150 PROCEEDINGS OF THE AMERICAN ACADEMY 



the ecliptic in the visible hemisphere should have been calculated 

 separately for the separate observations. As it is, they apply only 

 approximately to the mean results at particular elongations ; but when 

 these results are derived from considerable numbers of observations, 

 the approximation will be suificicnt for the present purpose. Means 

 are provided in the Appendix for the further reduction of the work, if 

 that should appear desirable, but I have not hitherto found time to 

 attempt it. 



The column of numbers is followed by two columns giving the 

 latitude of the axis of the light, and its half extent in latitude. By 

 the latitude of the axis is meant the mean of the latitudes of the 

 northern and southern boundaries. These data are derived exclu- 

 sively from those observations from which the latitude of the axis 

 can be separately determined. In the cases of Heis and "Weber the 

 number of observations for the half extent in latitude is sometimes 

 slightly inferior to that here given, which relates to the latitude 

 of the axis, as will be seen in Table XXVIII. The next two 

 columns of Table I. contain the zenith distances of the northern and 

 southern boundaries corresponding to the previous data. They are 

 followed by the amounts of atmospheric absorption due to each of these 

 zenith distances according to Miiller, expressed in terms of stellar 

 magnitude by dividing each logarithm of the original table (M. 59) 

 by 0.4. The last column of Table I. gives the difference between 

 the amounts of absorption in the preceding columns, expressed as a 

 positive quantity when the absorption at the southern boundary ex- 

 ceeds that at the northern. 



In the discussion of these results I shall first consider the question 

 raised by Serpieri (Sp. Ill), whether Jones was right in referring the 

 apparent changes in the place of the light to the corresponding changes 

 in the place of the ecliptic in the visible" hemisi^here, and whether they 

 ought not rather to be referred to the geographical position of the 

 observer in latitude. In the case of Jones, which Serpieri was dis- 

 cussing, the changes in the place of the ecliptic are very generally in 

 accordance with the movements of the observer. But the second 

 group of observations in Table I. exhibits a decided interruption of 

 this accordance. Since it may perhaps be conjectured that the 

 method of reduction is here at fault, I have determined the position 

 of the ecliptic for each of the twenty-four observations of this group. 

 The new reduction makes no change in any of the data, except that 

 the mean result for the " elongation of the zenith" becomes 122°, 

 instead of 123°, as in the table. This alteration is entirely insignifi- 



