562 osteoarthritic knees rated as stage 1or more according to Kellgren’s osteoarthritic knee classification were selected randomly and analyzed radiologically. Eighty cases with the height of 155 cm-160 cm, for which a large number of male and female cases are available (34 male cases, 46 female cases) were extracted for analysis. The values measured were significantly larger in male than in female in any region. In order to clarify differences in morphology between the sexes, the ratio between the values measured of various regions was computed. As a result, the value(AP/ML ratio) obtained by dividing the length of medial femoral condyle in anterior-posterior direction and the depth of medial femoral condyle in proximal-distal direction by the width of femur at articular level was 0.87±0.03, 0.56±0.03 in female against 0.81±0.04, 0.52±0.03 in male, respectively. A statistically negative correlation was found between femoral width and AP/ML ratio. The value (AP/ML ratio) obtained by dividing the length of medial tibia condyle in anterior-posterior direction by the width of tibia at articular level was 0.61±0.05 in female against 0.59±0.04 in male. A statistically negative correlation was found between femoral width and AP/ML ratio. That is, the larger the medial-lateral width of the tibia becomes, the smaller becomes the AP/ML ratio. When the differences between the sexes were studied, the values measured of various regions were significantly larger in males than in females even in the group of the same height. Moreover, The AP/ML ratio of the current components does not follow the negative correlation between the width and AP/ML ratio. It was concluded that the size variation of the currently available components should be reconsidered.