146 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1937 



Photography was employed with success at the ecUpse of 1860, and 

 by its means it was proved that the prominences belonged to the sun. 

 The first triumph of the spectroscope at eclipse time was the discovery 

 of helium in 1868, though it was not isolated as an inert gas until 1895. 

 Coronium was discovered in 1869, but even today we do not know its 

 physical constitution. Before the eclipse of 1870, Young foretold a 

 sudden change in the appearance of the spectrum of the sun at the 

 time of the beginning of totality, from dark lines on a bright back- 

 ground to the sudden flasliing out of bright lines on a dark background. 

 His were the first eyes to see the spectrum of the chromosphere called 

 by liim the "flash spectrum" wliich was first photogra])hed but very 

 imperfectly at the eclipse of 1893. With better and better photo- 

 graphic plates, each eclipse, especially since 1900, has been assiduously 

 observed, distances away from home and difficulty of access being no 

 msurmountable obstacles to the eclipse observer. One of the most 

 spectacular discoveries from eclipses came in 1919 ^\^th the verification 

 of the relativity shift as predicted by Emstein. Succeeding eclipses 

 gave a more complete verification. In the following pages will be 

 given a more detailed account of the chief discoveries from the most 

 recent of eclipse observations. 



There are many different kinds of eclipses: Eclipses of the sun and 

 moon, eclipses of a star by the moon, eclipses of Jupiter's and Saturn's 

 satellites, and eclipses of one star by another of a binary system. 

 Observations of eclipsing binaries have given valuable information 

 regarding the masses and densities of stars — but for lack of space 

 the subject cannot be treated here. 



The accuracy with wliich the times of solar and lunar eclipses can 

 be predicted depends on the reliability of the work of astronomers of 

 all ages, and the manner in which the torch of learning has been handed 

 on from one generation to the next. The motion of the earth about 

 the sun is now laiown with a very high degree of accuracy. In spite 

 of the refined researches on the motions of the moon, the place in the 

 sky of our unruly neighbor is yet not known with a sulhcient degree of 

 precision. Tables of the moon are used for calculating the place of 

 the moon in the nautical almanacs prepared by different national 

 governments some 3 years in advance. For these almanac positions 

 it is desirable to keep the theory of the moon as free as possible from 

 arbitrar}'^ empirical terms so that the lunar theory may not be cluttered 

 up. When observations are secured, the results can be compared 

 directly with theory. Between the theoretical position of the moon 

 given by the almanac and the observed place there may be a difference 

 of several seconds of arc. 



For long-range predictions of lunar or solar ecUpses the simplest 

 method is to refer to Oppolzer's Canon der Finsternisse where are 

 given the elements of all the eclipses (8,000 solar and 5,200 lunar) 



