MADE AT THE APARTMENTS OF THE ROYAL SOCIETY. 
309 
were made, was so great, that, unassisted, I hesitated to begin the reductions, although 
the results, as printed in the Philosophical Transactions, are unfit for application to 
useful purposes, there having been a departure in the observations themselves, from 
an absolutely necessary condition, viz. that of taking the observations at stated times, 
when the diurnal variations were unknown. 
About two months after this, William Farr, Esq., who is at the head of the Statistical 
Department in the office of the Registrar-General, wished me to supply the meteoro- 
logical particulars of as many years as could be given with certainty, to accompany 
statistical tables upon which he is engaged. Finding therefore that the demand for 
the results from trustworthy observations, extending backwards many years, was in- 
creasing, with the knowledge that such could be deduced from the observations of 
this Society, together with the hope of connecting the Greenwich series of observa- 
tions with these, I readily undertook to perform the work. 
I have now the honour of presenting to the Society the results from all the Ther- 
mometrical Observations which have been taken at Somerset ITouse. I have chosen 
these in preference to the Barometrical, as being at present more important, and more 
immediately useful. The prevalence of epidemic complaints renders it desirable to 
compare the simultaneous meteorological conditions with those, when no particular 
disease prevailed. The cholera epidemic now prevalent has caused me to prepare 
this paper as quickly as possible. 
I shall now proceed to explain the manner in which the annexed Tables were 
formed. 
Table I. was made by applying the corrections to the mean of the observations 
made during every month, according to the times of the day at which they had been 
taken, and thus determining, from the observations taken during the day, one mean 
temperature for the month, which has been used as the true mean at all times when 
the self-registering thermometers have not been used ; at times when they were in 
use a second mean has been found by applying the corrections as mentioned in my 
paper in the Philosophical Transactions in 1848. Thus two values of the same ele- 
ment have for the most part been found monthly ; the difference between these re- 
sults in most cases was a quantity less than a degree. As both determinations rested 
upon two observed readings daily, I considered them entitled to equal weight, and in 
most cases I have taken a simple arithmetical mean betvveen them, which I have 
adopted as the true mean for the month. From the accordance thus found by these 
two methods, I inferred that either could be used with safety at times when both sets 
of instruments were not in use. 
2 s 
MDCCCXLIX. 
