57 . 



intellectual feat that the path of an eclipse can be marked 

 out at all before the event is to occur." A slight error in 

 the position of the moon would cause a large difference 

 in the path of the shadow on the earth. "In point of 

 f\ict, the time of the late eclipse as predicted was nearly 

 half a minute in error, while the actunl path of the shadow 

 was some ten miles from the predicted one." In order to 

 obtain the errors of the eclipse calculations, it is custom- 

 ary for the observers to note the time when the moon 

 first notches the limb of the sun, the duration of totality 

 and the time of last contact. When these are collected 

 together from the different stations of observation, they 

 are used to correct the lunar tables. That this may be 

 done, the errors of the chronometers used and the lati- 

 tude and longitude of each station must be known. Ob- 

 servations for time and latitude are usually made with 

 a sextant as is done at sea, or still better by a portable 

 transit instrument such as surveying parties or watch- 

 makers use. If a place whose latitude and longitude are 

 known is near l)y, the position of the observing station 

 may be referred to it by the methods used in surveying. 

 If there is telegraphic connection with a place whose 

 position is known, clock signals may be sent from one to 

 the other, and the difference in time which is the same as 

 diflerence in longitude be easily determined. At all the 

 eclipse stations connected Avith the Western Union tele- 

 graph lines, the clock signals from Washington were 

 received on several days, which was of great advantage. 

 Some of the observers made extensive observations 

 during the partial phase, by measuring the angle which 

 the line joining the cusps makes with the apparent path 

 of the sun from east to west, which can also be used for 

 correcting the moon's place. A series of photographs is 

 sometimes taken for the same purpose. As it takes about 



