308 I. P. Kocu. 
The term THD is tabulated for the и D ша similar 
manner as shown on p. 305 for the term I E pp, after which the 
computation of the distances is performed = me engineer’s scale as 
mentioned above. 
The importance of a careful levelling at the sea horizon is thus 
to be found in the fact that in this manner it becomes possible to 
eliminate the strongly varying co-efficient of refraction (see, however, 
also the remark on p. 302). The dip of the sea horizon is, on the 
other hand, not so well suited for the determination of the altitude 
of the telescope above the level of the sea, when k is not known 
beforehand, which by the way will never occur in practice. 
By differentiating 
R 
h= eo a. 
we get 
R 2 h x 
И sum 0113. == en 
If for altitudes between 500 and 1000 m we put k — 0.15 
— — — 100 - 500 - — К = 02 
= = — 20 - 100 - — К = 0.3 
it may be supposed that dk аз far аз the greater altitudes are соп- 
cerned will rarely exceed 0.1, whereas dk as far as the lesser altitudes 
are concerned may easily become as high as 0.2. If one further 
takes into consideration that the error of the altitude determined by 
a is proportionate to — it will quickly be realized that this manner 
of determining the altitude will generally, as far as certainty is con- 
cerned, be considerably inferior to the above-mentioned primitive 
barometric measurement. 
In those cases where every other form of the measuring of alti- 
tudes fails, or where one wishes to apply a controi to the altitude 
determined in a different way, it may, however, be necessary or 
justifiable to compute the altitude by means of the dip of the sea 
horizon. 
For this purpose, one may, as far as North Greenland is con- 
cerned, make use of the following simple term: 
For altitudes between 500 and 1000 m (k — 0.15) 
h = 0.318 tan? a 
— — between 100 and 500m (К — 0.2) 
h = 0.338 tan? а 
= — between 20 and 100m (k = 0.3) 
в 20,386 tan? а. 
