688 KANSAS CITY REVIEW OF SCIENCE, 



tation more tedious and accurate than can be done by the Saros. And so I find 

 that in eighteen years more, viz., 2365, May 21st, at two P. M., the shadow of 

 the moon seems to miss the earth entirely. And now it must continue to traverse 

 the void expanse of space for the long period of nearly thirteen thousand years 

 before it again touches the south pole of the earth to begin another series of spiral 

 visits northward on the mundane sphere. 



Now, we may find when eclipses of this series occurred previous to 1806 by 

 subtracting the Saros period. Then from 1806, six months, fifteen days and twen- 

 ty-two hours take eighteen years, twelve days and eight hours, and we have 1788, 

 June 3d, two hours past midnight, Kansas City time. By reference to a long 

 catalogue of eclipses in Ferguson's Astronomy, I find that this eclipse — that is, 

 one eclipse in this series — was visible at London 1788, June 4th, at nine A. M.; 

 which is very near our Saros time — allowing for difference of longitude. Then 

 from this date we subtract eighteen years, ten days and eight hours and get 1770, 

 May 26th at 6 P. M., which would be visible in western Asia and the Pacific 

 (each preceding eclipse being visible eastward instead of westward, as when we 

 added the Saros). So we might trace backward period after period for several 

 centuries, and obtain, nearly, the times when this eclipse has been visible in the 

 remote past. But remember, it has been going northward at every succeeding 

 period ; and as we go backward, it was farther south in the preceding periods. 

 It was central on the equator, I think, about 1700. The farthest back, or earliest 

 record, that I find of this eclipse was 1572, January 25th, at seven A. M., Lon- 

 don time. At this hour in the morning the sun had not risen at London ; but 

 the eclipse must have been visible as a total eclipse in southern Africa. I have 

 not applied the Saros back to this date, but calculation by a different process 

 agrees exactly with the record. By a careful and somewhat tedious computation 

 I obtain the date of 1049, March 12th, when this eclipse first came on the earth 

 at the south pole. In 103 1, March 2d, the moon's shadow seemed to nearly 

 graze the earth, and at the next return in 1049 was fairly on ; so that, from the 

 south pole, the moon would seem to cover about one-twentieth of the sun's diam- 

 eter. Although this eclipse was largely total when it passed over Ohio in 1806, 

 yet my computations indicate that it was annular when it first struck the earth in 

 1049. That it became total about the time of its return in April, 1103, and will 

 continue so until 1969, when it will again become annular; that is, the moon will 

 get farther from the earth and appear less than the sun. 



Not having found in any work on astronomy an account of this change from 

 annular to total eclipse, I had always thought that if the eclipse was annular when 

 it began to appear on the earth, it would continue so during its whole periodical 

 round of a thousand years or more. But I find in these calculations a slow and 

 continuous change in the relative distances of the sun and moon from the earth, 

 Saros after Saros causing a change from the ring to the total eclipse. Then, to sum- 

 marize a few of my results in reference to the solar eclipses of 1806, 1878, etc., 

 viz: It came on at the south pole of the earth as an annular eclipse March 12, 

 1049; became total April 15, 1103; reached the equator — middle of its spiral 



