524 
= 98.= 
In the present investigation an attempt was made to determine the velocity of the pressure wave 
near the charge as accurately as possible in crder to obtain the variation of velocity with pressure- 
amplitude. The method finally adopted employed two exactly similar circular p.e. gauges placed broadside— 
on to the charge and fixed on a rigid iron bar (of girder section) at a distance apart usually 5 feet 
(approx.) which could be accurately measured. The pair of gauges were suspended at a known distance 
from the charge on a continuation of the line of the bar passing through the charge. The depth of the 
charge, 184 G.C. and the gauges was kept constant at 15 feet the distances of the two gauges being 15 and 
20 feet respectively from the charge. The gauges were connected in paralle? to the oscillograph recording 
plates, the usual circuits containing the pressure switch being employed to traverse the spot across the 
plate. In order to obtain an accurate measure of the time=interval between the two pressure peaks (due 
to the two gauges) to be recorded, the potential variations in a valve circuit oscillating at a frequency 
of 2000 ~ (or in some cases 1000 ~/second) were applied to one of the deflecting plates of the oscillograph 
Whilst the record was being made. In this manner tne record consists of an oscillation of 2000 ~/ second 
on which the two pressure peaks are superposed. Since the rise of pressure is very sudden, the instant 
of arrival of the explosion wave at each gauge can be read off the record very accurately. The frequency 
of oscillation was standardised at the time of firing the charge by comparison with a tuning form® of 
frequency 1000 ~/second. A steel tape graduated in feet (tenths and hundredths) was employed to measure 
the distance apart of the gauges. {t is considered that the distance apart of the gauges is probably 
correct to 1 part in 1000, The probable error in measurement of the velocity due to errors measurements 
of length of base is not greater than + 0.1%, i.e. to+ 5 feet ina velocity of the order of 5000 feet/ 
second. it must be pointed out at this stage, however, that the probable error of the final result is 
considerably greatat than this due to errors introduced in the actual measurement of the records. Further 
reference to this will be made later. Temperature observations, using a reversing thermometer were made 
on each occasion of firing a charge. The preliminary observations by the methoc indicated above gave a 
value which agreed closely with the ordinary velocity of small amplitude waves, i.e. in accordance with 
the St. Margarets Bay results. This unexpected resu)t led us to suspect a possible error due to the 
movement of the iron girder-bar which carried the gauges, in the direction of the pressure wave, such 
movement being transmitted (at a velocity higher than that in the water) through the bar itself to the 
second gauge. Such a movement would have two effects, (1) to increase the effective length of the base 
line between tne gauges and, (2) to-develop a small pressure on the second gauge before the arrival of 
the pressure pulse through the water. The entire-absence of (2) made it seem improbable that (1) was 
of a serious nature. In order to leave no room for doubt, however, the experiment was repeated with the 
two gauges suspended freely (though at a definite distance apart) by means of spunyarn from short cross 
bars fixed on the girder bar. The values of the velocity were, however, unaffected by the change in the 
method of mounting the gauges, thereby eliminating the possibility of the error mentioned above. 
Further measurements were therefore made with the gauges rigidly clamped to the iron girder-bar. 
The velocity observations may conveniently be divided into two series viz. (1) those made March — 
May 1923 when the sea temperature was about 10°C and, (2) those made in December 1923 the temperature 
then betng 7°C approx. In the first series of observations the time scale was generated by an oscillation 
of 2000 ~/second with high harmonies (see Sheet 8 Records a to d), it was thought possible in this manner 
to obtain ‘t' the time interval between the gauges more accurately on account of the increased 
probability of obtaining sharply defined turning points on the time scale occurring at or near the instants 
of commencement of the pressure rise at the two gauges. In this case it was arranged that the ‘pressure 
deflection’ on the record should be large compared with the A.C. oscillations. The frequency of 
oscillation was adjusted to 2000 ~/second by tuning aurally to a 1000 ~/second fork. It is estimated 
that the error introduced in this manner does not exceed 1 in 250 in frequency, i.e. + 20 feet in velocity. 
In the second series of experiments a large amplitude sine wave of 1000 ~/second was used as the time 
scale, the pressure deflection being kept relatively small (see Sheet 8 Records e, f and g). The 
frequency in this case was adjusted to that of the fork by beats and is considered accurate to 1 in 1000, 
equivalent to + 5 feet in velocity. The cesults of all observations are included in the following table, 
where a Comparison js made with the values of velocity for small-amplitude waves at the same temperatures. 
The distances of the two gauges from the 184 lbs. charges were in all cases 15 feet and 20 feet respectively, 
corresponding to a value of Fog at the mid point of the base-line of 2000 Ibs /sq.in. 
Table weesee 
This fork had previously been standardised by direct comparison with a similar fork having 
an N.P.L. certificate its frequency being 1000.35 at 16°C. 
