6 
PHYSICS: E: THOMSON 
the determination of the total height reached by them, curvature of 
the earth being allowed for. The effect of this curvature will be less 
the higher in altitude the auroral arch, or the nearer it is to being over- 
head. The streamers are often observed to rise from the arch first as 
short streamers, gradually developing and extending upward until 
their upper ends are a few degrees from the zenith. When they origin- 
ate in an arch which is of low altitude and extend nearly to the zenith, 
as they appear to do in the greater displays, their length must be hun- 
dreds of miles, possibly in some instances reaching one thousand or 
more. It would appear that no limit can be set for their possible height. 
In most auroras, however, the visible extent of the streamers is more 
limited. A low altitude auroral arch implies great distance north from 
the observer and for a given actual streamer length a less apparent 
height or length. Paulsen's class of auroras without streamers may 
mean either that the streamers are too short and too many to be noted 
separately, or that the electrification is too feeble for their development, 
the observed luminous glow being due to flow of current in the con- 
ducting layer itself, an arch forming horizontally, but without outward 
projection. 
If our assumptions are approximately correct the arch of an aurora, 
if located farther north than about 600 miles from an observer will be 
below his horizon but the streamers extending upward from it if long 
enough may be seen. Auroras far north of this will probably be invisible 
or be seen merely as a luminous glow well down on the northern horizon. 
When the breadth in latitude of the auroral zone is great and the dis- 
play is seen from the south, the streamers may overlap or be arranged 
in apparent curtains or folds, the lower ends of the streamers being in 
that case at varying apparent altitudes above the northern horizon 
even when in the same general line of view. In such case they may be 
superposed in the line of sight and therefore be composite, or increased 
in apparent length owing to imperfect superposition in their lengths 
with respect to those back or front of them. It is believed that these 
and like considerations will suffice for the explanation of observed 
appearances of auroras in spite of their great variety. 
The direction of streamers, as indicated, being vertical to the earth's 
surface, is suggestive of electrical discharges, ions or electrons projected 
outward into space from the conducting layer of our atmosphere; dis- 
charge into space in which the mean free path is unlimited. These 
discharges might have their origin in a charged conducting layer in 
which the potential suddenly rose to a critical value, or might be brought 
about inductively by the presence in the space around the earth of 
