127 
mencement of the series of observations in 1848 to its 
termination in 1853. On the other hand Mr. Harrison’s 
results appear to support very strongly the hypothesis of a 
lunar period; but it is easy to show that they form no 
exception to the same general law. For this purpose I have 
made use of the St. Petersburg observations for 1856 — 64. 
being the years comprised in Mr. Harrison’s Oxford series ; 
and as the diagrams which accompany Mr. Harrison’s paper 
show a maximum of mean daily temperature on the 7th 
day after new moon, and a minimum on the 5th day after 
full moon, I have arranged the St. Petersburg mean daily 
temperatures for the five days from the 5th to the 9th day 
after new moon, and the five days from the 3rd to the 
7th after full moon in two tables. There were 111 luna- 
tions during the nine years, and the mean temperatures of 
the selected days were found to be as follows : — 
Days after New Moon. Mean Tem. Days after Full Moon. Mean Tem. 
5 th day 
. 38*32° F. 
3rd day 
. 38-88° F. 
6th „ 
. 38-30 
4th 
55 
. 38-79 
7th „ 
. 38-57 
5th 
55 
. 39-29 
8th „ 
. 38-16 
6th 
,, ........ 
. 39-62 
9th „ 
. 37-65 
7th 
55 
. 39-96 
Mean.... 
... 38-20 
Mean.... 
.. 39-31 
The temperatures of the days in the group after new 
moon, are all below the lowest in the group after full moon; 
and the mean of the first group is 1T1° below that of the 
second ; but at Oxford, during the same series of years, the 
highest temperatures occurred on the days comprised in the 
first group, and the lowest on those in the second. It is 
evident, therefore, that Mr. Harrison’s results cannot be 
