rise time <10~7 sec, but the pulse 
height increased by about 10% when 
the voltage was raised from 2.6 to 4.6. 
Of this increase about half took place 
between 2.6 and 2.9 kv before the start 
of the plateau region. The rest oc- 
curred over the entire plateau region; 
the quiescent corona current increases 
with voltage, so that the effective field 
between the wire and the plate increases 
only very slightly. Indeed, depending 
upon the external resistor used, Veg can 
even decrease with increasing V, in cer- 
tain cases, where a fall in the pulse 
heights can be anticipated. 
Quiescent Current 
A corona discharge is set up around 
the anode wire as the applied voltage is 
raised above the counting threshold. 
A bluish glow in the form of a fine 
streak of light is observed all along the 
wire’s length. The visible glow ex- 
tends about one wire diameter toward 
the plate. 
With the source removed we studied 
the variation of corona current with 
applied voltage by placing a microam- 
meter in series with the cathode resistor 
(R = R, + Ro, Fig. 2). The quiescent- 
current characteristic for the counting 
arrangement described earlier is shown 
in Fig. 3. It is seen that the quiescent 
current sets in at the beginning of the 
hump in the counting characteristic. 
Then there is a very small region where 
it increases nonlinearly with increasing 
applied voltage. Beyond about 2.9 kv 
it increases linearly. The start of the 
linear region corresponds clearly to the 
start of the counting plateau. 
Examining the quiescent condition 
on an oscilloscope, we observe corona 
pulses, their number increasing with 
the applied voltage. Near the applied 
voltage corresponding to the hump in 
Fig. 3, corona pulses of comparatively 
large magnitude are observed and 
corona oscillations were loud enough 
to be clearly audible. Probably the 
slight increase in the counting rate at 
the hump is due to these corona oscilla- 
tions. Once the corona discharge has 
extended over the whole length of the 
wire, the counting condition seems to 
stabilize and the observed rate settles 
down to the normal value. 
The quiescent current characteristics 
were also observed for four different 
values of R (4.2, 7.7, 11.5, and 19 
megohms). The results are shown in 
Fig. 4. These curves, obtained for 
different values of R, fit well with the 
40 
ideas outlined earlier. For each of 
these curves, Ver can be calculated 
easily and the previous predictions re- 
garding the variation of Veg with V, 
examined. The plots of Veg vs Va ob- 
tained in this way are given in Fig. 5. 
Clearly the higher curves correspond to 
the case when k(R) < 1/R and Ver 
increases with V,. As expected, the 
rate of increase of Veg with V, gradu- 
ally fallsas Risincreased. The bottom 
curve corresponds to the case where 
k(R) > 1/R (R = 19 megohms), and 
here Ver decreases with increasing V,. 
For R between 11.5 megohms and 19 
megohms, there should also be a critical 
resistance R, for which Ve would re- 
main practically constant (dotted line 
AB) over a large range of Vs. For all 
values of R greater than R., Ver and 
hence the counting rate are expected to 
decrease with the applied voltage, as 
has been actually observed by Connor. 
For all our subsequent measurements 
we have a total quenching resistance 
R = 4.2 megohms. 
Rate vs Distance 
The number-distance curve for an 
alpha source was measured with this 
counter using a collimated old polonium 
source. Figure 6 shows the number- 
distance relation at room temperature. 
The shape of the curve is similar to that 
obseryed with other alpha detectors 
near the end of the range. The con- 
siderable spread of the curve observed 
between the peak and the maximum 
distance at which counts were regis- 
tered (~3.6 cm) is due to our using a 
fairly old polonium source (8). With 
the source nearer to the counter there 
is observed, however, a remarkable fall 
in the counting rate with decreasing 
distance, making the monochromatic 
alpha peak singularly fine. This sud- 
den fall seems to arise from two causes. 
First, the specific ionization of the alpha 
particles is known to fall somewhat for 
distances less than those corresponding 
to the maximum of the Bragg curve. 
This will amount to a reduction in a 
priori sensitivity and a consequent fall 
in the counting rate as the source ap- 
proaches the counter. However, this 
factor does not seem to account for the 
rather steep fall observed. The in- 
crease in the number of oblique rays as 
the source is moved toward the coun- 
ter seems to be the dominant cause, the 
a priori sensitivity being lower for the 
oblique rays than for the normal rays 
for the same working voltage. The 
range, however, of monochromatic 
alpha particles from a freshly prepared 
source can be determined by this coun- 
ter with reasonable accuracy. 
Temperature Dependence 
None of the earlier workers have 
tried to study the temperature depend- 
ence of the spark counter. Connor has 
expressed the view that it is likely to be 
slight, but no exact data on tempera- 
ture effect appear to exist. 
To study the temperature depend- 
ence it is essential to allow for the ap- 
parent change in the range of particles 
caused by the change in the density of 
the surrounding air as its temperature 
is changed. We thought it best to 
study the temperature dependence by 
observing the number-distance curve 
for the polonium source at two different 
temperatures. The first observation 
was at room temperature (20° C). 
Then the counter was placed in a ther- 
mostatically controlled volume where 
the temperature was maintained at 
49 +1°C. The results are shown in 
Fig. 6, where one solid curve corre- 
sponds to room temperature and an- 
other to the higher temperature. The 
dotted curve was drawn by shifting the 
room-temperature curve toward higher 
d values by an amount equal to the 
apparent increase in the range of the 
alpha particles expected at the higher 
temperature. The dotted curve may 
be called the extrapolated number-dis- 
tance curve for the higher temperature. 
It will be seen from the figure that a 
better fit between the shifted curve and 
the measured one might result from a 
simple ‘stretching’ of the Bragg 
curve; that is, by increasing all 
abscissas by the same fraction of them- 
selves so that extrapolated ranges (or 
the maxima) would coincide. We have 
not attempted this slightly more com- 
plicated analysis, but it appears that 
a small temperature effect exists. 
* * * 
We are greatly indebted to Professor D. 8S. 
Kothari for stimulating this work and his 
kind and continued interest during tts prog- 
ress. This work was carried out under the 
financial support of the Department of 
Atomic Energy, Government of India. 
BIBLIOGRAPHY 
1. H. Greinacher, Z. tech. Phys. 16, 165 (1935) 
W. V. Chang, S. Rosenblum, Phys. Rev. 67, 
222 (1945) 
R. D. Connor, Nature 163, 540 (1949) 
R. D. Connor, Proc. Phys. Soc. 64, 30 (1951) 
R. D. Connor, J. Sci. Instr. 29, 12 (1952) 
R. M. Payne, J. Sci. Instr. 26, 321 (1949) 
M. P. Savel, Comptes Rend. 234, 2596 (1952) 
N. K. Saha, K. L. Kaila, Indian J. Phys. 29, 
417 (1955) 
wt 
DRAIWRG 
