General considerations about the use of convex mirrors indicate that they may be very useful to enlarge the field of view. There seems to be no need to use smaller values for the radius of the mirror than 1000 mm.A reaction time experiment is described in which plane mirrors of equal size at different positions on the fender are used. There are indications that the position of about 20° out of the line of sight straight ahead might be optimal. Whether this also holds for convex mirrors is not investigated.An experiment is described in which a driver must decide whether he can overtake a car in front of him, while a car behind him is approaching. This is done with several speeds of the car behind and with different curvatures of the side mirrors. Especially if the radius of curvature is not less than 1200 mm, no serious effects on driver behavior is observed.The instrumented car used in the studies is briefly described.The research described in this paper is carried out partly under contract with Donnelly Mirrors Inc., Holland, Michigan.