OCTOBEB 1955 MONTHLY WEATHER REVIEW 225

HAS THE AMOUNTOF CARBON DIOXIDE IN THE ATMOSPHERE CHANGED

SIGNIFICANTLY SINCE THE BEGINNING OF THE TWENTIETH CENTURY?
GILES SLOCUM
U. S. Weather Bureau, Washington, D. C.
manuscript received November 4, 1954; revised October 7,19551

ABSTRACT
The search for causes of the rising temperatures in some geographic areas during the twentieth century has
directed interest toward the amount of atmosphericcarbon dioxide (GOa). If the carbon dioxide added by the
combustion of fossil fuels remains as a net increase, any temperature-changing effects of its presence as a minor
constituent of the atmosphere should be cumulatively operative as the amountincreases.
In thispaper, the physical knowledge of atmospheric GO2 is examined and theavailable nineteenth and twentieth
century observations of the atmospheric COa concentration are summarized to ascertain the extent t o which they
corroborate claims that the amount of atmospheric COa has increased since the nineteenth century. In the light
of the uncertainty of both physical knowledge and of statistical analysis, it is concluded that the question of a trend
in atmospheric GO, concentration remains an open subject.

1.INTRODUCTION In 1938,Callendar [6] suggested that the combustion of

the fossil fuels, such as coal, lignite, petroleum, etc. may be
This report examines the physical knowledge of atmos- causing such an increase. At that time, according to his
pheric CO, and summarizes the available nineteenth and estimate, 4.3 x lo9 tons of COz per annumwere being added
twentieth century observations of the atmospheric to the atmosphere in this way. He gave the total added
content ofCO,, to ascertain how fartheycorroborate between 1887 and 1937, after allowing for an accelerated
claims that the amount of carbon dioxide in &heatmos- rate of burning as time went on, as about 1.5 x 10” tons.
phere has increased since the nineteenth century. Charts So large an amount addedso quickly would, he suggested,
and tables are included, showing the locations, periods be absorbed into the earth’s watersa t a much slower rate.
of record, numbers of measurements, and the ranges of Assuming that other natural processes, such as the bio-
values of the observed CO, concentration. logical exchange, be in balance, the result would be an
Theamount of carbon dioxide in the atmosphereis increase of atmospheric CO, with time. He estimates that
actually little more than a trace [lS, 231. That dissolved 2000 to 5000 years will be required before we may expect
in the waters, or combined as carbonates, etc.,in the crust the atmospheric content toreach equilibrium with therate
of the earth, is much greater. Goldschmidt [15] presents of oceanic abs0rption.l Since the acceleration in the rate
a table, showing the location of the earth’s COaand poten- of industrial combustion may not cease for some time [17],
tialCO,. It is the source for table 1. Fromthistable, the consequently increasing CO, in the atmosphere would
it appears that about 0.005 percent of the earth’s crustal increase the absorption of outgoing radiation, and the
carbon is in the atmosphere as CO,. surface layer of the atmosphere would become warmer.
Important climatic effects are attributed to this small In support of this view, Callendar [7] selected from the
percentage of carbon dioxide in the air, and, according to published records of determinations of CO,, made between
Callendar[6,7, 91 and Plass [24], a significant increase 1867 and 1935, those which he considered the most
in the concentration of COzwould noticeably raise the accurate. On the evidence of these records, he found that
surface temperature of the earth because of the “green- the CO, had apparently increased since 1900 by about 6
house effect.” percent.
Figure 1, after Glueckauf[14], shows the increase in
content, according tocallendar’s selected data. To Glueck-
auf’s plotted points have been added more recent data,
including Callendar’s 1949 computed value [9], and other
1 Suess [a]however,
, estimates that the “average lifetime of a Cot molecule in the
atmosphere before it is dissolved in thesea willlie between 20 and 50 yeas.”
226 MON!CHLY WEATHER
REVIEW OOrOBICB 19511

450 Hutchinson [19] has stated that, “There can be little

440 - K doubt that during the first half of the twentieth century
-E 430 - a
the mean CO, content of the air in north temperate
latitudes has increased.” Callendar hasthus had a
-:420 - number of supporters in whole or in part.
410 - Independent opinion has not,however, been unanimous
W

$400
- D
e
in support.At the time Callendar delivered his 1938
paper [6], Mr. J. H. Coste suggested that the accepted
390 -
CO, content hadat the turn of the centurybeen considered
W
380 - to be about 0.04 percent, and not the 0.029 percent indi-
370 - cated by the measurements Callendar cited. Mr. Coste
z - then asked, since the value, 0.04 percent, is a higher per-
,360 centage than the average value of about 0.032 percent
350 -
0 0
0 Callendar found for the 1937 CO, content, can we be sure
0
,340
- 0
that there has been any net increase a t all in the per-
0
- centage of CO, in the atmosphere?
F 330
a
-
320 2. PHYSICAL EVIDENCE
E 310 -
0

There are processes which may deplete the increased

300 concentration of C02produced by combustion, and others
290 ““-7
-+:
’+
+ J which may be more important than combustion of fossil
fuel in increasing the concentration, at least temporarily.
1 1 1 1 1 1 1 1 1 1 1 1 1 1
18 6D 1950
1900 1920 2000 Callendar does not consider them relatively important,
DATE
but it seems logically tenable to suppose that a relatively
FIGURE1.-Trend in C02 content of the earth’s atmosphere slight increase in the rate of biologic absorption of C02
(selected data). Adapted fromGlueckauf[15]. To Glueokauf’s mightnearly, or even more than, compensate for any
values (+) have been added those of Duerst (D), 500 observations
[la]; Kreutz (K) 25,000 observations [21];Callendar’s latest esti-
increase in its evolution fromother sources, such as
mate, 1949 (C) [9]; and the means derived in the present paper (S). industrial combustion.
Moreover, Hutchinson [19] suggests that with the ex-
spansion of industry through the past century, agriculture
recent estimates mentioned at various points inthe present also expanded, and that there is a f a r greater opportunity
paper. for loss of respiratory CO, from soil in arable land thanin
Buch [5], used his own observations (1932-1935) taken forest land. He therefore doubts the validity of Callen-
in scatteredhigh latitudes in the North Atlantic ocean and dar’s explanation of the source of an increase in the
its estuaries, as representative of the concentration of amount of CO,.
atmospheric carbon dioxide at that time, and the same Dingle [ll], in discussing Callendar’s fiDdings,points
sources as those selected by Callendar to represent the out the complexities of determining the amount of COz in
content in the late nineteenth and early twentieth cen- the world’s atmosphere as a whole. Measurements at one
turies. Then, comparing these latter data with his own or a few localities over a limited period of time are in-
newly observed values, he came to substantially the same adequate, since the concentration of carbon dioxide varies
conclusion as did CallendLar. in air masses of differing trajectories. He suggests that
Recently there havebeen independent studieswhich are increases in the concentration of observed CO, might be
at least consistent with Callendar’s. Among these recent due to changes in the general atmospheric circulation
studies may bementioned those of Brown [3], who, in 1952, rather thannecessarily mainly dueto a worldwide increase
determined the C12/C13ratiointreering samples. He in C02concentration. He holds this to be a more attrac-
found evidence that this ratio is, on the average, greater in tive physical hypothesis to explain any increase in the
the younger samples than in the older. This indicates, observed CO, value than is Callendar’s thesis that the
he suggests, that “carbon dioxide in the air has been higher temperatures are due to an increasing concentra-
diluted in recent years by carbon dioxide from industrial tion in the atmosphere as a whole.
sources,” and [4] that the total “Carbon dioxide content Since this paper was initially prepared for publication,
of the atmosphere may be increasing, or at least may not two studies by Suess [25, 261 have become available. He
be in equilibrium with the oceans.” Dingle [ll],by phys- cites the fact that fossil carbon does not contain appreci-
ical reasoning, arrives at the conclusion that the C02 able C1*,and presents evidence that the proportion of C“
content of the atmosphere a t present exceeds 0.03 percent, contained in tree rings has decreased slightly since the
which is in excess of the proportion Callendar estimates nineteenth century. I n its place is a greater proportion
for the nineteenth century. of C12. The decrease in the ratio is, however, greatest
OCrOBEB 1955 MONTHLY WEATHER REVIEW 227
TABLES.-Physical and chemical processes adding or extracting mas- to the beginning of the middle third of the twentieth.
sive quantities of COa to or from the atmosphere (after Hutchinson
1191) Their comparisons are, however, based on a narrow selec-
c tion of values from a much larger body of data, scattered
Qross emission
(x 101%gramslyear)
Gross 8bSOrptiOR through the scientific literature of the past century. It
(x 101: grams/year)
Process or category may be granted that the data they used are probably
1
At leastmore Butnot
than 7
At leastmore than
quite accurate averages for the time, place, meteorological
conditions, etc., of observation. The question remains,
however, are all the measurements which they did not
use, inaccurate?
Buch who followedCallendar, accepts without challenge
Callendar's selection of data, and merely adds his own
observational material. Hutchinson [19] bases his state-
ment that CO, has increased on Callendar's results,
~~ ~~~~ ~

'Adding to the maximum photosynthesis estimate,8 half of the highest estimate of

although he limits his corroboration to the case of the air
production by industrial consumptionof fuels. in the North Temperate zone. Brown and Dingle alone
offer any new evidence on theamount of C02 in the
near concentrations of industry, and is much less in the atmosphere, and their evidence, if it be not negative, is
case
of a tree which grew in Alaska thaninthe not necessarily confirmatory.
casewof trees near dense population centers. Suess con- Since Callendar, by basing his hypotheses on statistical
cludes that ". . . the world-wide contamination of the data, has tacitly invoked the laws of statistical evidence,
earth's atmosphere with artificial CO, probably amounts it is fitting to examine the validity of his procedure, that
to less than one percent." of using only the datahe believed to beof the best quality
Brown's [3,4]evidence from tree rings of the dilution of available, rejecting the rest.
CO, has been suggested as showing that there has been an The mathematics of statistics, and the experience of
increase in the amount of GO2. This does not necessarily statisticiansbothindicate, as a general principle, that
follow. It can be shown that dilution would be the ex- arbitrary rejection of data, without specific knowledge of
pected result of any 1s.rgereplacement of natural by fossil their unreliability orunapplicability, is questionable.
carbondioxide. Such dilution would occur whether Although the purpose of such a procedure may be to re-
the totalamountin
the atmosphere be gradually move an observational or sampling bias that is known to
increasing, remaining approximately stationary,or de- be present,selection of the data tobe used will often intro-
creasing. Indeed, the replacement would be most pro- duce agreater source of error than that which it was
nounced, and, therefore, most detectable inan atmosphere intended to remove.
with decreasing CO, concentration. Thus f a r only At best, the omission of part of the datais notas neces-
Brown's abstracts are available. I n them, he states his sary or as helpful as might appear at first thought, since
conclusions in strictly qualitative terms. Demonstration it can be shown that when the means of two sets of data
that the trend is either up or down awaits a quantitative are compared, the presence of a given average bias in
discussion of his findings and of whatever compensating each set will not affect the difference nor the standard
forces in either direction enter as complications. What error of this difference, except as an added contribution
Brownseems to have confirmed thus far is the already to the variance of the sample. If, however, some of the
established fact that great quantities of fossil carbon have data be selected to the exclusion of the rest, for the pur-
been turned into atmospheric GOz. pose, perhaps, of reducing the magnitude of the residual
Thus, students of the subject differ. That is, physical variance, due to crudity in some of the measurements,
reasoning has not as yetshown that CO, is necessarily in- then, in addition to any unintentionalbias that might be
creasing as a result of the addition of combustion gases. introduced in the comparison of the means, there might
There remains thestatisticalapproach,that used by also result an underestimate of the standard error of the
Callendar, Thecurrent knowledge from a quantitative difference, due to the mistaken rejection of those of the
standpoint is summarized in table 2. extreme values which actually belong to the distribution.
The result may be an entirely spurious accuracy in the
3. STATISTICAL EVIDENCE means, which leads to unjustified conclusions.
I n the light of these considerations, a reexamination of
With a dependence on statistical evidence, the mathe- the entire body of available measurements of the relative
matically established statistical criteria for significance of proportion of CO, in theatmosphere may have some value.
results must rule the degree of confidence with which con- Fortunately, Effenberger [13] has compiled what seems to
clusions may be drawn from the original data. Callendar's be a fairly complete list of the published observations up
and Buch's averages appear, as presented in figure 1, to to 1940. He hasindicated the setsof determinations used
show an increase in COz from the late nineteenth century by Callendar [8]. More recently, the American Meteoro-
228 REVIEW WEATHER
MONTHLY OC~OBEB
1965

Symbol Period of Location No. C% content of atmosphere Symbol Period of Location No. C% content of atmosphere
Observatians of (parts per million) Observations of (parts per million)
as . Min.
Mean Max. ObS. Min. Mean b5x.
370620 o 1880-1889 France 64 287
* France 410 " "
1816-1827 "

1844 Prussia 210 420 400 0 1880-1889 France " " 290 "

--
"

v 1866-1879 45"N, 30"W 4 " 270 0 1890-1898 Ireland 64 " 280 "

0 1866-1879 Greenland 3 480 550 640 0 1890-1898 England 92 " 294 "

V 1866-1879 England I 26 "

310 -- 0 1890-1898 Austria "
200550 380
303
Q 1866-1879 England 53 210 296 410 V 1904-1919 us. 645 "

480
"

700
V 1866-1879 France 350 92 802 270 V 1904-1919 Greenland 59 "

v
P
1866-1879
1866-1879
France
Germany
89
1,034
292
"

270
291 "

350
1920-1929 France
I ;:1 180 1 329
1 590
V 1866-1879 Germany 330341 210 420 1930-1939 US. " " "

0 1930-1939 45"N, 31"W 28 152 318 568

V
v
1866-1879 Austria
1866-1879 Switzerland "
295 300410
210420
340
330 0 1930-1939 45"N, 29"W 53 -- 320 "

V 1866-1879 France "

410 0 1930-1939 Scotland 152 " 324 "

. 350
-
" "

V 1866-1879 France " " 300 " 0 1930-1939 England " 310
o 1880-1889 Belgium 525 350
260 294 1930-1939 Finland 95 " 321 "

Q 1880-1889 Baltic
E. 266 " 300 350.. 0 1930-1939 Germany 25,000 " 438.5 "

0 1880-1889 France 1,000 240 292 360 0 1930-1939 Italy 790 400500 240

FIGUBE 2.-Geoma~hic distribution of selected CO, content measurements that have been made in the Northern Hemisphere and the
data available foreach location.

logical Society [l]has published a "Bibliography on Car- data for the observations which were made earlier than
bon Dioxide in the Atmosphere." From this source and the first of those selected by Callendar. One set of obser-
elsewhere, references have been found and some addi- vations was made in 1816, theotherin 1844. The ob-
tional, more recent, data have been compiled. The geo- servations shown by open symbols represent the period
graphic distribution of these observations of CO, atmos- 1866-1901 from which Callendar selected his values for
pheric content and other data listed by Effenberger and thelatterpart of the nineteenthcentury.During this
the othersources used in these summary tables, are shown period, the consumption of fossil fuel had not become M
in figure 2. This figure shows the means and the highest great as it was between 1901 and 1930, a period represented
and the lowest values of the atmospheric concentration roughly by the bulk of the data charted as solid symbols.
found during each of the observational programs repre- Where a closed symbol appears inside an open symbol,
sented. Where available, the numbers of observations, observations were made during both of the latter periods.
on which the means were based, are given. In table 3, the mean values shown in figure 2 are re-
The asterisk and boxed plus sign in figure 2 show the classified to show the values used by Callendar for the
OCTOBEB 1955 MONTHLY WEATHER REvlEW 229
.O?
TABLE3.-Mean COa values, in parts per million. Determined by
A
observation i n the period 1816 to 1940. Compares COa content for
.Od
observations used by Callendar with that f o r observations not used b y
him.[Arrangement i s i n order of magnatude.](after Effenberger
[13]except as noted) 0.02

Nineteenth Century Twent.ieth Century

OB
e
1 1 1 1 1
. 1 1 1 1 1 1 1
(1816 to 1901)
-
(1904 t o 1935) (1936 t o 1940) 1 f W

T Means not usedby Callendar 0

"
N
05 -
.
-
1
.04
Dorm- Means I +
nenteda used b5 z
somalot
data
Call-
endar inacrm-
Published DOCU-
Rejected too late mented
tobe anomalous
Y
V
w
.03 "--
etc. available data [B]
0 02

287 270 460 303 6 480 .05

289 290 b660 318
291 300 320
292 300 311 04
292 320 324
282 3330
294 330 I 03
204 340
295 350
286 380 0.02
400 1810 20 30 40 50 60 70 BO 90 1900 10 20 30 40
410 DATE
410
470 FIGURE
3.-Proportional amounts of atmospheric CO,, in parts per
650
Means 292 363 317 303 419 million, measured in (A) the British Isles, (B) France and Switzer-
~ land, and (C) Germany, Denmark, East Baltic States, and Awtria-
Combinedmeans 335 334 Hungary.Length of line denotes length of time observational
.
b
Observed on rainy days
Observed on dags with snow
program continued. Dots are used for periods less than one year.
Theline segment in (C), at 1939-40 and showing 438.5 ppm,is
e
d Not listed &
Erogh's 17' 2 data
Effenberger. Cited by Callendar, butnot included in his means
based on more observations than all other points and line segments
on all three charts combined.

nineteenth and for thetwentieth centuries, andthree

categories of observations not used by him. Each value,
as in figures 1 and 2, is a mean of a group of observations, are so many possible sources of variability, that there is
varying from 3 to about 25,000. no basis for any claim, based on these data, that the CO,
Figure 3 shows the majority of these determinations content of the atmosphere has remained anywhere near
grouped in another way. Here, the means of the sets of constant. Similarly, there is inadequate basis for a claim
observations, for each of the principal regions where that Glueckauf's trend line approximatesthe recent trend
measurementswere made,are shown for: British Isles of the actual carbon dioxide content of the earth as a
(fig. 3A); Franceand Switzerland (fig. 3B); Central whole.
Europe, including Germany, Austria-Hungary, the eastern The means that Callendar rejected from the nineteenth
Baltic States, and Denmark (fig. 3C). The length of line, century records are, inthe main,indicative of higher
representing each mean, shows the length of time the values than those he accepted. He points out that the
observational program continued. It can be seen from accuracy of observations improved as time went on, and
this figure that the majority of programs were of short that early techniques tended to give too high values.
duration, and from thetable accompanying figure 2 Statistically speaking, the data in table 3 could well be
that in some cases only a few observations were made. drawn from a population having these properties.
Reference to the three charts in figure 3 does not reveal The three values for the twentieth century, however,
any significant trend in COacontent, such as is so clearly which Callendar rejected average lower than those he
shown in figure 1. Indeed, after excluding values which accepted. This does not demonstrate that his choice
the observers themselves have designated as non-repre- was bad, but the fact that he considers so many nineteenth
sentative, but not anyof the others, then the mean value century values to beoverestimates and two twentieth
for the nineteenth century is 335, and for the &st third century values tobeunderestimates raises a question
of the twentieth century 334 parts per million. Such a about his method of selection.
close approach to identity of values for the two periods is, Since techniques have been improving, the latest
of course, an accident. Referring tothetexts of the observations should be the mostaccurate.Duerst [12]
papers from which Effenberger made his tabulations, it and Ereutz [21] found values of 400 and 438.5 parts per
appears that there has been wide variability in the means million, respectively, from observations made in 1936
found for differing geographical regions, on land and on and 1939. Duerst bases his mean on 500 observations,
sea, and from one synoptic weather condition to another. a reasonably large number, if his techniques are correct.
The data-gathering programs were conducted by mutually Kreutz made about25,000 observations. This is more than
independent observers, using differing techniques. There were made in all other herein listed observing programs
230 MONTHLY WEATHER REVIEW OCTOBEB 10%

combined. He expressesconfidence inthe accuracy of its cause ascertained, it will then be valid to base physical
his measurements and of his computed mean values. explanations of atmospheric events on the assumption
Admittedly Duerst’s and Ereutz’ values may be more that COa is increasing. Meanwhile, Callendar’s interest-
representative of the atmospheric concentration of CO, a t ing extrapolations (through the 22d century) of the effects
the time and place of observation than of the earth, or of burning up of the world’s fuel, stimulate the interestof
even the Northern Hemisphere, as a whole. By the same the speculatively minded.
token, however, might not some of the high nineteenth
century values and low twentieth century values be as ACKNOWLEDGMENTS
representative as those Callendar accepted? The present review of scientific progress in this field
Callendar’s presentation of his 1938 paper on the subject was prepared under the direction of Dr. Harry Wexler.
of increasing C 0 2 occurred just after a succession of five To him, andtoDr. Sigmund Fritz, acknowledgements
warm years in western Europe. Since then, this positive are due for their suggestions. Similar debts are owed to
anomaly has been persistent in some densely populated Mr. Glenn Brier, Mr.Isadore Enger, andDr. Lewis
districts in the United States and western Europe; on the Kaplan.
other hand, temperatures have been lower, rather than
higher, in recentdecades, than they were in the nineteenth REFERENCES
century in some Southern Hemisphere regions [lo], and
elsewhere. Can we be entirely sure that the earth as a 1. American Meteorological Society (Malcolm Rigby,
whole has warmed up enough to require an increase in Editor), “Bibliography on Carbon Dioxide in the
C02in the air toexplain it? Atmosphere,” Meteorological Abstracts and Bibliog-
At any rate,it is apparent that, if we use the statistical raphy, vol. 3, No. 2, pp. 137-170.
approach, different degrees of selectivity in determining 2. Bonner, “PhotosyntheticCarbon Dioxide Fixation
which data to include areproductive of differing hal Turnover,” Paper
submitted at California In-
results. stitute of Technology Conference on the Bio-
4. CONCLUSIONS chemistry of Carbon, May 1955.
3. H. S. Brown, “Variations in Trees: Observed Frac-
Is the C02 increasing? Much seems to depend on the tionations,” California Institute of Technology
objectivity of Callendar’s decisions as to which data to Conference, May 1955.
keep. 4. H. S. Brown, “Has the Carbon Dioxide Content of
In thelight of the uncertainty of both physical knowl- the Atmosphere been Changing?”, California In-
edge and of statistical analysis in determining whether stitute of Technology Conference, May 1955.
the relative proportion of carbon dioxide in the air is in- 5. E. Buch, “Der Eohlendioxydgehalt derLuft als
creasing significantly, remaining almost constant, or even Indikator
der Meteorologischen Luftqualitat,”
decreasing slightly, the h a 1 word cannot as yet be con- Geophysica, vol. 3, 1948, pp. 63-79.
sidered to have been said. Instead, the subject remains 6. G. S. Callendar, “The Artificial Production of Carbon
open, either until another chemist critically evaluates the Dioxide and Its Influence on Temperature,”
accuracy of the existing data, or else until more and Quarterly Journal of the Royal Meteorologied
better-organized data are available. Society, v01. 64, NO.275, April 1938, pp. 223-240.
All this does not refute Callendar’s thesis. The avail- 7. G. S. Callendar, “The Composition of the Atmosphere
able data merely fail to conikm it. The positive evidence through the Ages,” The Meteorological Magazine,
that the C02 has increased is inconclusive, but seems vol. 74, No. 878, March 1939, pp. 33-39.
strong enough to reward furtherstudy,andthetime 8. G. S. Callendar, “Variations of the Amount of Carbon
seems ripe for new research. Dioxide in Different Air Currents,” Quarterly
It may be hoped that the collection of standardimd Journal of the Royal Meteorological Society, vol. 66,
measurements of C 0 2 can be made a part of the 1957-58 No. 287, October 1940, pp. 395-400.
International Geophysical Year program. Once a de- 9. G. S. Callendar, “CanCarbon Dioxide Influence
pendable set of observational data has been assembled, Climate?” Weather, vol. 4, No. 10, October 1949,
the evidence of the old observations can perhaps be re- pp. 310-314.
evaluated. If such new reevaluation proves impracticable, 10. E. L. Deacon, “Climatic Change in Australia since
even then a reliable set of new worldwide observations can 1880,” Australian Journal of Physics, vol. 6, No.
serve as a basis for comparison in future years. 2, June 1953, pp. 209-218.
In summary, the data, at present available, are inade- 11. A. N. Dingle, “TheCarbon Dioxide Exchange be-
quate as they now stand toprove or disprove a statistically tween theNorthAtlantic Ocean andthe At-
significant trend in C 0 2 concentration in the atmosphere. mosphere,” Tellus, vol. 6, No. 4, Nov. 1954, pp.
If and when an upward trend hasbeen demonstrated, and 342-350.
OOTOBEB 19E MONTHLY WEATmR REVIEW 231
12.2U. Duerst, “Neue Forschungen uber Verteilung und 19. G. E. Hutchinson,“The Biochemistry of the Ter-
Analytische Bestimmung der wichtigsten Luftgase restrial Atmosphere,” chapter 8 of The Earth as a
als Grundlage fur deren hygienische und tier- Planet, vol. 11, G. P. Kuiper, Editor, 1954, pp.
zuchterische Wertung,” Schweizer Archiv fiir Tier- 375-392.
heilkunde, vol. 81, No. 7/8, August 1939, pp. 20. K. Kalle, “Die Stoffhaushalt des Meeres,” Problem
305-3 17. der kosmischen Physik, vol. XXIII, 1945, 263 pp.
13. E. Effenberger, “Messmethoden zur Bestimmung 21. W. Kreutz,“ Kohlensaure Gehalt der unteren Luft-
des C02-Gehaltes derAtmosphare und die Be- schichten in Abhangigkeit von Witterungsfaktoren,”
deutungderartiger Messungen fiir die Biomete- Angewandte Botanik, vol. 2, 1941, pp. 89-117,
orologie und Meteorologie. Zweite Ted: Ergeb- 22. August Erogh,“The Abnormal Carbon Dioxide
nisse der bisheren C02-Messungen,” A n d e n der Percentage in the Air in Greenland, and the
Meteorologie, Vierte Jahrgang, Heft 10 bis 12, General Relations between Atmospheric and
1951, pp. 417-427. Oceanic Carbonic Acid,” Meddelelser omGrphtland,
14. E. Glueckauf, “The Composition of Atmospheric Air”, vol. 26,1904, pp. 407435.
Compendium of Meteorology, American Meteoro- 23. G. P. Kuiper,Editor, TheAtmosphere of the Earth
logical Society, Boston, 1951, pp. 3-10. and Planets. 2d Edition,University of Chicago
15. V. M. Goldschmidt, “Grundlagen der Quantitativen Press, Chicago 1952, pp. 250-260.
Geochemie,” FortschrittederMineralogie, Kristal- 24, G. N. Plass, “The Carbon Dioxide Theory of Climatic
lographie, und Petrographie, vol. 17, 1934, pp. 112- Change,” Read before the American Geophysical
156. Union, 34th Annual meeting, Washington, D. C.,
16. V. M. Goldschmidt, “Drei Vortrage uber Geochemie,” May 4, 1953. (Available only in Program Ab-
GeologiskaForeningen i Stockholm, vol. 56,1934, stract.)
pp. 385-427. 25. Hans E. Suess, “Radiocarbon in Modern Wood,”
17. M. E. Hubbert, “Energy from Fossil Fuels,” Annual Science, vol. 122, No. 3166, Sept. 2, 1955, pp.
Report of the Smithsonian Institution, 1950, Wash- 415-417.
ington, D. C. 1951, pp. 255-272. 26. Hans E. Suess, Natural Radiocarbon and the Rate of
18. W. J. Humphreys, Physics of the Air, Third Edition, Exchange of Carbon Dioxide between the At-
McGraw-Hill Book Company, 1940, page 81. mosphere and theSea (Unpublished manuscript).