On archeomagnetic secular variation curves and archeomagnetic dating