ON SEISMOLOGICAL INVESTIGATIONS. AS 
During the first period the records accord fairly well with what might 
be expected, the small moment of C accounting for the small number of 
its records. 
In the second period, when the load on C was increased threefold, we 
find that it gives the largest number of records and the largest number of 
early commencements. The large increase in the records on B may be 
due to the influence of C swinging on the same support. In the last 
period, when B carried a load practically equal to that on C, and had its 
time of swing increased to thirty seconds, we see that it gave the greatest 
number of records and also most frequently was disturbed before the 
others. 
Although these records are not strictly comparable, and for the most 
part only refer to mere thickenings of the photographic trace, they suggest 
that an increase in load and of period in the type instruments would result 
in increased sensibility. 
Ill. Improved Record Receiver for Horizontal Pendulum Seismograph. 
The accompanying illustrations, figs. 1 and 2, show two views of 
a new seismograph recorder. 
The instrument consists of a light brass cylinder, D, 1 metre in circum- 
ference and 160 millimetres wide, mounted upon a steel spindle. One of 
the projecting ends of this spindle has a deep-threaded helix of 6 millimetres 
pitch cut in it ; this being suitably mounted upon roller bearings, advances 
the cylinder 6 millimetres for one turn in four hours, by a gear connection 
with a clock. The bromide paper carried on the cylinder is changed every 
3°5 or 4 days. 
A cylindrical mirror has been introduced to give a greater concentra- 
tion of the light on to the boom-plate. 
For the time record mark upon the bromide paper a shutter actuated 
by an electro-magnet is employed, the light being shut off from seven to ten 
seconds every hour. For this purpose a regulating clock with suitable 
electric contacts is required. An example of records from the new and 
old form of receiver is shown in Plate I. 
The advantages of the new arrangement are :— 
1, Although the paper moves beneath the end of the boom at more 
than four times the rate (250 millimetres per hour) that it does in the 
original receiver, only one-half the quantity of paper is used. This implies 
a large reduction in expense for paper and developer, the latter being 
applied by a brush. 
2. An open diagram is obtained on which wave-periods can be 
measured. 
3. Movements of small amplitude are easily recognised. 
4, Records can be quickly inspected and are easily stored. 
IV. The Origins of Large Earthquakes recorded in 1903 and since 1899. 
The origins of the large earthquakes recorded in 1903 are indicated by 
this Shide register number upon the accompanying map, Plate II. In the 
registers (Circulars 8 and 9) there are 135 entries for this year, whilst on 
the map only sixty-four origins are indicated, which means that there were 
about seventy-one earthquakes the materials relating to which were insuf- 
ficient to enable their origins to be determined. Even with the origins 
which have been determined, the notes of interrogation attached to 
