Originally posted by Ollie
The vector of the air velocity can be resolved into a component at right angles to the 25% chord line and a spanwise vector. Neglecting frictional drag, the lift and drag are produced by the velocity perpendicular to the 25% chord line and the airfoil should be oriented accordingly. The spanwise vector does nothing but produce some small frictional drag. Since the effective velocity vectors are not parallel to each other, where the 25% chord lines meet at the center of the wing, there may be a middle effect which decreases lift and increases drag in that region. The Horton brothers designs added a bat tail at the wing center in an attempt to curve the 25% chord line from one panel to the other across the centerline in an attempt to reduce the middle effect . The Swiss SB-13 curved the leading and trailing edges in planform to attempt the same reduction in middle effect.
AS Ollie suggests, the square of the velocity component perpendicular to the quarter cord should be used to calculate the lift, but that is only part of the impact of sweeping the wing. There is a slightly offsetting effect of an increase in the effective camber. For instance, with a 30% sweep measured from the quarter cord, the velocity component perpendicular to the quarter cord will be 86.6% of the free stream velocity (cosine(30)). But the cord perpendicular to the quarter cord will also be reduced to 86.6% of the cord parallel to the free stream. As a result, the wing appears to be 1/.866 thicker (e.g., a 10% foil will behave like a 15.5% foil, with proportional increase in camber an angle of attack.) The ultimate impact is that the lift drops off as the cosine of the sweep angle, rather than as the much higher cosine squared rate implied by lift being proportional to velocity squared.