Major solar flares during 1967--1970 are signficantly more likely to occur in active regions whose inferred overlying large-scale (~100,000 km) magnetic flux is oriented preferentially north-to-south than for south-to-north or indeterminant orientations. This purely solar effect may be the dominant cause of previously reported correlations between southward solar active region magnetic fields and enhancements in geomagnetic disturbances, solar wind velocities and (newly reported here) solar flare proton fluxes. If this effect may be generalized to other portions of the solar cycle, it could offer a fundamental clue to the flare mechanism which should also prove useful in the prediction of solar-terrestrial disturbances.