36 
of his paper, read at the previous meeting, on the Theory of the instru- 
ment. He exhibited his instrument, which was one of the most perfect 
yet constructed by Messrs. Negretti and Zambra. It was very portable 
and could be carried-about without fear of derangement. In making an 
observation, it was first to be suspended perpendicularly ; if this could be 
done from a nail driven into a wall, it should be, but if there were no such 
sheltered locality, the speaker used a light tripod which he had contrived 
for that purpose, and exhibited to the meeting, and which was to be carried 
with the barometer. Having hung up the instrument, the first thing to be 
done was to lower the mercury in the cistern, by a screw at the bottom, 
and bring it to a certain fixed level, which was determined by causing a 
little ivory cone to form one continuous line with its reflection in the sur- 
face of the mercury, perfect contact being thereby ensured. 
At this stage of the proceedings, Rev. Canon Moseley, M.A., F.R.S., 
one of the Vice-Presidents, entered the room, and on the motion of Mr. 
Thomas Pease, was called to the chair. 
Continuing his description of the instrument, Mr. Sanders said that the 
top of the cistern was closed with porous wood, and that it was necessary, 
in order to equalise the atmospheric pressure and temperature through the 
whole instrument, to allow ten minutes to elapse after bringing it into 
position, before taking the observation of the exact height of the column 
of mercury. This was ascertained by a scale of brass sliding over the 
glass tube, and he was accustomed always to allow the thinnest possible 
streak of light between the top of the mercury and the bottom of the 
scale. This gave a constant error, the same in all cases, but "exceedingly 
slight. The exact height was then read off by a vernier, to 0*001 inch. 
In ascertaining the difference of height between two stations, it was 
necessary to repeat the observation at the first station a second time, after 
having taken a reading at the second station, and if there were any 
difference between the two observations at the same place, the error was 
to be divided into as many parts as there were hours of interval, and a 
proportionate correction of the observation at the second station was 
to be made, according to the time which had elapsed between it and the 
first observation at the first station. 
He had explained the corrections to be made for temperature and gravity, 
and he now urged the importance of taking into account the diurnal 
variations in the barometer, according to the time of day, which might 
cause an error of 0*020 inch, or in other words, of 19 feet. This point 
had been noticed in no published treatise, and was therefore fully explained 
and illustrated by a diagram constructed from some results kindly for- 
warded in MSS. to the author by Mr. James Glaisher, showing the average 
hourly variation for every month in the year. The author then stated that 
he worked out his results by Mr. Phillips's formula, but that, in order to save 
