October, 1903.] 



KNOWLEDGE, 



225 



exact number of ceiitiiiics, and it frequently repeats tlie 

 (rack of totality with considerable accuracy. In fact we 

 can say with confidence that whenever the sun is totally 

 eclipsed at a consideiable altitude (totality rot beiu^' very 

 brief) there wiU be a total eclipse 300* later not very far 

 away. "W'l-.en the sun is very low, foreshortening' may 

 shift the track consideiably, as, for example, 1598 total 

 in Scotland with low sun, 1898 total in India. But the 

 other three eclipses on the diagram are repeated much 

 more closely ; notice in particular the resemblance of IGdO 

 to 19(10. It may be added that in 22o0 this will be total 

 in the North of England (Rev. S. J. Johnson). As other 

 exam2)les of the cycle, we may give 16(30 (Spain), 1860 

 (Spain), 2160 (France). 1551 '(Norway), 1861 (Norway), 



useful, besides lieing much harder to remember. Mr. 

 Stockwell suggested the use of a 372^ cycle (that is SOty 

 plus four Saroses). He thus satisfies condition D, but at 

 the cost of B, C, which it satisfies so badly that it does not 

 need further consideration. 



Note. — Mr. E. W. Maunder has drawn my attention to the fact that 

 a (500 year period (viz., two of these periodt) was known to theatcient 

 astronomers, though there does not seem to be much evidence as to 

 how it was difcoveied. Josephus (" intiquities of the 7cw6," 

 Chap. III.) says that the aneieijts had " a longer time of life on 

 account of .... the good use they made of it in astronomical 

 and geometrical discoveries, which would not have aflordcd tlie time 

 of foretelling [the peiiods of the stars] unless they liad lived 600 

 years; for the Great Year is completed in that interval." 



Thk Mboalosaeos. — This is by far the best long 



Plate III. — Tho 300-year Cycle. The last four Total Solar Eclipses and their counterparts three centuries earlier. 



2161 (England). 1415 (France), 1715 (England), 2015 

 (near Shcflands). 1424 (Germany), 1721. (England), 

 202-1. (Atlantic). 1781 (N.Africa), 2081 (France), 2381 

 (England ; this is from Rev. S. J. Johnson). 1406 

 (France; given above in the 521'' cycle), 1706 (Switzer- 

 land), 2006 (Egypt). 



Tho above ari^ all total eclipses; examples might be 

 multiplied ml lihihnn, hnt the above will suliice to establish 

 the utility of the cycle. 



It may be noted that a 246 year cycle (viz., 300* 

 minus the triple Saros) satisfies B, C, E better than the 

 SOO'. But it satisfies D so badly that it is not really so 



eclipse period known, and bears a remarkable analogy to 

 the Saros, as will be seen by the schedule above. It 

 was discovered a few years ago by M. Oppert ; he 

 apparently deduced it from theoretical reasoning, not 

 from a study of tabulated eclipses, as the Chaldwans 

 discovered the Saros. He considers that this period was 

 known to the ancients, which, however, seems to be very 

 improbable on the face of it, and to require convincing 

 evidence to establish it. 



In this long period the secular acceleration of the moon, 

 which causes her to go quicker and quicker each century, 

 becomes au important factor. It is possible that New- 



