DECEMBER 17, 1903] 
NATURE 
161 
En EEE ——————————— nn 
pressure. Tremors occur most frequently, but by no means 
always, with marked changes in pressure. An observation 
more novel in its character is that days when pendulums 
were much disturbed in Trieste coincided with correspond- 
ing disturbances in Strassburg. It is apparently taken for 
granted that the tremors recorded have a seismic origin. 
In No. 16 Dr. J. Knett publishes a list of 507 shocks which 
were recorded between February 13 and May 6 in north- 
west Bohemia, which he follows in No. 18 by an account of 
an earthquake which on November 26, 1902, disturbed the 
same country. 
No. 17, by Adolf Faidiga, is a lengthy description of 
the earthquake which, on July 2, 1898, created considerable 
destruction on the coast of Dalmatia, the vibrations from 
which reached Great Britain. It is largely of local interest. 
In Nos. 2 and 3 of vol. ix. of the Bolletino Della Socicta 
Sismologica Italiana, the well-known earthquake register 
published by this body is brought up to March 5, 1902. 
This is supplemented by two papers. In No. 2 Dr. A. 
Ricco adds to the knowledge we possess respecting the 
crater of Etna, and changes which are taking place in the 
‘same. On August 23, 1900, the interior of the crater was 
described and photographed. Its depth was then 282 m. 
On July 21, 1903, it was again photographed, and its depth 
was found to be 490 m. The supplementary contribution 
to No. 3, by S. Costanzo, is on the relationship or want 
of relationship between the wind and tromometric move- 
ments. The author apparently holds with Bertelli and 
others that the pendulum movements have an endogenous 
origin, and are not produced by the wind, but that they 
accompany, precede, or follow falls in barometric pressure. 
Dr. Agamennone regards the movements in question as 
being in great measure influenced by the wind. Our own 
experience, which is detailed in the Transactions of the 
Seismological Society of Japan, reports to the British 
Association, and in other works, is that vertical or hori- 
zontal pendulums, chemical balances, and like apparatus 
are set in movement when there is a marked fall in baro- 
metric pressure, a steep barometric gradient, or a marked 
fall in temperature; the movements more frequently occur 
during the night and in winter than during the day and 
in summer, and they may or may not occur with heavy 
gales. Instruments in a close atmosphere, as, for example, 
in a cellar, are more likely to be disturbed than similar 
installations in a well ventilated or even draughty room. 
Burning a lamp or a gas jet in a room frequently brings 
so-called microseismic storms to an end, and what occurs 
and creates annoyance in one room may not be observed in 
a neighbouring apartment. 
In vol. ii. of the reports of the Mathematical-Physical 
Society of Tokio, amongst forty-seven papers which for 
the most part are mathematical we find five by Dr. F. 
Omori and one by Mr. A. Imamura which relate to seismo- 
logy. The former of these contributions may, to a large 
extent, be regarded as epitomised reproductions of papers 
previously published by the Tokio Earthquake Investigation 
Committee, which we have already noticed (NATURE, 
April 30, July 9, 1903). _In a note on the seismograms of dis- 
tant earthquakes, Dr. Omori tells us that the ‘‘ motion con- 
sists of a series of different epochs, in each of which the 
period remains essentially constant, while the amplitude, on 
the whole, is also constant, except for the occurrence of maxi- 
mum and minimum groups.’’ In a table we find ten groups, 
in which the periods vary between 1-02 and 66 second. 
After discussing the amplitudes and durations of these 
various phases, it is shown that the arcual velocities vary 
between 2 and 11-3 km. per second; as the result of more 
accurate work, this last quantity is now raised to 14-1 km. 
per second. One argument in favour of the supposition 
that the preliminary tremors and other phases of earthquake 
motion follow arcual paths is that the durations of these 
successive phases are proportional to the arcual distances 
they have travelled, and these durations are approximately 
equal to each other. The speeds for certain earthquakes 
were determined by dividing the difference of the distances 
of Tokio and certain European stations from an origin by 
the difference of the times of the arrival of seismic waves 
in Tokio and Europe. This, it must be observed, involves 
the idea that the velocity of preliminary tremors on short 
paths or on long paths is equal. Lastly, it is assumed that 
NO. 1781, VOL. 69] 
each phase recorded at a distant station originated simul- 
taneously at the earthquake origin. 
The results to which we are led by assumptions of this 
nature are well illustrated in the paper, ‘‘ Notes on Milne 
Horizontal Pendulum Seismograms,’’ by Mr. A. Imamura. 
In this paper records of the Guatemala earthquake of April 
19. 1902, as published in circular No. 6 by the Seismological 
Investigation Committee of the British Association, are 
analysed. One result is to show that the preliminary 
tremors had an arcual velocity of 15-6, whilst the eighth 
phase of motion had only a velocity of 2.1 km. per second. 
These determinations, amongst other things, depend upon 
the time at which this earthquake originated, which Mr. 
Rockstroh, in Guatemala, gives as 2-27 a.m. G.M.T. This, 
it must be observed, apparently depends upon a single observ- 
ation made at some distance from the epicentre. As being 
more probably correct, Mr. Imamura adopts 2.26 a.m. as 
the time of origin. Curiously enough, the determination of 
the exact time at which this particular disturbance took 
place became a matter to be considered by the law. At or 
about the time of the earthquake a certain block of build- 
ings insured against fire, but not, as was stipulated in the 
policy, against fire occasioned by an earthquake, was 
burned down. The owners of the block claimed that the 
destruction was occasioned by the overturning of a lamp 
immediately before the earthquake, and if this were the 
case, the loss naturally fell upon the insurance company. 
The result was that a careful inquiry was instituted to 
determine the time when Quezaltenango and other cities 
were wrecked, and, so far as the writer knows, the time 
given for this occurrence was 2h. 21m. or 2h. 22m., which 
is a time that falls in line with what we know of earthquake 
speeds, and the times at which this earthquake was recorded 
at stations cooperating with the British Association in North 
America and round the Atlantic. If this latter time is fairly 
correct, and we therefore add 4 or 5 minutes to the time 
ordinates given by Mr. Imamura, we see that the times for 
transit of the first phase of earthquake motion are not pro- 
portional to arcual distances. The speed over short paths 
is less than it is over long paths. When we look at phase 
8, which occurs about twenty minutes after the maximum, 
and see the many phases which follow, we do not sce the 
reason why the transit velocity of phase 9, or even of phase 
20, should not have been considered. 
Following the maximum, as an earthquake dies there is 
almost invariably a series of fairly rhythmical impulses 
which gradually grow less in amplitude. These are separ- 
ated by intervals of from two to five minutes, and may 
extend over two or three hours. Inasmuch as_ these 
phenomena may be observed equally well near to an earth- 
quake origin as at a distance from the same, it would be 
unreasonable to suppose that if we were at a short distance 
from a centrum the movements last recorded there had 
been two or three hours longer on their journey than those 
first recorded. In rocky materials waves of small ampli- 
tude may travel more quickly than those of larger ampli- 
tudes ; in an earthquake there may be ripples due to surface 
tension together with the more pronounced compressional, 
distortional, and gravitational waves, all of which may 
originate at practically the same time, but if. we consider 
this to be the case for the followers of the main portion 
of a seismic disturbance, we are led to conclusions apparently 
unacceptable. The fairly uniform time spacing between the 
expiring efforts of an earthquake as recorded at a distance 
from an origin or near to the same, rather than leading 
to the conclusion that paths are arcual, suggests a rhyth- 
mical series of surgings possibly due to interferences or 
reflections at or near a centrum. A collapse takes place 
and a mass is launched upon some substratum. Each has 
its natural period, and when these coincide, it seems possible 
that at approximately equally spaced intervals a more 
vigorous set of waves starts out and adds to the train of 
its predecessors. 
The first response to the primary disturbance may be the 
well-known Uri Kaishi, or return shaking, whilst its 
followers, which die down rapidly in amplitude, like the 
swinging of a damped pendulum, resemble a family series, 
children with children’s children decreasing in vigour, not 
born simultaneously, but successively. 
The general criticisms on the behaviour of the Milne 
horizontal pendulum are made without any reference to the 
