145 
Survey of Northeast Greenland. 
limb might differ by as much. as six seconds of time; however, as 
a rule, the difference was only one or two seconds. 
Ås will appear from the table below the determinations of time 
by means of the sun were carried out by zenith distances varying 
from 82° to 86”. The uncertainty caused by the refraction must 
here be supposed to be very great. As an error of 3” in the zenith 
distance near the prime vertical, as far as Danmarks Havn is con- 
cerned, gives a clock error of about 1°, the clock corrections 
computed from the observations of the sun must be rather unreliable. 
On the other hand it was only to be expected that a systematic 
error of the computation of the refraction would be comparatively 
more obvious in the case of the large zenith distances than otherwise. 
Results of determinations of time 
in September 1907. 
Astronomical | Clock time | Approximate | FOR pe: | Rate for 
date 57310 zenith distance | 57310 | 24 hours 
8. Sept, .... 58.9 84° —L 2803.1 | 
8. — (10.1) | Polaris’ vertical (26.1) | 
8. — 18.5 83 27.5 | 4 038 
Du 58 | 84 27.3 | 
9. — (9.9) | Polaris’ vertical (26.9) | 
pipe: 18.5 83 27.4 | 
11. — 5.8 85 271 À +083 
12. — 5.5 84 27.9 | 
12. — (9.4) | Polaris’ vertical (27.8) | 
DRE 19.1 82 25.5 | 
15. — 5.6 86 PAGLS | + 03.4 
Eu 18.7 84 | 24.8 | 
Ile == (8.6) | Polaris’ vertical | (25.8) | 
17. — 19.0 84 | 25.4 || AA 
ic 5.5 | 86 | 24.3 |) KR 
je (14.0) | Polaris’ vertical | (24.1) | 
18. — 19.0 84 | 22.6 | 
The rate is deduced from the transits on the vertical of Polaris, 
which in this case may be considered faultless. It will appear that 
the clock rate, though even, has been rather variable. 
The series of observations is very short. However, it shows 
with sufficient clearness that there is no pronounced systematic dif- 
ference between the observations of the forenoon and those of the 
afternoon. The zenith distances being so great it is rather surprising 
that the clock corrections computed from the observations of the 
sun are not more uneven or subject to greater errors. 
10* 
