302 PROFESSOR FORBES ON THE CLIMATE OF EDINBURGH. 
height above the sea not having materially varied) has very slightly, if at all, 
affected the indications of the thermometer. 
20. The observations are recorded in seven quarto paper books, all apparently 
original. My attention has been almost entirely confined to the thermometrical 
observations, though I have also used the monthly totals of the rain-fall which 
were first collected by Mr Joun ADIE. 
21. My first business was to obtain the mean temperature of each day of the 
forty years’ observations. This was done with great care by Mr Grassick, clerk 
in Messrs Apiz and Son’s establishment. For the ten and a half years, 1795-1805, 
the mean of the eight o’clock morning and evening observations of the thermo- 
meter was taken and entered in vertical columns under each day of the year. 
22. Beginning with 1822, the mean of the maximum and minimum tempera- 
ture is taken to represent the mean of the day; for 1821 on/y the mean of the 
10 a.m. and 10 p.m. observations was used. I have not attempted any correction 
for the more accurate estimation of the mean temperature than is given by these 
several times of observation. They are known all to coincide closely with the 
mean in the long-run. The degree of coincidence of the mean of the 10 a.m. and 
10 p.m. observations with the mean of the maxima and minima may be appre- 
ciated by an examination of Mr Aprs’s printed registers in the Edinburgh 
Journal of Science. As a general rule, I am averse to the application of correc- 
tions to particular meteorological data, depending upon averages merely, because 
they fail to reproduce the very peculiarities of which we may be in search. In 
deducing probable results for the climate at one place from that at a neighbour- 
ing locality, the possible error arising from such corrections must be submitted to. 
But in most other cases I prefer using, when admissible, those observations in 
their pure and simple form which give most nearly the mean temperature of the 
day (omitting any others) to applying to individual results any considerable 
correction or factor derived from long averages merely. Such corrections may, 
if wished, be easily applied to the jinal average of a long course of observations, 
to which alone they are strictly applicable. 
23. The sums and means of each column (Art. 21) of daily mean tem- 
perature were taken for four periods of ten years each, into which the whole 
of Mr Apir’s observations were divided; viz., for the periods 1795-1804, 1821-30, 
1831-40, 1841-50. The total averages were then taken. 
24. In like manner, the monthly averages for each year were taken, and the 
whole combined in the usual way. The mean temperature of the month, for each 
period of ten years, was first taken by the mean of the averages for each day 
of the month (or means of the vertical columns), and again, by the mean of 
averages of the months for each year separately. 
25. Another check on the accuracy of the. computer was obtained by the 
circumstance that Mr Apr had already, in by far the majority of instances, taken 
the means himself. Mr Grassicx’s calculations were made independently of these 
—— ee we hl — 
o Os Bei wl eee. 
