ON ELECTRICAL OBSERVATIONS AT KEW. 119 
has ears much Meher ta than 50 div. at the hours of 0, 2 and 4. 
The inference undoubtedly is, that 
the means at those hours are too low, 
and as a consequence, the mean of 
each year, as well as the mean of all 
the observations, is also foo low. With 
regard to the hour of 6 a.m., the 
value appertains only to the summer, 
very few observations occurring at 
this hour in the winter. When we 
come to discuss the seasons, it will be 
seen that the higher tensions inva- 
riably occur in the winter; the value 
at 6 A.M., upon the whole year, is 
therefore also too /ow ; consequently, 
were we in possession cf either an un- 
interrupted series of personal observa- 
tions during the day and night, or care- 
fully executed photographic registers 
for the same period, we should doubt- 
less have a curve which would exhibit 
neither so rapid a rise from 6 A.M. to 
8 A.M., nor so great a fall from 10 
P.M. to midnight, but would at these 
hours be more in accordance with its 
other portions. Of course it is im- 
portant, in reference to this point, to 
bear in mind the circumstances under 
which the observations were made, 
the personal establishment not having 
enabled the observer to continue the 
observations during the night, and the 
uncertain diminution of the charges 
of the night-registering electrometers. 
above 50 diy. rendering it preferable 
ie ae ta ae not to record the indications of the 
PI 4 instruments above 50 div., rather than | 
a 5 insert numbers likely to vary from 
the truth, and for which there are no certain means of correction. From a 
consideration of the tables and curves, it will be apparent, that the hour most 
suitable for observing the mean electrical tension during the entire year is 
§ a.o., the difference from the mean at this hour being 1°3 div. in excess. — 
Diurnal period. Summer.—The 10,176 observations from which the 
diurnal period having reference to the entire year has been deduced, are 
thus divided :— 
1845. 
1846. 
Fig. 2. 
1847. 

3 years. 
Mean diurnal curves of the electrical tension for the years 1845, 1846 and 1847, with the mean curve of the three years. 

SUMMEESs cl ea ek wee 5,514 
Winter ............ 4,662 
10,176 
The following table exhibits the distribution of the summer siseteaiions 
among the twelve daily readings: the months considered to constitute the 
summer half-year are, April, May, June, July, August and September :— 
