162 REPORT—1849. 
as compared with the sunset mean tension, is nearly two-tenths more than it 
is in summer, é.e. in summer the value of the entire mean tension is about 
seven-tenths of the sunset mean tension, while in winter it is nearer nine- 
tenths: there is a greater proportional difference in summer than in winter. 
That this ought to obtain is evident from the consideration that the epoch of 
mean tension as before-mentioned is4 p.m. In the summer the epoch of 
sunset is nearer 8 p.M., at which hour the tension is higher than at 4 P.M., 
consequently the differences of the two sets of mean tension should from this 
variability of the epoch of sunset be greater in summer than in winter; and 
although this is not apparent when we contemplate the differences only, 
because of the great increase of tension in the winter, yet upon ascertaining 
the ratios of the one series to the other it becomes apparent, the proportional 
differences as we have seen being greater in summer than in winter. 
The irregularity of the curve of entire mean tension renders it doubtful 
whether these ratios ought to be regarded as at all sufficiently approximate 
to justify their employment in deducing from a series of sunset observations 
the entire mean tension. The month of July presents its usual anomalous 
character. October also presents a higher tension than the usual flexure of 
the curve would indicate as the mean, and. the displacement of the apex 
renders it difficult to apply any correction at present to the February mean ; 
nevertheless it is highly probable that a scale of corrections founded on the 
distance of the epoch of sunset from that of the mean tension of the entire 
year, and applied to the deductions from five or more years’ observations, 
would furnish a tolerable approximation to the annual period of the entire 
mean tension. 
The lowest curve in fig. 177 is that derived from the observations at sun- 
rise ; it partakes greatly of the character of the sunset curve, the flowing of 
the ascending branch only being interrupted by a greater increment of 
electrical tension in November than the mean monthly increment at this 
period of the year, and this would appear to be confined to November 1843 
(see Table LX VI. on page159). With this exception the sunrise curve follows 
the sunset very closely, the principal difference being in range, the range of 
the sunset exceeding that of sunrise by 32°2 div. 
We now come to examine the differences between the annual periods of 
sunrise and sunset. ‘These periods differ from each other not only in value, 
but, as we have just observed, in range ; the consequence is an inequality of 
the monthly differences between them. We have already alluded to the ap- 
proximation of the sunset curve to the mean annual curve derived from the 
observations of three years at the twelve observation-hours, the necessary cor- 
rections being comparatively small; it is however probable that the curves of 
sunrise and sunset approach much nearer in form to the true annual curve, 
which in value would come between them, and it is also likely that both curves 
may furnish true representatives of the annual period when certain corrections 
are applied, the value and range of the sunrise being necessarily lower than 
those of the sunset, from the observations contributing to its determination 
being made at a portion of the day characterized by a feebler development of 
the electrical tension. As the tension increases towards sunset in a certain ratio 
and according to a certain law which is most probably preserved during the 
annual progression of the electrical tension, the consequence would be that 
with increasing tensions at sunset we should have increasing differences 
from summer to winter and decreasing differences from winter to summer, 
and that from a sufficiently long series of observations either at sunset, sun- 
rise, or any selected hour, the mean annual period might be deduced. The 
following are the differences between the two series. 

