706 PROCEEDINGS OF SECTION E. 
writer in Quarterly Journal of the Royal Meteorological Society 
vol. 23, page 305). Boiling- ‘point thermometers are, of course, 
fragile instruments, and require more care in transport than an 
aneroid barometer, though much less than a mercurial. On one 
of the writer's expeditions a bundle of blankets in which the 
thermometers packed tumbled down the side of a mountain 
for some 30 feet without injuring the instruments. <A further 
advantage with the thermometer is that if injured it discloses 
the fact at once ; whilst unbroken it is trustworthy, when broken 
it refuses to give any indication. On the contrary, a mercurial 
barometer with air in tube, or an aneroid barometer, the index 
of which has been deranged by a severe fall, will continue to give 
some readings and not of itself disclose the injury. 
But little need be said as to the methods of taking observations 
save that for accurate work simultaneous readings at upper and 
lower stations are essential. The writer does not believe in the 
system of reading the aneroid in the morning prior to leaving 
home, taking readings at different places during the day, again 
reading the aneroid on returning home at night, and expecting 
accurate determinations of the altitudes of the stations visited 
from such observations. In settled weather the ordinary atmos- 
pheric tides alone are sufficient to introduce most serious dis- 
erepancies, whilst in unsettled weather irregular fluctuations of 
pressure will upset the results. Care must be taken to observe 
the air temperature a the shade, and if the humidity formulz are 
used in the computations the temperature of the wet bulb must 
also be noted. With reference to the use of the boiling-point 
apparatus fresh water should be used for each observation, as if 
this be not done the percentage of impurities in the water becomes 
increased as portions of it are evaporated, and the boiling point 
is altered in consequence. The writer has not in his work 
experienced any difficulty in obtaining water of sufficient purity, 
and it should be remembered that water of uniform quality and 
not necessarily of absolute purity is required. For determining 
with fair accuracy the height of a point situated (say) 30 or 40 
miles from the base station, the writer considers that at least 
six or seven sets of simultaneous observations should be taken at 
(say) hourly intervals. Should the readings at each station 
remain fairly constant the work may be considered reasonably 
trustworthy, but should the barometers be rapidly altering accurate 
results cannot be expected. In such a case the work should be 
postponed til] a more suitable occasion. 
As to computing the results from the observations we are 
practically reduced to the adoption of a formula of one of two 
types, the first that in which the air is assumed to be of an average 
moisture throughout, and the other in which the actual amount of 
moisture at each station is observed and used in the computations. 
