area of the cathode surface of the sensi- 
tive volume of the counter is 2,100 cm?, 
so that the alpha rate is 1.6 & 10-4/ 
cm?*/min, which agrees closely with the 
lowest figures reported by Bearden (6). 
It is therefore unlikely that this back- 
ground component could be decreased 
appreciably by trying to obtain radio- 
actively purer material for the con- 
struction of the counter. 
2. Cosmic-ray mesons not detected 
by the ring of counters. Calculations 
based on the geometrical relationship 
of the proportional counter and the 
surrounding ring of counters indicate 
that approximately 2 cpm of the back- 
ground counting rate will be due to 
mesons not detected by the ring of 
counters. With more counters in the 
ring this component could be eliminated 
at the expense of a slight increase in 
equipment complexity. 
3. Gamma rays from radioactive 
contaminants in the gamma-ray shield. 
It has been noted (Fig. 7) that the 
background counting rate increases 
with filling pressure for ‘“‘dead’’ CO,.— 
being 10, 12, and 14 cpm for 1, 2, and 3 
atmospheres respectively. This could 
be due either to a radioactive contami- 
nant in the gas or to an increase in 
efficiency of detection of gamma rays. 
The relative efficiency of detection of a 
strong gamma ray source outside the 
gamma ray shield was therefore meas- 
ured as a function of pressure of CO» in 
the counter. An increase of 30%/ 
atmosphere was observed. 
The 2 cpm due to counter contami- 
nants and the 2 cpm due to undetected 
mesons will be independent of pressure. 
The variation of the remainder of the 
background counting rate with pres- 
sure is then 6, 8, and 10 cpm for 1, 2, 
and 3 atmospheres respectively. As 
this agrees with the change of gamma- 
ray efficiency with pressure it may be 
concluded that, at 1 atm filling pres- 
sure, 6 cpm are due to gamma rays 
passing through the counter. 
The gamma-ray shield provides 8 in. 
of iron and 1 in. of mercury as shielding 
around the counter. The background 
counting rate at 1 atm filling pressure 
is reduced from 14 to 10 cpm by the 
addition of the mercury, so a reduction 
of 4 cpm is due to absorption in the 
mercury of y rays from the iron. One 
in. of mercury will have a shielding fac- 
tor of approximately 6, so less than 1 
epm of the background will be due to y 
rays from the iron. 
There are two reasons for believing 
that the gamma rays from radioactive 
contaminants in the mercury cause 
only a small (<5%) contribution to 
the background countingrate. Firstly, 
Kulp and Tryon, in their experiments 
with a mercury shield (7), obtained 
only a small improvement between the 
use of commercial-grade mercury and 
triply distilled mercury. Secondly, the 
400 lb of mercury used in our shield 
was obtained in batches from several 
sources of supply. Batches were com- 
pared in lots of 50 lb for radioactive 
contamination in a low-background 
Geiger-counter arrangement. No dif- 
ference between the various batches 
was detectable. 
4. Gamma rays associated with the 
cosmic-ray flux passing through the 
gamma-ray shield. The background 
counting rate appears to depend on 
barometric pressure, but unfortunately 
statistical variations of the counting 
rate and the small percentage change 
of barometric pressure prevent this cor- 
Contemporary Assay Results 
The CO,-filled-proportional-counter method for radiocarbon dating will help 
analyze the fine structure of the contemporary C1! assay. 
Using only 12 gm of 
carbon, an accuracy of 0.3% in C™ content is possible. 
Already the specific activity of natural radiocarbon in living wood samples 
has been measured as 12.5 + 0.2 dpm/gm. 
This figure is considerably lower 
than the currently accepted 15.3 dpm/gm (1) but agrees with the 12.9 + 0.2 
dpm/gm obtained by others using scintillation counting of organic compounds 
(8). 
mean value. 
Various species of New Zealand woods tested agree within +14 % of the 
The specific activity of natural radiocarbon in the carbonate of shellfish has 
been found to be (6 + 1)% greater than for their flesh (two samplesonly). 
This is in close agreement with the theoretical figure of 6% and lies approxi- 
mately in the middle of the range of O-10% obtained by others (1,9, 10). 
These 
results will be published in greater detail at a later date. 
relation from being established accu- 
rately. The pressure coefficient ob- 
served for mesons (coincidence counts 
from counter) is —2%/em Hg. For 
the total background counting rate 
(anti-coincidence counts from counter) 
it is (—3 + 1)%/cm Hg. Therefore, 
it is reasonable to assume that the re- 
maining 5 cpm of background due to 
gamma rays is associated with cosmic 
rays. 
The amount of radon and thoron 
present in the atmosphere is known to 
be dependent on the barometric pres- 
sure. The possibility of this variation 
affecting the counting rate has been 
eliminated by continuous flushing of 
the space between the proportional 
counter and the mercury shield with 
nitrogen gas. 
If the preceding reasoning is correct, 
a considerable decrease in the back- 
ground counting rate should be ob- 
tained by setting up the equipment 
underground so as to reduce the compo- 
nents (2 and 4) associated with cosmic 
rays. At 100 ft underground, back- 
ground components associated with 
cosmic rays would be reduced suff- 
ciently to be negligible compared to 
shield and counter contaminant back- 
ground. The background counting 
rate of the equipment would then be 
approximately 3 cpm. 
* * * 
The author wishes to acknowledge the col- 
laboration of T. A. Rafter who developed the 
chemical procedures for burning the samples 
and purifying the CO2, and of W. McCabe 
for his careful preparation of the numerous 
samples that have been necessary during de- 
velopmental work. Construction of nearly 
all electronic equipment has been by K. A. 
Bargh whose attention to detail has con- 
tributed to its reliability. This paper is 
published with permission of the Director of 
the Dominion Physical Laboratory. 
BIBLIOGRAPHY 
1. W. F. Libby, ‘‘ Radiocarbon Dating'"’ (Uni- 
versity of Chicago Press, Chicago, IIl., 
1952) 
2. G. J. Fergusson, 
Technol. B36, 90 (1953) 
New Zealand J. Sci. 
38. T. A. Rafter, New Zealand J. of Sci. 
Technol. B36, 64 (1953) 
4. E. C. Anderson, H. Levi, Kgl. Danske 
Videnskab. Selskab., Matfys Medd., 27, No. 
6 (1952) 
5. H. L. DeVries, G. W. Barendsen, Physica 
19, 987 (1953) 
6. J. A. Bearden, Rev. Sct. Instr, 4, 271 (1933) 
7. J. L. Kulp, L. E. Tryon, Rev. Sct. Instr. 
23, 296 (1952) 
8. F. N. Hayes, D.:L. Williams, B. Rogers, 
Phys. Rev, 92, 512 (1953) 
9. M. Blau, E. 8. Deevey, M.8. Gross, Science 
118, 1 (1953) 
10. J. L. Kulp, Atomics 4, 96 (1953) 
11. W. N. English, G. C. Hanna, Can. J. 
Phys. $1, 768 (1953) 
12. A. J. Freedman, E. C. Anderson, Nvu- 
cCLEonNIcs 10, No. 8, 57 (1952) 
13. A. R. Crathorn, Nature 172, 632 (1953) 
14. H. Barker, Nature 172, 631 (1953) 
16. J. R. Arnold, Science 119, 155 (1954) 
195 
