Five hundred and seventy one matches in the Australian Football League [AFL] had ground hardness measured using a Penetrometer, over the period 1997-2000. The method used was 3 drops at each of 20 locations over the playing field on the morning before games. Anterior Cruciate Ligament [ACL] injuries were recorded using an ongoing injury surveillance system. There was a non-significant trend towards a higher risk of ACL injury when the 3-drop average of the Penetrometer was less (harder) than 4.5cm, RR 2.36 (95% CI 0.90-6.24). When the first drop average of the Penetrometer was less (harder) than 2.5cm, the relative risk was 2.60 (95% CI 0.94-7.20). There was also a nonsignificant trend towards an increased risk of ACL injury in games where the predominant grass type was couch (Bermuda) grass, as opposed to rye grass, RR 2.37 (95% CI 0.89-6.36). This study confirms previous findings from the AFL that early season matches and matches played at northern (warmer) venues have a higher risk of ACL injury. It is likely that ground-related variables are partially responsible for these observations, but to date, the relative contributions of ground hardness, grass type, shoe-surface traction and other confounding factors are not certain. There was a significant fall in the number of ACL injuries in the AFL (to approximately half the previous level of incidence) during seasons 1999 and 2000. It is possible that reaction to this study and related publicity has led to the preparation of ground conditions in the AFL that are less likely to produce ACL injuries.