Prof. Challis on the Zodiacal Light. 1 19 



January 6 and December 25, 1685. The line joining these 

 apices will be approximately coincident with the line joining the 

 apices observed at the Cape near the summer solstice. 



The foregoing observations, taken together with all those col- 

 lected by M. Houzeau, indicate that the zodiacal light is always 

 visible before sunrise and after sunset, whenever the ecliptic 

 makes a large angle with the horizon of the place of observation, 

 and accord very well with the supposition that its form is that 

 of a double convex lens having the sun at its centre, and its 

 principal plane inclined at a small angle to the plane of the 

 ecliptic. After making the hypothesis that the position of the 

 apex in space is always on a straight line through the sun perpen- 

 dicular to the earth's radius vector, M. Houzeau has found by 

 the method of least squares applied to the fifty-eight positions 

 in his list, that the longitude of the node of the principal plane 

 is 1° 59', and the inclination to the plane of the ecliptic 3° 35'. 

 He has also compared the apparent inclinations to the ecliptic of 

 the planes passing through the earth, the sun, and the observed 

 positions of the apex, with the corresponding inclinations calcu- 

 lated from these elements on the above hypothesis. And having 

 made the same comparison on the supposition that the principal 

 plane coincides with the plane of the sun's equator, adopting 

 75° 8' for the longitude of the ascending node, and 7° 9' for the 

 inclination, as determined by M. Laugier {Comptes Rendus, 

 vol. xii. p. 649), he finds that the sum of the squares of the dif- 

 ferences in the former comparison is to that in the other as 100 

 to 121. Notwithstanding that this result seems to be in favour 

 of M. Houzeau' s elements, it appears, on inspecting the respective 

 differences between the calculated and observed apparent incli- 

 nations, that the law of the inclinations is not so well represented 

 by these elements as by the elements of the sun's equator. This 

 will become still more apparent after correcting the signs of the 

 inclinations deduced from the observations taken in the evenings 

 of December 25 and January 5, 1685, which by some mistake 

 are plus instead of minus. It is also to be considered that the 

 hypothesis on which the calculations rest is inexact except for 

 positions of the earth very near the nodes, and that the ten- 

 dency of the error, in the comparison with the elements of the 

 sun's equator, is to augment the plus differences in August and 

 September, and the minus differences in February and March. 

 When these circumstances are taken into account, it will be found 

 that the apparent inclinations corresponding to positions of the 

 apex on the north side of the ecliptic in the evening and the 

 south side in the morning, prevail generally from September to 

 March, and the opposite inclinations during the rest of the year ; 

 which is just what would take place if the nodes of the zodiacal 



