MAGNETIC INCLINATION AT GOTTINGEN. 661 
All the observations were made in the forenoon. 
In the calculation the values of 2x", y", e', found above, were 
employed for needle 3, but for needle 4 the values of x", y!", el” 
had to be deduced as well as they could from the observations 
themselves, whereby it was found that 
yl" —— eee 1103" 
v= + 556! 
el” —e= 424", 
The determination of y/”, resting on so few observations, is no 
doubt very uncertain ; but the influence of the uncertainty on 
the reduction of + (f+ g) is quite insignificant, as the greatest 
coefficient of y'", in the equatigns I., is only 0°00341. The results 
for 7 then stand thus :— 
Needle 3. Needle 4. 
° / LA [e} ‘ “ 
q May 15. | 67 38 57 | 67 40 0 
ve 40 36 4] 36 
18. 41 15 40 15 
19. 41 19 40 16 
The weight of the determination of e!"—e is here only twice as 
great as the weight of 3 (f+ g), and as the observations them- 
selves have considerably less exactness than the later ones, it is 
obvious that the value now found says as little for the reality of 
a constant error as that which was deduced from the later ob- 
servations. 
The great departure of the values of «and y'" from those 
found in Art. 26, 27, only proves that the revolving part of the 
needle, considered in itself, has not its centre of gravity in the 
axis of rotation, which, however, is of little consequence. 
30. 
I now bring together the final results for the inclination, de- 
duced from all the observations which have been treated of, and 
include among them the observations of the 1st and 7th of 
August, which have been already mentioned, and which were 
made with needle 1 in a similar manner in all respects to those 
on the 23rd of September. These observations were the fol- 
lowing :— 
August 1. August 7. 
4 tZA / 
r | 673012 | 67 2241 
4411 | 67 42 8 
7! 59 53 | 68 156 
g | 35 53 | 67 35 46 
