A Note on Saari’s Treatment of Rotation Curve Analysis