757 
of Edinburgh, Session 1871-72. 
great change which happens for the sun, between September and 
November, occurs earlier, or between September and October for 
the moon. 
5. The range of the variation is greatest in January, and is least 
in May and October ; the arc, including the mean diurnal variation 
for January, from eleven years’ observations, being nearly 0'*5, 
while in the latter months the ranges were nearly C018 and 0'*14 
respectively; the range for July being O'^G. 
The author states, that, in a paper already published,* be has 
shown that the range of the diurnal variation amounts sometimes 
to five minutes (5'‘0), which, from the less value of the horizontal 
force, would he equivalent to about twelve minutes (12'*0) in Eng¬ 
land ; and that the diminution of range appearing in the mean of 
many lunations is due to the combination of variations following 
different laws. 
6. The ranges of the mean lunar and mean solar diurnal varia¬ 
tions thus obey different laws with reference to the period of the 
year; the range of the former in January being nearly double that 
in any month from May to September, while the range of the latter 
in August is nearly double that in January. 
In the discussion for the change of the law which might be due 
to the moon’s passing from one hemisphere to the other, the author 
found different results for different months of the year; this led 
him to perform the calculations in a new way, described by him, 
in which the law derived from observations made during the day 
is separated from that obtained from observations made during the 
night. From this discussion it follows— 
7. That the action of the moon on the declination needle is, in 
every month of the year, greater during the day than during the 
night; the range of the oscillation in January and June being nearly 
four times greater during the day than during the night, the ratio 
being less in the intermediate months. 
When the results are derived from the forenoon hours only, or 
from the afternoon hours only, the range in January is six times 
greater than that derived from the night hours only. 
It also appears that the law derived from the night hours varies 
little in the course of the year; it is only that derived from the 
* Trans. Roy. Soc., Edin. vol. xxiv. p. 673 
5 h 
VOL. VII. 
