Background: Traumatic chondral lesions of the knee are common in football and rugby players, The diagnosis is often confirmed by arthroscopy, considered appropriate because of persistent pain and effusion. The natural course of these injuries is not well known. Clinical diagnosis is difficult and MRI is not always reliable. Aims: 1. To introduce a simplified arthroscopic mapping system of the weight-bearing surfaces of the knee which can be used to describe the location of these chondral lesions. 2. To correlate the location and severity of these lesions with a novel knee function score designe~ to reflect the demands of football and rugby. 3. To assess the accuracy of different MRI sequences in diagnosing chondral lesions using the arthroscopic mapping system as a standard. 4. To evaluate the short-term functional outcome of microfractured lesions using MRI and function scores. Methods: Forty two consecutive football and rugby players with traumatic isolated chondral lesions observed at arthroscopy were included after appropriate consent. Lesion size and grade were recorded with the mapping system. All subjects were scanned two to three weeks after surgery using a 3-Tesla MRI. At eight to 12 weeks from surgery they were tested with the functional knee score. Twenty four out of 42 subjects with grade III IV lesions underwent microfracture at the time of arthroscopy. They were assessed at 3, 6, 12 and 18 months by functional knee score and MRI. A second look arthroscopy was carried out in 10 players five to seven months after surgery to evaluate lesion healing because there was discrepancy between. a 'normal' MRI and persistent clinical symptoms. Results: Fifty five lesions on weight-bearing surfaces were found in the 42 subjects. The average size of the lesion was 197 square mm. Pain, effusion, tenderness on palpation and positive compression rotation test were the predominant symptoms and signs. The medial femoral condyle (MFC) was affected most with 36 (65 %) of the lesions. the lesions were concentrated in the B areas (p < 0.05). Grade IV lesions were the most common with 26 lesions (47.3 %). These lesions were concentrated in the B areas (p < 0.05). Cartilage specific sequences (CSS) showed a sensitivity of 89 percent and specificity of 98 percent to identify the chondral lesions. Lesion location and grade determined by MRI were comparable to arthroscopy, but size was underestimated by MRI (p < 0.05). Both the functional knee score and MRI showed good correlation in assessing healing after microfracture at six, 12 and 18 months (r2 =0.993,0.986 and 0.993, respectively). Conclusion: The distribution of the traumatic chondral lesions over the weight-bearing surfaces of the knee is unequal, and neither location nor grade predict functional outcome. Cartilage specific sequences have relatively high sensitivity but are not reliable enough to replace arthroscopy in diagnosing cases with typical symptoms and signs. Microfracture shows excellent short term out-comes. Both the functional knee score and MRI are reliable enough on average to confirm healing at the defect site, and a second look arthroscopy may be required in some cases.