THE APPARENT POSITION OF THE ZODIACAL LIGHT. 141 



half the sum of the latitudes of the boundaries, here designated as half the extent 

 of the light; second, the difference of absorption at the boundaines, expressed as 

 in Table I. ; third, the zenith distance of the north pole of the ecliptic, here 

 called the inclination of the ecliptic ; and fourth, the latitude of the axis, as above 

 defined. The next three columns, headed " Remainders," contain differences found 

 from the three preceding columns, by subtracting the quantities given in each 

 line from those in the next line below. The remainders in the last lines of the 

 evening and morning observations are found by subtracting the quantities con- 

 tained in them from those in the corresponding first lines. A system of correc- 

 tions, which will be explained below, was derived from these remainders. The 

 three final columns of Table II. contain corrected values for the latitude of the 

 axis. The first two of these columns result from connections determined by 

 the difference of absorption ; the same system of corrections is respectively ap- 

 plied in the two cases to the mean values in Table II., and to the separate 

 observations from which these mean values were derived. There is no material 

 difference in the result of these methods. The corrected latitudes in the last 

 column are determined by the inclination of the ecliptic, instead of by the differ- 

 ence of absorption ; the corrections were applied to the mean values in Table II. 

 As in Table I., negative quantities are indicated by Italic figures. 



We know already that the latitude of the axis varies to some extent in 

 accordance with the difference of absorption at the boundaries of the light, as 

 well as with the inclination of the ecliptic^ and it is obvious that Table II. ex- 

 hibits variations of this kind. We should presumably find similar relations be- 

 tween the latitude of the axis and the mean of the amounts of absorption at the 

 boundaries, or their ratio ; but there is little reason to suppose that any function 

 of the absorption would be better adapted to the present purpose than the differ- 

 ence here employed. As there is some probability that atmospheric absorption is 

 a phj'sical cause of the changes of position observed in the zodiacal light, it may 

 be well to begin by making the proposed system of corrections depend upon the 

 difference of absorption, and afterwards to form another system, dependent upon 

 the inclination of the ecliptic. After some allowance for the effect of absorption 

 has been made by estimation, Table II. seems to indicate the existence of addi- 

 tional changes in the latitude of the axis, dependent upon the longitude. In 

 attempting to obtain numerical corrections for the effect of absorption, we must 

 accordingly compare the results furnished by groups of observations at longitudes 

 not too widely separated. At the .same time it does not seem advisable to con- 



