The length of the sidereal revolution varies to the extent of three hours, and that of the synodical revolution to the extent of thirteen hours. But, whichever measure of the month we take, it is incommensurate with the year, _i.e._ there is not an even number of months in a year. In ancient times ceaseless efforts were made to adjust the months to the measure of the year, but we have practically given up the attempt, and in our calendar the lunar months shift along as they will, while the ordinary months are periods of a certain number of days, having no relation to the movements of the moon. It has been thought that the period called a week, which has been used from time immemorial, may have originated from the fact that the interval from new moon to the first quarter and from first quarter to full moon, etc., is very nearly seven days. But the week is as incorrigible as all its sisters in the discordant family of time, and there is no more difficult problem for human ingenuity than that of inventing a system of reckoning, in which the days, the weeks, the months, and the years shall be adjusted to the closest possible harmony. [Illustration: =Spiral Nebula in Ursa Major (M 101)= Photographed at the Lick Observatory by J. E. Keeler, with the Crossley reflector. Exposure four hours. Note the appearance of swift revolution, as if the nebula were throwing itself to pieces like a spinning pin-wheel. ] [Illustration: =The Whirlpool Nebula in Canes Venatici= Photographed at the Lick Observatory by J. E. Keeler, with the Crossley reflector. Exposure four hours. Note the “beading” of the arms of the whirling nebula. ] ------------------------------------------------------------------------ PART III. THE SOLAR SYSTEM. ------------------------------------------------------------------------ PART III. THE SOLAR SYSTEM.