OcTOBEu 2, 1893.] 



KNOWLEDGE 



183 



combine to make the eclipse at one return a little different 

 from what it was the time before. And since these 

 differences continually increase, there is a steady and 

 progressive change in the character of the eclipse. 



To take one point alone — the fact that the moon arrives 

 at conjunction about an hour sooner than she arrives at 

 the node. The maximum distance from the node at which 

 an eclipse can take place is 18'^ ; 18° on either side of the 

 node gives a range of 3G°, or one-tenth of the circle, for 

 the extent of the region wherein an eclipse is possible. 

 One-tenth of 27 days 5 hours is 65 hours. If the laggiug 

 in the return to the node were exactly an hour, we should 

 then have the life-time of an eclipse extending to G5 or 0(5 

 returns ; as it is, it is a little longer than this, and like 

 the life-time of man himself it reaches three-score and 

 ten, sometimes a little more, but no eclipse " by reason of 

 strength " can attain to four-score returns. The usual 

 duration is seventy or seventy-one,* and curiously enough 

 the total number of eclipses in a particular Saras is also 

 seventy or seventy-one ; solar and lunar eclipses being both 

 included. During this time the eclipse will have travelled 

 down the year, eleven days at a time, twice and a little 

 over ; the last occurrence falling some thirty-six days later 

 in the year than the first. 



As has been already mentioned, an eclipse will generally 

 take place if the moon when " new " is not more than 17^ 

 from the node. Let us suppose that a time comes in the 

 history of a decaying eclipse when the moon is 18'' from 

 the node at conjunction, as in Fig. 2, where N represents 

 the node, and S and M the positions of sun and moon at 

 mid-eclipse. Wliat will follow? At conjunction next 

 month the moon will be some 30° further on in her orbit, 

 and will be about 17° from the node on the other side, as 

 at Mj, and in the general way we shall have a grazing 

 eclipse. We shall have, therefore, two partial eclipses only 

 a month apart. The next Saros will bring about a repetition 

 of the same state of affairs, only that the earlier eclipse of 

 the pair will take place further from the node, whilst the 

 second eclipse, the new one, which had been a mere 

 grazing eclipse at the previous Suros, would fall nearer 

 the node, and hence will give a somewhat larger phase 

 than before. This state of things is represented in Fig. 3, 

 where the new eclipse at M^ S.j has manifestly more body 

 than in Fig. 2. This process would be repeated several 

 times, the old eclipse continually receding from the node 

 and becoming smaller, the new one approaching it and 

 growing larger, until at last the moon would have drifted 

 too far from the node in the earlier month to cause an 

 eclipse. Henceforward there would be only one eclipse, 

 that of the later month. This would increase in phase 

 time after time ; would become annular or total ; would at 

 length fall close to the node ; would pass it ; and then, 

 receding from it on the further side, would begin to 

 diminish in phase. Lastly, as the seventy returns began 

 to run out, a time would come when this eclipse would fall 

 a full day or more before the node was reached. Then the 

 next month we should have the moon at conjunction just 

 near enough to the node to cause a very small jsartial 

 echpse, and once again a pair of eclipses would be formed. 

 As before, the earlier eclipse would grow smaller and 

 smaller, waning as the new eclipse waxed, and ere long it, 

 too, would in its turn cease to exist, like the eclipse 

 which it had superseded. 



This is the law, then, of eclipse life. Each is born as 

 the second member of a pair of eclipses ; each dies as the 



* The paragi'apli referred to at the beginning of tlie paper was 

 therefore wholly wrong. No eclipse lasts 2755 years. The usual 

 duration is 1262 years. 



first member of such a pair. The solar eclipse cycle is 

 the true Phaiiiix of the ancients — the Arabian bird which 

 sprang from the ashes of its dying parent, and which will, 

 in hke manner, give birth to its solitary offspring in its 

 own dying moments. It lives for 1262 or 1280 years, 

 forty generations of men, seventy of its own returns, and 

 passes away as the new existence, to which its decay has 

 given being, attains its iirst strength and vigour. 



As an instance of the life-history of an eclipse, we may 

 take the one which fell on New Year's Day, 1889. This 

 eclipse goes back to May 27tli, 933, when it was the second 

 of a pair of eclipses. The other member of the pair only 

 lasted for two ,Sanit:oi later, and on .June 28th, 9(S7, our 

 eclipse appeared fur the first time alone. All this time it 

 was a small partial eclipse, but the phase increased in 

 magnitude at every return ; and at its eighth appearance, 

 August 11th, 10.59, it was an annular eclipse. It con- 

 tinued as an annular eclipse until June 8th, 1564, the 

 thirty-sixth appearance, but at the three previous occur- 

 rences it had been partly total and partly annular ; the 

 thirty-fourth eclipse, that of May 18th, 1528, being the 

 occasion when the moon was closest to the node at con- 

 junction. The thirty-seventh eclipse fell on June 20th, 

 1582, and was the last eclipse before the Gregorian reform 

 of the calendar ; the calendar date of the thirty-eighth, 

 therefore, is July 10th, 1600, that being the New Style 

 date corresponding to June 30th, 1600, Old Style. The 

 eclipse has been a total one ever since ; its fifty-third 

 occurrence being the Algerian eclipse of December 22nd, 

 1870, and its fifty-fourth the Galifornian eclipse of January 

 1st, 1K89. It will still continue to be a total eclipse for 

 one hundred and forty years, and will be last seen as such 

 at the sixty-second appearance, March 00th, 2033. It is 

 a very large partial eclipse at the next return, April 11th, 

 2051, and at the following return, April 21st, 2069, the 

 dying Phtenix gives birth to its successor, for it is followed 

 on May 20th, 2069, by the first occurrence of a new series 

 of eclipses. The two series run on side by side for eight 

 returns, until the old eclipse reaches its end, as a very 

 small partial eclipse, on July 7tli, 2195 — its seventy-first 

 occurrence. 



If we trace the shadow-tracks of the eclipse from its 

 first occurrence as an annular eclipse in 1059, we find that 

 it began in the southern hemisphere close to the pole ; 

 indeed the shadow-track intersected the earth's axis beyond 

 the pole, so that there was no place where central eclipse 

 took place at noon. The centre of the eclipse on the next 

 occasion lay in south latitude 76° ; that is to say, the place 

 where central eclipse took place at noon, or the place where 

 the shadow-track cut the earth's axis as seen from the 

 moon, was in that latitude. The centre then moved 

 swiftly noi'th. In 1095 it was in south latitude 64° ; in 1113 

 in south latitude 60°. Then for one hundred and sixty- 

 two years it practically remained stationary, the reason 

 being that the 1113 eclipse occurred near the autumnal 

 equinox, and the iSnros bringing the following eclipses a 

 little further on in the year every time, the south pole was 

 tiu'ned more and more towards the sun almost exactly 

 at the same pace as the central shadow moved north. 

 Directly the winter solstice was passed, however — the 

 winter solstice of the northern hemisphere, that is — and 

 the south pole began to turn away from the sim, the north- 

 ward movement of the central shadow began to be felt ; 

 indeed the change in the presentation of the earth's axis 

 exaggerated its effect, and every successive return of the 

 eclipse showed a rapid rise in latitude, until the date of 

 the eclipse had travelled on past the summer solstice. 

 This fell in 1564, by which time north latitude 31° had 

 been reached, the equator having been crossed before the 



