which lens has the smallest percent error between the bear-object method and the distanct-object method? which has the largest? does this make sense given how the distant object method is performed?