COMMON THINGS THE ALMANACK. 



perceptible extent from the equinoctial year, this coincidence is 

 not continual and results from the compensation produced by the 

 device of bissextile or leap years, for a full explanation of the 

 origin and purpose of which, see our Tract on " Time." By this 

 expedient the civil years form a quadrennial cycle, consisting of 

 three years of 365 days, and one year of 366 days. Now it will 

 be easy to see that the consequence of this is, that the commence- 

 ment of spring oscillates forward and backward alternately, being 

 for the first three years of the cycle continually 5 h 48 ra later, and 

 in the last year 18 h 12 m earlier. On this account the first day 

 of spring sometimes falls upon the 20th and sometimes on the 

 21st March. 



As the length of the seasons respectively remains the same, it 

 follows that the commencement of them severally is subject to a 

 like variation. The commencement of summer oscillates between 

 the 21st and 22nd June, that of autumn between the 22nd and 

 23rd September, and that of winter between the 21st and 22nd 

 December. 



To make this more evident let us take for example the qua- 

 drennial cycle, which commenced with the year 1853. The 

 commencement of spring in the year 1853 was at 4 h 40 m in the 

 afternoon of the 20th March. The year 1853, having only 

 365 days, while the interval between the two successive equinoxes 

 is 365 d 5 h 48 m , it follows that the commencement of spring in 1854 

 was 5 h 48 m later, and consequently took place on the 20th March, 

 at 10 U 28 m in the evening. In the same manner, 1855 having 

 only 365 days, the next commencement of spring is again 5 h 48 m 

 later, and consequently takes place at 16 minutes past 4 o'clock in 

 the morning of the 21st March. The following year, 1856, is, 

 however, leap year, and has 366 days, while the interval between 

 the equinoxes being only 365 d 5 h 48 m , is 18 h 12 ra less, and con- 

 sequently the commencement of spring will be 18 h 12 m earlier in 

 1856 than it was in 1855, and, therefore, will take place at 

 4 minutes past 10 o'clock in the morning of the 20th March. 



Thus the commencement of spring alternately advances and 

 retrogrades ; but the Julian cycle of four years, modified by the 

 Gregorian cycle of 400 years, produces such a compensation, that 

 for many thousands of years it cannot be earlier than the 20th or 

 late* than the 21st March, and the variation of the commencement 

 of the other seasons is subject to similar limits.* 



11. It will be evident, therefore, that although the year, in 

 respect to its length, has a relation to the course of the seasons, it 

 has no such relation in respect to its beginning and end. It 



* See Tract on " Time." 

 8 



