In the Old Style, every year of an exact century—every centurial (or, as it was sometimes called, secular) year—was a leap-year of 366 days. What would be the effect of treating every centurial year as a common year of 365 days? We should have suppressed four days at the end of four centuries when we ought to suppress only three in 399 years. So it was suggested that while three successive centurial years should be regarded as common years, the fourth centurial year should be treated as a leap-year. Thus, in both Old and New Style the years 1600 and 2000 are leap-years; but 1700, 1800, and 1900, which in the Old Style were leap-years, are in the New Style treated as common years of 365 days. And the rule laid down in the Gregorian system was that if the number expressed by the first two figures of the century was exactly divisible by 4 it should be a leap-year, but if not exactly divisible by 4 it should be treated as a common year. The numbers 16 and 20 are exactly divisible by 4, but 17, 18, and 19 are not so divisible. The years 1600 and 2000 are in the New Style leap-years, but the years 1700, 1800, and 1900 are in the New Style common years. It is true that the adoption of 133 years, instead of 130 years, as the time in which in the Old Style one day was gained by the Kalendar on the sun, imports an error into the system, which causes the Kalendar to fall behind the sun. This error, as has been said, will accumulate to the extent of one day in A.D. 5200. It may be thought that, if men be on the earth at that date, they will know how to deal with the case. Yet it is suggested for the instruction of our remote posterity that they will have only to make A.D. 5200 a common year, instead of a leap-year, to bring things back to correctness[166]. For the Sunday letters in the New Style and for the Cycle of Epacts in the Gregorian Kalendar, see Dr Seabury, _Theory and Use of the Church Calendar_. The work of the Gregorian reformation is marvellous in its elaborate ingenuity. It even provides for a case which will not occur till Dec. 31, A.D. 8600. Yet it does not reach the attainment of an exact correspondence with astronomical phenomena. And it has been frequently observed that the new moons of the Kalendar may occur one, two, or even three days _later_ than the new moons of the astronomer. In fact the astronomical new moon rarely occurs on the date marked for the ecclesiastical new moon. But care has been taken that the new moon of the Kalendar never occurs _earlier_ than the new moon of astronomy. _The adoption of the New Style._