Survey of Northeast Greenland. 175 
temperature in the lowest strata of air. Under these circumstances 
I thought that the means would express more correctly the actual 
state of affairs, if I left out the values of June 15th 1908 at 10P10 
from the computations. These values are given within brackets in 
the table below. 
Refraction conditions in the line of vision: 
the Observatory—the “Skerry”. 
Date | Hour | b, | 7, + 0.0342 | Ts | x k 
| | | 
| RTE | 
23. Aug. 07..... | 2p30 757 0.0775 273.6 | 0.529 0.296 
_-. ae | 8730 758 0.0542 | 271.9 0.372 0.466 
и | 315 | 759 | 0.0642 | 2717 | 0445 | 0.583 
u ete | 10240 | 759 0.0675 | 2723 | 0.466 0.449 
13. Sept.07..... 7200 | : 756 0.0475 269.6 | 0.338 | 0.324 
dé ee | 5240 | 758 | 0.0442 2682 | 0.314 | 0.243 
ET | 7210 758 0.0375 268.0 0.267 | 0.304 
EE GORE 11210 758 | 0.0409 | 269.1 0.289 0.254 
Cu 3210 | 758 0.0475 269.8 | 0.334 0.275 
SEM RES REPT! 120 | 757 | 0.0542 | 2664 0.390 0.364 
BROT: 8250 | 763 0.0542 | 2527 | 0.437 0.344 
|. re 1230 | 760 0.0442 258.0, | 0.341 0.332 
15. June 08..... 9210 156 | 0.0675 | 2752 0.454 | 0.448 
Г. 6210 156 | 0.0942 | 275.3 | 0.634 | 0.577 
ie 10210 | 756 0.1175: | 213.8 (0.795) | (1.719) 
der en: | 2a00 756 | 0.0609 | 2740 0.414 | 0.780 
OR MES 6210 | 756 | 0.1542 BBD >| ЛБА SE 920 
a ет NARRE 10830 | 756 | 01375 | 2746 | 0930 | 0.641 
ED. А | 2000 | 155 | 00509 | 2736 0.346 0.475 
Mean.. | 0.520 | 0.524 
The agreement between the means of x and k is, no doubt, to 
a certain degree an accidental one, but still points to the correctness of 
the value of k as determined by the reciprocal simultaneous zenith 
distances, so that Ir: — 0.516 may be considered as expressing in a 
reliable manner the co-efficient of refraction in the line of vision; 
the Observatory —the “Skerry”. In this manner the value determined 
for the line of vision: the Observatory —the Horizon II i. e. kl — 0.584 
also gains in certainty. 
A glance at the three tables of simultaneous values of x and k 
not only shows a very marked disagreement between the single 
values, but also shows that k varies very much more than x. For 
the values given in the tables which have been used for the compu- 
tation of the means, we get 
XVLI. 12 
= 
