Survey of Northeast Greenland. 165 
those generally indicated in other places. This is perhaps more 
apparent than real. The determination of the co-efficient of refraction 
is generally a secondary result of big levelling works, where it follows 
as a matter of course that efforts have been made to avoid condi- 
tions of refraction, which deviated very much from the normal. As 
far as the Danmark-Ekspedition was concerned the chief object was 
a general investigation of the terrestrial refraction, and here the ab- 
normal cases could only claim our notice to the same extent as the 
normal ones. 
From the above-mentioned work of ZACHARIAE I am going to 
quote a few extreme single values of k for comparison with the 
corresponding ones given by the Danmark-Ekspedition. 
Maximum Minimum 
DANCE OMR RENE an + 0.595 — 0.007 
Ostpreussische Gradmessung .......... — 0.154 0126 
Levelling Swinemünde —Berlin ........ — 0.267 — 0.083 
Kusienvermessung ©. olay. nn u an sauna. — 0.388 — 0.096 
Eandestmianeulalion #40. Le ua +- 0.173 —- 0.058 
CUS Ken ne ec sk — 1.208 + 0.114 
Danmark-Ekspedition: 
Pryelline a6 Gain ое. — 0.359 + 0.019 
IX - ver 
— — ID, Ma oi EEE ae kn + 0.427 k- 0.340 
У War 
— — be RER ER et о Е 08 
— Slee ren + 1.720 + 0.252 
— О Пе u a 209 + 0.173 
I have not had the opportunity of acquainting myself with the 
determinations made by Gauss, but judging by the great difference 
between the maximum and minimum values of k I think it most 
likely that here — as in the case of the Danmark-Ekspedition — a 
general knowledge of the terrestrial refraction was the object of the 
investigation. The minimum value given by Gauss, — 0.114, corre- 
sponds quite well with the minimum value of Cairn IX, in that 
Ken; as will be shown below may be put at 0.237. The extremely 
high maximum value of the “Skerry” is an isolated one. The second 
greatest value, also in the case of the “Skerry” is 1.509, and after 
that comes the maximum value 1.219 given for Horizon II. 
The mean value of the co-efficient of refraction at the sea hori- 
zon is determined as the mean of forty-five single values, distributed 
over the period 31st May to 21st October. Within this period ke 
— 0.584 may be considered a fairly reliable value. In contradistinc- 
