220 REPORT—1896. 
causes, it seems likely that we should add to our records of the movements 
of the earth’s crust forms of vibration which horizontal pendulums and 
seismographs are incapable of recording. 
VI. Earthquake Frequency. (Extract from a letter written by 
Dr. C. G. Knorr.) 
_ In my paper on Earthquake Frequency (‘ Trans. Seis. Soc. Japan,’ vol. ix. 
1884), in which, probably for the first time, a sound mathematical treat- 
ment of periodicity was insisted upon, various possible causes of periodi- 
city in earthquake frequency were considered. Next to the solar annual 
and diurnal periods, the most important are the lunar monthly, fortnightly, 
and daily periods. From lack of completeness of information at that time, 
it was impossible to search for these. But the great eight years’ list of 
8,331 Japanese earthquakes, prepared recently by Professor Milne, seemed 
eminently suitable for harmonic treatment. Other necessary work has 
prevented me getting the investigation carried out so quickly as I had 
wished, but enough has been done to show the probable results in certain 
directions. 
The idea is that the tidal stresses due to the moon influence the perio- 
dicity. The lunar day gives a periodic tidal stress of comparatively short 
period ; the anomalistic month (from apogee to apogee) and the nodical 
month (from ascending node to ascending node), give periodic tidal 
stresses of long period. 
Tabulating the earthquakes according to the number of days each has 
happened after apogee, or after ascending node, we get two statistical 
tables of monthly means, one nearly 100 months. The anomalistic month 
is longer than the nodical month by almost exactly one-third of a day— 
in the hundred months, therefore, one will have gained upon the other by 
thirty days, or fully one month. The curves obtained, when created by 
harmonic analysis, give monthly, fortnightly, and weekly periods ; but 
the fortnightly is more marked in the nodical curve than in the anoma- 
listic. 
In discussing the daily lunar period, we must take account of the dis- 
tricts in which the earthquakes occur, for only in this way can we compare 
their times of occurrence with the time of meridian passage, or the time 
of high water. In the case of the Tokio and Yokohama district, there is 
evidence of a half daily period ; but the investigation is still far from 
complete. 
VII. Instruments used in Italy. By Dr. C. Davison. 
In the following pages a description is given of a few of the principal 
instruments used in Italy for the registration of pulsations proceeding 
from more or less distant origins. 
Many of the instruments erected in that country are long vertical 
perdulums, the movements of which are magnified and registered in 
different ways. The length is made as great as circumstances will allow, 
so that for rapid vibrations the bob may be practically a steady point, and 
the bob is made as heavy as possible, so as to lessen the friction intro- 
duced by the mechanical registration. Those who have used these pendu- 
lums claim that they possess the following advantages over the horizonta} 
pendulum and other instruments designed for photographic registration. 
1. They are much less expensive to work ; the cost of the paper on™ 
