low plateau coincident with the onset of rapid growth in collagen fibril diameter. 5. Dermatan sulphate/hydroxyproline ratios suggest that the proteoglycan orthogonal array around the fibril is largely dermatan sulphate. In the foetus hyaluronate and chondroitin sulphate exceed that expected to be bound to collagen. 6. An inhibiting action of chondroitin sulphate-rich proteoglycan on fibril diameter growth is suggested. 7. The distributions of hyaluronate, chondroitin sulphate and dermatan sulphate are discussed in the light of secondary structures suggested to be present in hyaluronate and chondroitin sulphate, but not in dermatan sulphate.

Two collagen-poor, ultramicroscopic layers are described at the surface of canine articular cartilage. They are distinguished by staining with an electron-dense cationic dye, Cupromeronic Blue, in a critical electrolyte concentration technique and by digestion with testicular hyaluronidase. The superficial layer, approximately 50 nm thick, stained at low electrolyte concentrations but failed to stain in conditions specific for sulphated glycosaminoglycans. It was hyaluronidase-resistant and may be either glycoprotein or protein in nature. The deeper layer, 100-400 nm thick, stained positively at electrolyte concentrations specific for sulphated glycosaminoglycans but not in conditions specific for keratan sulphate. It was removed by hyaluronidase digestion. This layer probably represents a chondroitin sulphate-rich proteoglycan. These surface layers may be important in the lubrication of the articular surface and in the permeability and compression resistance of the superficial cartilage zone.

SUMMARY Thin (100 urm) perpendicular slices of canine femoral condylar cartilage were placed horizontally on the stage of a Nachet microscope and viewed by transmitted light in the differential interference contrast mode. Each slice was held on the microscope stage by a loading rig and tested mechanically in compression. Measured loads to a maximum of -2-3 MN/M2 were applied to the end of the slice corresponding to the articular surface. Photographs were taken of the cartilage before and during loading, and the distance by which selected chondrocytes were displaced was used as an index of mechanical strain, i.e., of change in length/original length. Maximum strains were observed in the superficial cartilage zone. Minimum strains were recorded in the mid-zone, at a depth corresponding to -75% of the total cartilage thickness. The relative concentrations of cartilage collagen (COL) and proteoglycan (PG) were assessed by the light and electron microscopic histochemical study of cartilage sections taken from contiguous blocks. Superficial cartilage, which deformed most, had high concentrations of orientated COL fibres, low concentrations of PG. Mid-zone cartilage, which deformed least, had lower concentrations of randomly arrayed COL fibres but relatively high concentrations of PG.Hyaline articular cartilage appears uniform and homogeneous and has unique mechanical properties that include high compression resistance. This property is determined by the retention of water within the domains of PG aggregates arranged within a gel reinforced with COL fibres. The physical properties of cartilage have been examined by many classical mechanical procedures.' Experiments have been made on whole joints,2 3 parts of disarticulated joints,4 and pieces of cartilage.Tests have been made in tension and in compression, but most have not distinguished between the properties of different cartilage zones.The apparent homogeneity of articular cartilage conceals remarkable molecular heterogeneity. It is useful to consider cartilage as a series of different zones, each with a distinct microscopic structure and varied composition. In the most simple analysis, articular cartilage can be said to comprise superficial, mid-, and deep zones.9 In a more sophisticated

Proteoglycans of canine articular cartilage were labelled for transmission electron microscopy using the cationic copper phthalocyanin dye, cupromeronic blue, in a critical electrolyte concentration method. Much of the proteoglycan appeared to be structurally unrelated to collagen but a small proportion was positioned close to fibrils. On demonstrating the characteristic collagen banding pattern with uranyl acetate and phosphotungstic acid, it was evident that proteoglycan interacted with collagen at the d band.