26 Professor J. D. Everett's Description of a Method 
By taking the sum and the difference of A and a, we shall 
obtain approximately the mean temperatures of the warmest 
and coldest groups of 30 days respectively at each of the 
stations ; or if the difference between the warmest and coldest 
group is required, it can be found by simply doubling a. 
These determinations will, however, be inferior in accuracy 
to those which the table contains, and this is my reason for 
omitting them. (See note, page 23.) 
With the joint purpose of testing the powers of the method, 
and comparing different years, I have calculated the values 
of a and e for single years for a few of the Society’s stations, 
including three which are not contained in the first table. 
The results are given, without any reservation, in the follow- 
ing table (Table II.). Bressay (Shetland) appears to be the 
latest of the Society’s stations, being about 13 days behind 
the mean of the 24 stations included in the first table. 
Sandwick (Orkney) precedes Bressay by about 2 days, and 
this interval is preserved nearly constant from 1857 to 1858, 
although the absolute times differ by nearly a fortnight. The 
amplitudes are also less for these two stations than for any 
others, the amplitude (and consequently the range) at Bressay 
being only about four-fifths of the average derived from the 
24 stations. The extreme lateness of Thurston (near Dun- 
bar) seems to be borne out by the results from single years, 
as appears from a comparison with the neighbouring station, 
East Linton. The extreme earliness of Yester cannot be so 
satisfactorily tested, as the interpolations (in defect of actual 
observations) at this station for the years 1857-58 are too 
numerous to admit of any safe inferences being drawn from 
these two years. In 1856 (which year is entirely free from 
interpolations), Yester appears to have been 16 days earlier 
than Thurston, and 11 earlier than East Linton, a remark- 
able difference, considering that all three places are in the 
same county (Hast Lothian). Comparing one year with an- 
other, it appears that the seasons were latest in 1857, being 
fully a week later than in 1856. The greatest difference 
appears at Thurston, where it amounts (in comparing the last 
two years) to nearly 21 days. All the inferences as to date, 
contained in this paragraph, are derived from mere inspection ° 
