eral one lies in the difficulty of ascertaining the proper
correction to be applied for the refraction of the atmosphere, and the
higher the mountain the greater the liability to this error; for not
much is positively known about the angle of refraction of the upper
regions of the air. The officers of the Trigonometrical Survey of India
have published their opinion that the heights of the great peaks of the
Himalayas will have to be revised on this account. The report of the
Coast Survey's determination of the height of Denali claims a
"co-efficient of refraction nearer the truth" than the figure used on a
previous occasion; but a very slight difference in this factor will make
a considerable difference in the result.

The particular source of error in the case of this mountain lies in the
circumstance that its summit is flat, and there is no culminating point
upon which the cross-hairs of the surveying instrument may intersect.

The barometric determination of heights is, of course, not without
similar troubles of its own. The tables of altitudes corresponding to
pressures do not agree, Airy's table giving relatively greater altitudes
for very low pressures than the Smithsonian. All such tables as
originally calculated are based upon the hypothesis of a temperature and
humidity which decrease regularly with the altitude, and this is not
always the case; nor is the "static equilibrium of the atmosphere" which
Laplace assumed always maintained; that is to say an equal difference of
pressure does not always correspond to an equal difference of altitude.
There is, in point of fact, no absolute way to determine altitude save
by running an actual line of levels; all other methods are
approximations at best. But there had never been a barometric
determination of the height of this mountain made, and it was resolved
to attempt it on this expedition.

To this end careful arrangements were made and much labor and trouble
undergone. The author carried his standard mercurial mountain barometer
to Fort Gibbon on the Yukon in September, 1912, and compared it with the
instrument belonging to the Signal Corps of the United States army at
that post. A very close agreement was found in the two instruments; the
reading of the one, by himself, and of the other, by the sergeant whose
regular duty it was to read and record the instrument, being identical
to two places of decimals at the same temperature.

[Sidenote: Readings on the Summi