4 
extremes and means of climatic temperature; the horticultuit ha 
to contend with so great a difference in the distribution of taper! 
at the same season in different years, that it is of the first iiporf 
to determine accurately the average of temperature of every ay h 
year, with the extremes to which it is liable. It is known ththe 
disturbance of the animal econcfrny takes place when the rmn te 
rature of the air is that of the average of the season, and tlj gre ' 
disturbances take place whenever the temperature of the ir de] 
the greatest, either above or below that average; and there iiuo d 
that the same general law affects all vegetation in a simila! mai 
and that the effect follows the cause after an interval of tinn It 
these considerations which induced me to confine my first renctic 
this long series of observations to temperature alone. 
The instruments used were 
A Daniell’s hygrometer, and maximum and minimum thenom 
of Rutherford’s construction made by Newman. (These thenom 
were placed in an open spot in the Arboretum, screened frorthe 
of the sun and sheltered from radiation by a kind of umbrell of ■ 
cloth; they were attached to the northern side of the post viicli 
ported the umbrella, and are four feet above the ground.— See-age 
Yol. vii. Trans. Sort. Soc .) 
There is no record of change of instruments. The scale sed 
that of Fahrenheit, except in the years 1885, 1886, and 1837, her 
centigrade scale was used. 
Till the year 1844 there are no MSS. observations, but thefbse 
tions are published in extenso in the Transactions of t. It 
Sorticultural Society. From the year 1845 the observatior an 
MSS. 
The first process in the reduction of the observations wa da^ 
day, to see that the several thermometrical observations in he 
were less than the maximum and greater than the minimal am 
note all discordant readings. 
The second step was to examine these discordant readings for 
purpose all the observations made at the Royal Observatory, Ghfenw 
on that day were consulted, for readings taken at about the sai.3 tii 
and to determine the general course of increasing and dereas 
readings, and amount of change. A great many errors we t 
found, principally in the readings of the minimum fhermomeh 
The third process was to take the daily sums and means of lie tl 
observations of morning, noon, and night temperatures, fora i 
approximation of mean daily temperature. 
The fourth was to take the daily sums and means of the irjxim 
and minimum temperatures, for a second approximation t d* ‘ 
daily temperature. 
