Non-great-circle propagation was frequently noted in the year's HF forward oblique soundings recorded at College from various sites during 1963 and 1964. No direction-finding equipment was available, so a statistical analysis was made to determine the type of sidescatter involved. The number of occurrences during several periods was a maximum at night. The excess propagation time on the Palo Alto to College path varied inversely with magnetic activity. A comparison of simultaneous College backscatter and Palo Alto to College off-path data showed that the locus of off-path sidescatter extended north of the ionospheric backscattering belts. Previous work has shown the backscatter and auroral belts to be closely related, so these results are interpreted as showing that the deviated modes were produced by sidescatter from the auroral belt.