DISCOVERY 



299 



eclipse would be both annular and total, the former 

 in one part of its track, the latter in another. This 

 is a very rare phenomenon, and is due to a slight 

 variation in the moon's distance from the earth 

 during the progress of the eclipse, the earth being all 

 the time close to the apex of the lunar shadow. Or if 

 the earth were actuaUy at the apex, the shadow would 

 just exactly reach it, with the result that a total 

 eclipse would be observed at the moment when a 

 perpendicular upward view could be obtained, but 

 this would be preceded and followed by some minutes 

 of the annular phase. One such eclipse took place 

 on December 12, 1890 (see Fig. 5). 



An interesting series in the twenty-nine-year period 

 is that containing the historical eclipse of Nineveh, 

 which was visible in the year 763 B.C. on June 15. 

 This is recorded in the Assyrian tablets. Adding the 

 period we are considering, we find other eclipses on 

 May 27, 734 B.C., May 7, 705 B.C., a series contiiuiing 



THE 



EARTH 



Annular 



Fig. 5.— showixg three positions of the shadow-cone. 



to our own times in the eclipses of 1814, July 17 (total) ; 

 1843, June 27 (annular) ; 1872, June 6 (annular) ; 

 1901, May 18 (total) ; the last-mentioned being the 

 ninety-third of the series since that of Nineveh. 

 Between each of these eclipses the lunar perigee — 

 that point in the moon's orbit at which it is nearest 

 to the earth — moves through about 242-4 degrees, 

 so that after three such intervals it has advanced 

 through 727 degrees, and, since 720 degrees is twice 

 round the circle, it is thus only 7 degrees in advance 

 of its original position. For a long period of, say, a 

 thousand years or more, every third eclipse of the 

 series — i.e. every one recurring at 3 x 29, i.e. 87 years — 

 will therefore be of the same character. Thus those 

 of 1814, 1901, 1988, in the above series are total. 



The dates mentioned above involve a change from 

 the so-called Old Style to the New. The latter was 

 introduced in 1582, and the formula for conversion 

 from one to the other can be set out in a word or 

 two. To get the New Style from the Old, if the year 

 under consideration comes anywhere — 



From 1582 to 1699 add 10 days. 

 1700 ,, 1799 ,, II ,, 



From iSoo to 1S99 add 12 davs 

 1900 ,, 2099 ,, 13 ,, 



Similarly, to change from the New Style to the 

 Old, the days must be subtracted. 



If a longer view be taken of this twenty-nine-year 

 period, it will be noticed that when it is multiplied by 

 eighteen, the few days short of twenty-nine years 

 amount to practically a whole year, the result being 

 that eighteen times the period is almost exactly 521 

 years. The consequence of this is that after this 

 lapse of time the eclipses recur on the same day of 

 the year. Thus if we take the Nineveh eclipse of 763 

 and add 521 a number of times, we get the years 

 242 B.C., 280, 801, 1322, and 1843 A.D. In each of 

 these years, therefore, the eclipse took place on June 15, 

 Old Style, or, using the above formula for 1843, on 

 June 27 in that year, a result agreeing with that already 

 given. 



Reference has been made earlier in this article to 

 the eclipse of 1927, June 29, the first to be total 

 in Britain for more than 200 j-ears. If we subtract 

 twice 521 years, i.e. 1,042 years, and convert from 

 New to Old Style we find that its predecessor occurred 

 in 885, June 16. This was total in Scotland, its 

 track being shown by Maguire in the Monthly Notices 

 of the Royal Astronomical Society (vol. xlv) to run 

 from S.W. to N.E., thus being practically parallel to 

 that of 1927. 



Since total eclipses are not very frequent, and when 

 they do occur are seen as such only from a narrow 

 strip on the earth, it will be realised at once that 

 they are ver}' uncommon in any locality, especially 

 in any particular town. Si.x only have touched any 

 part of the British Isles for the last five hundred 

 years, the dates being : 



1424 June 26 

 1433 June 17 

 1598 Mar. 6 



1652 April 8 

 1715 May 2 

 1724 May 22 



If we extend this list backwards for another 

 thousand years, we find eleven more total eclipses 

 visible in the British Isles : 



1023 Jan. 24 



1 133 Aug. I (C) 



1140 Mar. 20 (C) 



1185 May ilC) 



1330 July 16 



594 July 23 



603 Aug. 12 



639 Sept. 3 



664 May I (C) 



878 Oct. 29 (C) 



885 June 15 



" C " indicates a reference in the .Anglo-Saxon 

 Chronicles. The eclipses have included London twice, 

 Dublin twice, and Edinburgh five times. Of the 

 seventeen in the list, that of 885 was of the longest 

 duration, lasting in England for 4 mins. 55 sees., 



