204 



KNOWLEDGE 



[September, 1903. 



Of course the irregularities in this cycle, e.g., those of 

 1889, 1892, 1925, 1928, are due to the fact that conditions 

 G, D are totally unsatisfied. For the same reason this 

 cycle ffives \is no clue to the longitude of an eclipse track, 

 or to the time of day at which the eclipse occurs. 



There is a point in which this 4^ cycle may sometimes 

 1)6 of use. The last of a series in the 4* is frequently the 

 first of a Sar. series ; when this is the case the 4^' serves as 

 an index, calling our attention to the birth of a new 

 eclipse in the Sar. It is not the case in 1953, for that 

 eclipse is bom two Sai-oses earlier, in 1917. 



The Saeos. — This cycle was discovered by the ancient 

 Chaldceans, at a very early date, and it still remains the 

 most perfect and serviceable of all eclipse cycles, on 



small northern partial eclipse, 1639, January 4, and in- 

 creased slowly each Sar. ; its companion eclipse (one mouth 

 earlier) lingering till 1837, April 5. Our eclipse became 

 central, 1891, June 6 (some readers may remember this 

 eclipse as the one visible from the Royal Observatory on 

 Visitation Day), but the tracks of 1891 and 1909 are too 

 far north to show conveniently. The diai4ram begins with 

 the eclipse of 1927, and shows all the tracks till 2179. 

 Even this date is not cpiite the middle of the series, but 

 that is only a few Saroses later. 



The southward motion of the tracks is affected by the 

 season of the year, which accelerates it from June to 

 December, and retards it for the other half-year. The 

 slope of the tracks also varies with the season. The 



Plate I.— Diagram illustrating the Saros Cycle. The tracks of a Total Eclipse are shown from 1927 to 2179. 



account of the wonderful accuracy with which conditions 

 B, C, D are satisfied. We can, in fact, get a very fair 

 idea of a coming eclipse without any calculation at all, by 

 tabulating the preceding ones of the series and carrying 

 on the differences. This appears to have been the way 

 that the Chaldseans predicted their eclipses. 



Mr. Maunder's paper, above referred to, has dealt so 

 fully with this cycle that it is unnecessary to add much 

 here. It mav, however, be interesting to give a diagram 

 showing the tracks of a series of eclipses, and I have 

 chosen the series to which the eclipses of 1927, 1999 (both 

 total in England) belong. This began its career as a 



western end of the ti-aek is the highest from June to 

 December, then the eastern end from December to June. 

 The tracks will continue to move southward, and the 

 eclipse will cease to be central about the year 260O, 

 remainiui;- jiurtial for two or three centuries more. 



The lialf-i-yrle lunar eclipse began its career 1900, June 

 13 (north limli) ; it becomes total 2044, September 7, and 

 reaches its climax about the same time as the solar eclipse, 

 ending about the same time as the latter ceases to be 

 central. 



It not infrequently happens that two eclipses, nine 

 years apart, at opposite nodes, pass one another about the 



