Page:Popular Science Monthly Volume 18.djvu/782

762 the error, after many other devices, by intercalating three months—every eight years, making every eighth year consist of fifteen months a method used by them for many years, perhaps centuries. It is said, however, that the length of the year was known, as early as the time of Solon, to be three hundred and sixty-five and one quarter days. Prior to the time of Numa, the Roman year consisted of ten months. He divided it into twelve lunar months, and to correct the error of eleven days a month was intercalated every second year. The management of this matter was intrusted to the priests, who added days whenever they deemed them necessary. But, owing to their ignorance of astronomy, this method proved irregular and erroneous, and the winter months were gradually carried back into autumn, the autumn months into summer, etc.

These errors had become so troublesome in the time of Julius Cæsar that he undertook, with the aid of an eminent astronomer (Sosigenes), to correct the calendar, which was done as follows: The Roman civil year had lost about ninety days. These were added to the year of three hundred and fifty-five days, making that year consist of four hundred and forty-five days, which is known in history as the "year of confusion." The year henceforward was made to consist of three hundred and sixty-five days, by adding ten days distributively to as many months. The odd quarter of a day was not noted until every fourth year, when the sum of these fourths made one day, and that year consisted of three hundred and sixty-six days. This odd day was inserted after the sixth day before the kalends of March, i. e., after the 24th of February, and was not counted as an addition to the year, but as a sort of appendix. Hence the sixth of the kalends of March was called bissextus, or double sixth, which root is still retained in our word bissextile, though the day is now added at the end of February. This arrangement would have been entirely correct had the year consisted of exactly three hundred and sixty-five and one quarter days. It was, however, in course of time, discovered to be erroneous, and another correction was made, which we will consider by and by. This new system went into effect on January 1, 46

The subdivisions of the Roman month were apparently arbitrary. The days were not numbered as we have them, but the first day of every month was called the calends, the fifth the nones, and the thirteenth the ides, except in the months of March, May, July, and October, when the nones fell on the seventh and the ides on the fifteenth. From these points the days were counted backward—i. e., the last of February was called Prid., Kal., Mar., etc.

—The division of days into weeks was invented at a very early day by the Chaldeans, and was afterward adopted by almost all civilized nations, and by the Romans about the third century The principle upon which the days were named is odd enough to deserve especial notice. The order of the planets, according to the