312 MR. J. WELSH’S ACCOUNT OF METEOROLOGICAL 
stopcocks, from which the air had been exhausted. All the instruments which were 
at all liable to accident were supplied in duplicate. The construction of the meteo- 
rological instruments was confided by the Committee to Mr. P. Adie of London, 
under my own general superintendence. They were executed by him in a very 
satisfactory manner, having been made with much accuracy and with an anxious 
wish to promote the success of the experiments ; many of the mechanical arrange- 
ments for the convenience of observation having also been devised by him. 
Barometers . — ^The barometer employed was of the siphon form, on the construc- 
tion generally known as Gay-Lussac’s. The tube was affixed to a brass scale in 
much the same way as a thermometer is attached to its scale. The brass scale was 
fixed within a stout rosewood frame furnished with a door which could be closed 
during carriage. The diameter of the tube was 0*25 inch. The graduation was 
made from the middle point upwards and downwards ; each division being ^th of 
an inch long, but representing twice that value ; so that an observation of either 
branch of the siphon would give the length of the column of mercury, subject to a 
correction for inequality of the tube and error in the position of the zero-point of the 
scale. A complete observation of the instrument required however readings of both 
branches of the siphon, the true height of the mercury being the mean of the two. 
In order to facilitate rapidity of observation, verniers were dispensed with, the height 
of the mercury being merely estimated with reference to the scale placed behind it, 
just as if it had been a thermometer of large calibre. As it would have been nearly 
impossible to obtain in the car of the balloon a complete reading of both branches of 
the siphon for each observation, the corrections to the readings of the upper branch 
alone were previously obtained, throughout the anticipated range of the mercury, by 
the help of a large vacuum apparatus at the Kew Observatory, which has been 
employed in the pendulum experiments of Colonel Sabine and Professor Stokes. 
The barometers having been suspended within the receiver, the air was exhausted 
by about half an inch of pressure at a time, and readings taken from which tables of 
corrections were computed for different heights of the mercury. These corrections 
have been applied to all the observations. The difference between the indications of 
the siphon barometers and those of the Kew standard was also observed : both 
barometers were found to read 0-025 inch higher than the standard. It was found, 
by intercomparisons made last year, that the standard barometer at the Royal Ob- 
servatory, Greenwich, reads lower than the Kew standard by 0-003 inch. The balloon 
barometers thus read 0-028 inch higher than the Greenwich standard ; and, as that 
barometer has been generally referred to in the computations of height, the equation 
-1-0-028 has been applied to the terrestrial observations to render them comparable 
with those of the balloon barometers. Each barometer was provided with a thermo- 
meter to indicate the temperature of the mercury. In order to obtain this tempe- 
rature more accurately, the bulb of the thermometer (which was cylindrical, about 
inch long and ^th of an inch diameter) was immersed in mercury contained in a 
