it occurred on different days of June, and so on through May, April, and the preceding months, so that in thirty-two and a half of our years their New Year's Day has run through all our months and comes back again to July. So much for New Year's Days; they are arbitrary selections, and though the Roman New Year's Day, or January 1st, has been precisely defined and fixed by the determination by astronomers of the position of the earth on that day in its revolution around the sun, yet the original selection of January 1st for the beginning of the year seems to have been merely the result of previous errors and negligence in attempting to fix the winter solstice (which now comes out as December 22nd). This is the day when the sun is lowest and the day shortest; after it has passed the sun appears gradually to acquire a new power, and increases the duration of his stay above the horizon until the longest day is reached--the summer solstice (June 21st). Julius Caesar took January 1st for New Year's Day as being the first day of a month nearest to the winter solstice. The ancient Greeks regarded the beginning of September as "New Year." Were mankind content with the measure of time by the completion of a cycle of revolution of the earth around the sun--that is the year--and by the revolution of the earth on its own axis--that is the day or day-night ([Greek: nychthemeron]) of the Greeks--the notation of time and of seasons would be comparatively simple. No one seems to know why or when the day was first divided into twenty-four hours, nor why sixty minutes were taken in the hour and sixty seconds in the minute. The ancient astronomers of Egypt and China, and their beliefs in mystical numbers, have to do with the first choosing of these intervals in unrecorded ages of antiquity (as much as 2000 or 3000 B.C.). The seven days of the week correspond to the five planets known to the ancients, with the addition of the sun and the moon. But the Greeks made three weeks of ten days each in a month. The true year--the exact period of a complete revolution of the earth around the sun--is 365 days 5 hours 18 minutes and 46 seconds. It was measured with a fair amount of accuracy by very ancient races of men, who fixed the position of the rising sun at the longest day by erecting big stones, one close at hand and one at a distance, so as to give a line pointing exactly to the rising spot of the sun on the horizon, as at Stonehenge. They recorded the number of days which elapsed before the longest day again appeared, and they marked also the division of that period by the two events of equally long sunlight and darkness--the spring and the autumn "equinox." It is obvious that if they took 365 days roughly as the period of revolution they would (owing to the odd hours and minutes left out) get about a day wrong in four years, and it was the business of the priests--even in ancient Rome the pontiffs were charged with this duty--to make the correction