AND SEMI-ANNUAL SEISMIC PERIODS, 
1119 
Period. 
Maximum. 
Amplitude. 
Annual. 
Semi-annual. 
March, h 
J March, a 1 
I September, a j 
•17 
•12 
Thus, (l) the amplitudes of both periods in all three classes are nearly the same, 
though somewhat less for the slight shocks than the others ; and (2) in the case of 
the annual period, the maximum epoch for the slight shocks agrees nearly with the 
minimum epoch for the more severe ones. 
16. In nearly every case examined below, in which we have two or more records 
for the same area, it will be seen that the amplitude increases with the number of 
registered shocks per annum. Now, two records can only differ in this respect by the 
omission or inclusion of slight shocks, for severe ones are almost certain to find a place 
in every list. Thus, in these cases, the slighter the earthquakes the greater is the' 
amplitude of each period. 
This is especially evident when, with the same catalogue, we use different defini¬ 
tions of the unit earthquake. This has been done in three cases, namely, Zante 
(§§ 45, 46), California, &c., (§§ 59, 60), and New Zealand (§§ 76, 77). 
Again, for Zante, the somewhat scanty list compiled from the catalogues of 
Schmidt and Fuchs (§47) gives the maximum of the annual period in December, 
and the amplitude '29 ; the much fuller catalogue of Barbiani (§§ 45, 46) gives the 
maximum in August, and the amplitude '10, or, using Barbiani’s definition of the 
unit earthquake, the maximum in August, and the amplitude '29. Thus, the effect 
of the first inclusion of slight shocks is to reverse approximately the maximum and 
minimum epochs, but to reduce the amplitude ; the effect of including a larger 
number of still slighter shocks is to maintain the reversal of epoch and to again 
increase the amplitude. 
17. Summary of Results. —(l) There appear to be two distinct classes of slight 
shocks, one having the maximum of its annual period in winter, the other in summer. 
Of the two classes, the former probably contains the stronger shocks, for their effect 
is evident in the results obtained from the less incomplete of two imperfect 
catalogues; while the effect of the latter is only visible in the results obtained from 
the fullest records, for example, those of the Imperial Meteorological Bureau for 
Japan, and of M. Barbiani for Zante, or from the list drawn up from Mallet’s 
catalogue in which only very slight shocks are included. (2) In both classes of slight 
shocks the annual period is more marked than for strong shocks, for the effect of 
including the first class of slight shocks is to keep the winter maximum epoch 
unaltered and to increase the amplitude ; that of including some only of the second 
class is to interchange the maximum and minimum epochs approximately, the 
