Survey of Northeast Greenland. 461 
On the stretch from s, and in an outward direction thus only 
directly appear the objects projecting above the bounding surface 
between the strata of air. This is the case with the rocky holm 
S,S,S,F shown in the figure; the latter is directly visible in the 
angle 5, OS, and reflected in the angle S,OS,””. The part of the rocky 
holm situated below the bounding surface between the strata of air 
ought theoretically to be directly visible, but separated from the 
remaining part, in the angle s,Os,. However, I do not know whether 
we have ever seen this part of the image, the reason presumably 
being that the bounding surface between the strata of air often only 
lies 1 to 2 metres above 
the level ofthe sea. The 
section S,S, .thus be- 
comes so small that it Fig. 144. Downward mirage of the “Skerry”. October. 
can only appear as a [From ALFRED WEGENER: “Terminbeobachtungen am 
line, which it is difficult Vena Havn”, Medd. om Grenl. XLII). 
to discern, and which therefore in the case in point coincides with 
the false horizon, whereas in other conditions it may lose itself in 
the reflected image of the upper part of the mountain. 
We may now leave the optics and return to geography. 
The immediate impression which the observer receives of the 
optical phenomena is that — corresponding to the direction Os, — 
there is a bounding line, where a change occurs in the landscape, 
but the manner in which the change is regarded varies according 
to the character of the surroundings, and in particular of the back- 
ground. With a practically free horizon, which in ordinary circum- 
stances is only broken by single icebergs or holms, the case becomes 
as mentioned above. The reflected image of the celestial vault then 
comes into contact with the directly seen celestial vault; the two 
images coalesce; the real horizon disappears, and a rather sharp 
and extremely deceptive false horizon may appear, lying considerably 
lower than that which has disappeared. The holms and icebergs 
are seen floating freely in the air above the false horizon, perhaps 
a trifle fuller in shape than one is accustomed to see them, but 
otherwise to a casual observer with their usual contours. Only on 
closer inspection does one discover that their images are displayed 
fairly symmetrically on both sides of a horizontal axis corresponding 
to the bounding surface between the two strata of air. Just because 
the attention is first and foremost attracted by the directly visible 
part of the holms and icebergs, which stand out calmly and rather 
sharply with their usual upper contours, it makes an all the more 
startling effect to see them floating up in the air. 
If on the other hand one looks across a closed fjord, where the 
