224 POOR 



vations with the accuracy that could be desired. As pointed 

 out by Bauschinger, the differences between the observed and 

 computed places of the comet varied periodically. In the first 

 appearance, 1889-90, the point of the comet which was meas- 

 ured, appeared a little south, and in the second appearance, 

 1896-97, a little north of the places computed from the mean 

 path, and the total differences amounted to about 3".o. Bau- 

 schinger explained these discrepancies as due to one of two 

 causes. First, that different portions of the comet were mea- 

 sured during the two appearances, -owing to local changes in 

 the brightness of the object. Second, actual changes in the 

 position of the nucleus of the comet, due to perturbations by the 

 smaller, and in the latter appearance, invisible, companions. 



The elements, as deduced by Bauschinger, were carried for- 

 ward by Neugebauer, and an ephemeris computed for the third 

 appearance in 1903. The comet was rediscovered on August 

 18, 1903, by Aitkin, at the Lick Observatory and was found 

 to be very close to the predicted place ; the difference between 

 observation and theory being about 24' in Right Ascension and 

 i' 30" in Declination. 



7. It seemed desirable to correct my work, heretofore men- 

 tioned, by means of the latest data, and to collect and publish 

 in final form the results of my various investigations upon the 

 appulse to Jupiter in 1886 and the consequent changes in the 

 orbit of this body. In this I was greatly aided by a grant of 

 money by the Trustees of Columbia University, which was 

 used in employing assistant computers. The first step was the 

 selection of the elements on which to base my investigation 

 and here at the outset I was confronted by a difficulty. A 

 comparison of Bauschinger's elements with the positions of the 

 body at the time of discovery in 1903 showed that these ele- 

 ments were not all that could be desired, and a redetermination 

 of the elements was decided upon as a preliminary. In this 

 redetermination all the data from the first two appearances were 

 used and the elements deduced therefrom were checked with 

 one or two normal places from the 1903 appearance. Thus 

 the elements may be regarded as definitive for the appearances 



