208 REPORT—1896. 
in which tremors for successive days have been placed under columns 
representing 24 hours, 12 hours being midnight. These tremors had 
values assigned to them equal in millimetres to the range of motion 
they exhibited on the photograms. By adding these columns up verti- 
eally a value was obtained for the period considered for each hour of the 
day and night. This value has been considered as proportional to the 
intensity of motion exhibited at various hours. By simply adding up 
the number of entries a set of numbers were obtained which may be 
regarded as proportional to the tremor frequency. 
These two sets of numbers obtained for the months of November and 
December 1895, when plotted on squared paper, give the curves shown in 
fig. 8. 
From these curves we see that for the period considered tremors have 
been least intense and least frequent between 3 p.m. and 7 P.m., but from 
Fig. 8.—Tremors November and December 1895. 
4 
7O =! Set) 
Hows0O 1234 5 6768 9 0 i1 12 1G HIS16 17 1879 20 7 2 23 24 
PM. AM. 
the latter hour there is a rapid increase in both these quantities. The 
intensity falls off rapidly from about 6 or7 A.m., whilst the frequency 
commences to diminish about five hours later. 
From these observations it would seem that the cause of tremors may 
possibiy be found in operations which grow in intensity during the night. 
and which become gradually enfeebled during the day. 
Tremors and Air Currents, 
Tnasmuch as the atmosphere may be calm, and the air inside an 
observatory may always be apparently quiescent, and yet an instrument 
