102 Astronomical and Nautical Collections. 
N. A. Nw A. 
Alt. es fe 52° Inte Ivory Diff. Bessel F. A. 
oO 0 238.72 381 239,16 $1.2 ‘2 28.9 
anh 9 4.29 5.1 2 4.65 +1.0. 2° 4,4 
30 0 1 40.5 1 40.1 1 40.85 +0.7 1 40.7 
35 0 23.0) 1) 225% 1 23.25 +0.5 123.1 
40 0 1, [9.841 $9.0 1 9.52 +0.5 1 9.4 
45 0 58.1 57.9 58.36 +0.5 58.27 
50 0 48 8 48.6 48.99 +0.4 48.91 
60 0 $3.6 33.0 33.72 +0.2 33.67 
70 O SN ty lie 3 Ness 21.26 +0.16 21.23 
80 0 10.2 10.2 10.30 +0.10 10.29 
The results of the Nautical Almanac are reduced to 52°, im 
order to compare them the more readily with those of Mr. Ivory. 
For the mean probable error of a single observation, Professor 
Bessel and Mr. Rosenberger have found at 
Alt. Error Alt. Error Alt. Error 
oO r a“ °o V/ , a“ 
630 26.0 230 53 £30. 2.0 
prego spg7eod3g)%6 23, oP syg me, 214.7 
130. 10.6.....3.30..sid@ Gadelpathondiian 
2 0 Py ea ae oa ae Boe Ge 
And the probable ultimate errors of their determination of the 
mean refraction, at 1° and 2°, are found to be 2”.5 and 1” re- 
spectively. : 
With respect to the refractions below the horizon to which the 
table of Bessel extends, it will always be amply sufficient to take 
the mean horizontal refraction, and to increase it by its excess 
above the refraction computed for an altitude equal to the de- 
pression, and for the actual state of the atmosphere; except that 
if the temperature at the surface of the sea were known to be ele~ 
vated or depressed, it would be proper to correct the mean hori- 
zontal refraction accordingly. 
The whole of this comparison has been instituted in order to 
ascertain the propriety of retaining or suppressing the remark sub- 
joined to the table in the Nautical Almanac, that it ‘‘ appears to 
agree more perfectly with the latest observations than any other table 
that has been published,” explained as it is by the admission in 
the preface, that the “ deviation from the French tables in the 
