Atmospheric Hydrocarbon and Carbon Monoxide Measurements at Point Barrow, Alaska Leonard A. Cavanagh, Conrad F. Schadt, and Elmer Robinson Stanford Research Institute, Menlo Park, Calif. 94025
The concentrations of low molecular weight hydrocarbons and carbon monoxide in uncontaminated Arctic air masses have been analyzed at Point Barrow, Alaska. Using gas chromatographic techniques, a variety of organic compounds were regularly observed in these air masses, and the following average concentrations were obtained : methane, 1.6 p.p.m.; butane, 0.06 p.p.b.; acetone, 1.0 p.p.b.; and nbutanol, 190 p.p.b. Carbon monoxide concentrations averaged about 90 p.p.b. Special investigations confirmed the unexpected presence of n-butanol in these samples. Sources of these organics are postulated. H
T
he presence of a variety of gaseous compounds in the remote or uncontaminated ambient atmosphere can be assumed because of the variety of natural processes that exist to introduce materials into the atmosphere. However, at present there are few data available on amounts of specific minor atmospheric components, except in polluted urban areas. The research discussed here was a trace gas measurement program to determine atmospheric concentration patterns of low molecular weight organic gases and carbon monoxide (CO). These studies were conducted a t the U. S. Navy Arctic Research Laboratory (NARL), Point Barrow, Alaska, between Aug. 24, 1967, and Sept. l l , 1967 (Kelley, 1967). Thus these measurements are indicative of conditions in the North American Arctic during the late summer. The Point Barrow area is well-suited for the study of the clean o r ambient atmosphere because the few sparsely populated settlements can be easily avoided and there are no other local sources of contaminants. In addition, excellent technical support facilities were available at NARL, Point Barrow. The program was planned to coincide with the latter part of the growing season of tundra flora and with the usual annual minimum concentrations of CO, in the atmosphere. This was done in the hope that, over a longer period of study, possible correlations of CO, concentration with processes in the biosphere could be obtained. Although the present single sampling period was too short to permit any such correlations to be developed, it is hoped that future work along these lines can be conducted and explanations developed for the observed patterns. Sampling Locutions This program called for the establishment of an analytical field installation. Although this created some troublesome logistic problems, it avoided the many problems associated
with sample transport and the verification that no changes have occurred in transit. I t is also possible to get many more data points from field instruments. The field installation was a small wanigan, 1 mile south of the N A R L complex. It was adjacent to the carbon dioxide (CO,) monitoring installation of the University of Washington, which has been in operation for a number of years (Kelley, 1964). These COz studies have established the fact that the area is generally unaffected by local installations. Figure 1 is a map of the local Point Barrow area and shows the locations of the various significant installations. The wanigan was heated by electricity so that there would be no contamination from gas combustion products. Rarely was there any vehicle travel across the open tundra except for personnel moving between N A R L and the two air sampling wanigans. In addition to the gaseous concentration measurements, periodic measurements were made of condensation nuclei concentrations. These were not only interesting in themselves, but they were also an excellent indication of those few cases when one of the local installations was up-wind of the sampling site. Although most air samples for organic analysis were obtained a t the laboratory site, some air samples were also collected in glass sample bottles during air trips to Meade River and Peard Bay. These sites are 50 miles south and 50 miles southwest of Point Barrow, respectively. The samples obtained during these trips were taken back to the wanigan laboratory and analyzed as soon as was practical to avoid any possible changes with storage. Analytical Procedures
Methane Analysis. The methane concentration was determined by a hydrogen flame detector using a continuous sample air flow. The procedure, shown schematically in Figure 2, was similar to that described by Altshuller et a/. (1966). One channel of a dual-channel Varian Aerograph Model 204B was used as the methane analyzer. The sample air was drawn through a port located about 12 feet above the ground and was pumped into the analyzer by a diaphragm pump using check valves and a diaphragm fabricated from Teflon. This Teflon-metal construction reduced the possibility of contamination of the sample air by the pump. Before the sample was introduced into the flame detector, the air passed through a bed of activated charcoal maintained a t 30" C. This bed of charcoal removed all organics except methane. The methane analyzer at Point Barrow was periodically calibrated by substituting compressed air with a known methane content for outside air. The methane content of the calibration air was determined at the SRI laboratory in Menlo Park before the field program. Volume 3, Number 3, March 1969
251
Organic Analysis. The concentrations of low molecular weight hqdrocarbons were determined by gas chromatography. 7 he second channel of the dual-channel Varian Aerograph Model 204B, in conjunction with a special inlet system, was used for these analqses. The gas chromatograph was modified to permit either a 6-foot 2 0 z Carbowax 20M or a 6-foot Porapak Q column to be Lalved in tandem with the organic analysis detector.
For the low concentrations encountered it was necessary to concentrate the organic components from the atmosphere in the inlet system. This was done using a cryogenically cooled freeze-out trap prior to the separation column. In practice the inlet system contains two injection ports and their associated cryogenic freeze-out traps. One freeze-out trap is packed with Carbowax 20M and the other with Porapak Q. The Carbowax 20M separation column is used to determine
+ 0
5
10
MILES
Figure 1. hlap of Point Barrow area 252 Environmental Science & Technolog?
AIR SAMPLE
I
1 AIR PUMP
I T H E R M OSTATE D OVEN
I
1
HYDROGEN FLAME DETECTOR
1
Figure 2. Methane detection procedure
the C2 to Clo organics. The Porapak Q column was designed to provide a separate means of measuring the methane content of the atmosphere and to allow increased resolution and accuracy in the determination of C2to C4alkanes and alkenes. The Porapak Q column did not prove to be feasible at Point Barrow, because the very low concentrations of C? to C4 hydrocarbons were obscured by tailing of the large methane peak. Therefore, the organic determinations were accomplished primarily using the Carbowax 20M separation column and concentration trap. Good separation of materials was obtained except for ethane and ethylene, which had to be treated as a composite compound. The air samples were obtained by gas-tight syringes from a windward location outside the wanigan and about 1 meter above the ground. A 200-ml. sample of air was used for each analysis. Four fillings of a 50-ml. syringe taken over a period of about 2 minutes were injected directly into the special inlet system of the gas chromatograph. This relatively large sample was needed because most organic contaminants were present at concentration of 1 p.p.b. or less. Helium was used as the carrier gas for the gas chromatograph at a flow rate of 22 ml. per minute. The gas chromatographic column was maintained at 78” to 82” C. Liquid oxygen was used as the cryogenic fluid. Although liquid oxygen could not be conveniently shipped to Point Barrow, it was condensed at the site as required from compressed oxygen cooled by liquid nitrogen. Thermal release of the cryogenic trap was accomplished by water at 95” C. Separate hydrogen supplies were used for the methane analyzer and the gas chromatograph. Detector sensitivity was determined by periodic calibration using acetone vapor in equilibrium with liquid acetone at a known temperature. Detection sensitivities for other organics relative to acetone sensitivity were also determined. Retention time calibration was accomplished using vapor in equilibrium with a standard mixture containing pentane, acetone, methane, benzene, and toluene. Although only 1 pl. of calibration standard was required, this quantity far exceeded the concentration of most organic components of the atmosphere. These relatively concentrated calibration standards did not completely elute from the cryogenic freeze-out trap during the initial thermal release. During succeeding heating and cooling cycles, minute quantities of the calibration standard were eluted. Therefore, prior to atmospheric analysis, 200 ml. of helium carrier gas was injected into the gas chromatograph to determine the concentration of helium im-
purities and the magnitude of the cryogenic trap “memory.” This background, if any, was subtracted from the atmospheric sample. The air samples taken at areas remote from the laboratory were collected in 1-liter borosilicate glass bottles with Teflon stopcocks and were returned to the laboratory at N A R L for analysis. A syringe was used to remove the air from the flask for analysis. Carbon Monoxide Analysis. The continuous measurement of C O in clean atmospheres is feasible because of the sensitivity of the C O detection instrument which was developed by Robbins et ctl. (1968) and has been in use for several years at SRI in studies of clean atmospheres. This carbon monoxide analyzer uses a rapid quantitative reaction between CO and hot mercuric oxide (HgO). The chemically produced mercury vapor is then measured by standard ultraviolet absorption spectroscopy using the strong 2537 A line. The HgO cell design provides for very low background mercury vapor from thermal dissociation. Therefore, the analyzer is capable of extreme sensitivity, and changes in C O concentration of 2 p.p.b. can be measured. The HgO cell, in combination with a series silver oxide oxidation cell (to provide a base or reference mercury vapor level), is sensitive to olefins and aldehydes. However, these are normally present in the ambient atmosphere in very low concentrations relative to CO. Hydrogen and methane in the atmosphere do not interfere with the C O analysis. Condensation Nuclei Analysis. The concentration of submicroscopic particles in the atmosphere is, in general, a measure of contamination from combustion processes. However, there is a low background level in the atmosphere which results from natural processes. The air samples collected near NARL contained 10,000 to 50,000 particles per cubic centimeter (N per cc.) because of the numerous combustion sources within the complex. The concentration of nuclei in uncontaminated Arctic air near our wanigan laboratory varied from an undetectable level to 200 N per cc. The concentration of condensation nuclei in the Arctic atmosphere was measured by a Rich expansion-type condensation nuclei counter (General Electric Co., Type CN). Calibration of the instrument was based on the work of Nolan and Pollak (1946). Although the SRI counter has been calibrated relative to a Scholz counter, no calibration points were obtained at nuclei concentrations as low as that present in the Arctic atmosphere, and the data presented are our extrapolation of prior calibrations at concentrations greater than 900 N per cc. Sumpling Results
It is not practical to list all of the data obtained during this program; however, Figure 3 illustrates the variations in mean daily concentration over the sampling period of CO, methane, and the five nonmethane organic components which were almost always in the atmosphere. With the exception of the unknown, the components were generally present in minimum concentrations at the beginning and end of the sample period, with maximum concentrations occurring from Sept. 5 to 6. The unknown showed a high concentration peak at the beginning of the sampling period as well. Also included in Figure 3 are plots of the resultant wind speed and average daily temperature during the sampling period. The average daily organic concentration patterns all tend to vary in a somewhat Volume 3, Number 3, March 1969
253
1
1
14
12
~
04
0 2 2 9 3 0 3 1 CJG ,1967
1
2
3
7 8 S E P T E M B E R , 1967 4
5
6
9
1011
1
2 4 25 26 27 2 8 29 3 3 31 A U G U S T , 1967
I
2
3
4 5 6 7 8 S E P T E M B E R , 1967
9
13
II
Figure 3. Mean daily temperature, wind speed, and concentrations of atmospheric components during sampling interval
similar manner, the maximum similarity occurring with nbutanol and acetone. There are insufficient data to establish any significant correlations of organic concentration variations with seasonal change. Table I lists all the organic components, major and minor, detected by gas chromatography, as well as methane, CO, and condensation nuclei determined on an hourly basis during a 24-hour period on Sept. 2 and 3. The 24-hour run was designed to explore possible diurnal changes in the concentration of contaminants in Arctic air. Of the five major nonmethane organics graphed in Figure 3, one of the compounds has not been identified. It is apparently a relatively light compound generally present at a concentration of several parts per billion. In Table I, of the 12 nonmethane compounds which were present in the samples, eight compounds were identified and four were classed as unknown peaks. The most apparent anomaly in the results shown in Figure 3 and Table I is the extremely high concentration of n-butanol relative to the concentration of other constituents. It was recognized in the field that this component was present in unusually high concentration. The n-butanol eluted from a Carbowax 20M column under our operating parameters at an elution time similar to toluene. Since toluene was a component of our calibration elution standard mixture, and because nbutanol was an unexpected component in these clean Arctic air samples, considerable effort was made to assure that the peak was not the result of faulty technique or contaminated sampling equipment. Samples were also collected at areas some distance from Point Barrow and provided indications that the Point Barrow 254
Environmental Science & Technology
data were generally representative of wide areas of the Alaskan Arctic. Before samples were taken, the collection bottles were flushed first with helium and then with ambient air. Gas chromatograms of the air within the bottles after this flushing indicated that no organic components were being contributed by the flasks themselves. These more remote air samples were collected at Meade River and Peard Bay after the bottles were flushed with about six volumes of air. These remote samples contained n-butanol and other organics at concentrations comparable to those found at Point Barrow. This would certainly indicate that the presence of n-butanol was not due to local contamination in the vicinity of the laboratory wanigan. The n-butanol peak could not be identified from retention times using the Carbowax 20M and SE-30 liquid phase columns available at Barrow. Therefore, two bottle samples were taken at Barrow on Sept. 11, 1967, and returned to the Menlo Park laboratories of SRI to provide substantiating identification. The atmosphere samples acquired for substantiating identification were collected during an interval when n-butanol concentration was minimal and insufficient sample was returned to permit identification by mass spectrometry. Therefore, the major component was identified as n-butanol rather than toluene by gas chromatographic retention times observed on Carbowax 20M and diglycerol liquid phase columns and on a solid absorption column of Porapak Q. Although it is not possible to identify unequivocally an organic by gas chromatography alone, the possibility is extremely remote that this major contaminant could possess the identical retention time as an n-butanol standard on three different separation columns and not be n-butanol.
The organics other than n-butanol listed in Figure 3 and Table I represent tentative identification based on gas chromatographic retention times observed with a Carbowax 20M liquid phase column. The concentrations of these organics were calculated with the detector sensitivity for the listed compounds except for the following : ethane-ethylene used butane calibration for detector sensitivity; and acetaldehyde, benzene, and all unknowns used detector sensitivity calibration for acetone. Discussion of Results Carbon Monoxide. The carbon monoxide concentration in the Arctic atmosphere was quite low and varied over a range from 5 5 to 260 p.p.b. It usually changed less than 20 p.p.b. during any given day, and most of the measured concentrations were within the range of 60 t o 150 p.p.b. The average was 90 p.p.b. I n general, the C O concentrations at Point Barrow were consistently low and less varied than those measured previously in clean atmospheres. The concentration of atmospheric C O at Point Barrow is considerably lower than that measured during a recent SRI study at a remote location o n the Greenland ice cap (Robinson and Robbins, 1967). During late July and early August, 1967, the data gathered 400 miles east-northeast of Thule, Greenland, at Inge Lehmann station (78" N , 39" W) indicated an average atmospheric CO con-
centration of 200 p.p.b. with a low of 60 p.p.b. and a high of 500 p.p.b. There is some evidence in biological literature that CO may be liberated by growing plants (Wilks, 1959). Thus, C O levels in Arctic areas might have seasonal variations as a function of both photosynthesis activity and contamination by manmade activities. In this regard, CO would be similar to CO?, which is strongly related to both natural and man-made sources. Figure 4 shows the mean daily concentration of CO and the mean daily Cor! index without the relatively small manometric correction. The COz concentration data collected during the interval of our research program were obtained from the University of Washington. The CO, sampling station was about 100 yards north of the SRI sampling site as shown in Figure 1. Figure 4 does not indicate any obvious correlation between C 0 2 and C O concentrations during this period. However, such a short sampling period would indicate only the most obvious of correlations. Figure 5 shows C O concentration variations and CO, index values over a 24-hour period on Sept. 2 and 3, 1967. Again, there is no obvious correlation in diurnal variations. Seasonal variations in CO concentration can only be determined with measurements over longer periods of time and during intervals when variations due to possible natural sources are anticipated.
~~
Date 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/2/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67 9/3/67
Time 1030 1140 1230 1340 1440 1500 1550 1630 1730 1820 1930 2050 2145 2230 2330 0030 01 30 0230 0330 0430 0530 0630 0730 0830 0930
Table I. Carbon Monoxide and Organics in the Atmosphere, Pt. Barrow, Alaska 24-Hr. Analysis, September 2-3, 1967 MethCarEthane,Ii Un- AcetUn- anol, UnUnbon CondenEthylBuPen- known alde- Ace- known Eth- Ben- known known n-Bu- Meth- Mon- sation (2), anol. zene, (3), (4), tanol, ane, oxide, Nuclei, ene. tane, tane. ( I ) . hyde, tone, P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.B. P.P.M. P.P.B. N/CC, 0.04 0.04 0.03 11
0.03 0.03 0.06 0.05 0.08 0.06 0.03 0.06 0.03 0.06 0.05 0.05 0.02 0.05 0.04 0.05 0.05 0.04 0.04 0.06 0.05
0,190.1 0 . 0 5 N.D. 0 . 0 4 N.D. b N.D. N.D. 0 . 0 3 N.D. 0 . 0 8 N.D. 0 . 0 6 N.D. 0 . 1 1 N.D. 0 . 0 4 N.D. 0 . 0 4 N.D. 0 . 0 8 N.D. 0 . 0 5 N.D. 0.06 0 . 0 9 0.030.1 0.05 0 . 2 0.030.1 0.05 0 . 2 0.030.1 0.05 N.D. 0.03 0 . 2 0.04 0 . 3 0 . 1 0.08 0 . 1 N.D. 0 . 1 N.D.
2.3 3.0 3.3 6 9 9.4 3.4 1.8 1.6 1.6 1.2 1.1 0.4 0.4 0.2 0.3 0.3 0.3 0.3 0.3 1.1 0.5 0.9 1.4 1.4 1.7
0.2 0.3 0.3 0.2
0.3 0.1 b
* * *
*
b
b
0.3 6 6
b b
1.6 N.D. 1.5 2.7 2.9 1.9 0.9 0.9 1.3 1.1 0.3 N.D. N.D. 0.8 0.9 1.2 0.9 1.0 0.7 0.7 1.1 0.9 1.0 1.1 1.3
0.3 0.2 0.2 0.3 0.5 0.2
b
0.4 0.4
0.5 0.5 0.9
b
b
b
b
h
*
b
0.4 0.4 0.3
b
0.7
[J
b
6
11
b
h
0.2
1.2
6
b
*
0.9 0.9
*
b
0.7 0.9 0.9 0.9 0.4
b 6
b b
b
0.4 0.5 0.4 0.4 0.2 b
0.1
0.4
0.4 0.1
b
b
0.7 0.7 0.7 0.7 0.9
0 1 0.1 b b
0.3 0.4 0.3 0.2 0.3 0 3 0.4 0.3 0.3 0.4 0.4
b
0.2 0.5 0.1 0.3 0.2 b
0.3 0.2 0.2 0.3 0.2 0.2 0.4 0.3 0.3 b b b
I, ?J
b b
b
96 96 73 91 91 81 97 111 126 121 83 99 83 76 74 68 57 51 73 86 107 83 93 92 90
1.4 1.52 N.D. 1.52 1.41 1.4 1.46 N.D. 1.39 N.D. 1.43 1.35 1.35 N.D. N.D. N.D. N.D. N.D. 1.48 1.50 1.50 N.D. 1.55 1.55 1.65
122 0 89 64 146 0 125 N.D. 134 0 105 127 102 100 119 N.D. 0 102 119 0 100 107 80 119 200 123 100 90 97 110 105 105 0 92 0
Composite peak. * Not detected. N.D. = no d a t a , instrument difficulties. a
Volume 3, Number 3, March 1969
255
Sample bottles of Arctic air returned to Menlo Park and chromatographed on the same equipment and under the same conditions contain comparable concentrations of n-butanol. Normal butanol was identified by retention times on three different columns.
Nonmethane Organic Gases. Most gaseous components of the Arctic atmosphere are present at concentrations of 1 p.p.b. or less. The most obvious anomaly in the organic gas measurements is the presence of n-butanol at 100 times the concentration of any other nonmethane organic. As previously indicated, this was completely unexpected. Other workers in the field have failed to report the presence of nbutanol in studies of clean atmospheres (Williams, 1965). Although Hoff and Kapsalopoulou (1964) have reported the presence of lower alcohols and 2-butanol as a component of wood smoke, n-butanol has not been specifically identified as a component. Since the presence of n-butanol was considered to be unique, considerable effort, as described previously, was undertaken to establish the validity of our results. It is useful to summarize the procedures used to ensure that n-butanol was neither an artifact nor a result of artificial contamination. They are as follows :
Ambient air from Menlo Park does not contain n-butanol at the concentrations found at Point Barrow. Marine air from Hawaii collected in the same sample bottles and analyzed under the same gas chromatographic conditions did not contain n-butanol. Since the purpose of the research program was to determine the concentration of a broad spectrum of organic contaminants in the Arctic atmosphere, no attempt was made to locate the source of the contaminants. However, there can probably be little doubt that the n-butanol source was local, judging both from the magnitude and range of observed concentrations: over one order of magnitude, from 34 to 445 p.p.b. with an average of 190 p.p.b. The most probable source of the nbutanol would seem to be a fermentation process of the tundra cover. At least two strains of bacteria, Clostridium butj-licum and Clostridium ucetobutylicum, are capable of fermenting starch with the production of n-butanol (Fieser and Fieser, 1966). Optimum fermentation for C. ucetobuty/icum occurs at temperatures of 98" to 107" F. Britton (1966) reports that within a few inches of the tundra surface, on sunny summer days, temperatures sometimes reach 100" F. or more. This can occur when air temperatures four to five feet above the ground are only in the high 60's. Other products of fermentation are acetone and ethanol, both of which were identified in the organic samples. Acetone
To ensure that n-butanol was not a contaminant of the interior of the wanigan laboratory introduced by error with the air sample, sample syringes were baked at 125" C. for 12 hours-no change noted. Carrier gas was injected into the gas chromatograph using the same techniques as for sample introduction-no n-butanol detected. Samples from 50 miles distance contain comparable concentrations of n-butanol; therefore, contamination was not local to area near the wanigan.
+
+
Neither carrier gas 20% oxygen nor nitrogen 20z oxygen injected into the gas chromatograph produced an artifact resembling elution peak of n-butanol.
200
L
3112 24
I
A
I
I
1
25
26
27
I
1
I
28 AUGUST, 1967
29
30
31
I
2
3
I
I
4
5
S E P T E M B E R , 1967
7
Figure 4. Mean daily concentration of CO and C 0 2 during sampling interval 256 Environmental Science & Technologj
I
1
6
8
9
IO
11
0
“
90 70
I
50
i
1
1
I
I
I
15 16 17 18 19 SEPTEMBER 2 , 1967
20
1
~
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
J
,
310
9
IO
/I
‘2 13
14
21
22
23
24
1
2
3 4 5 6 7 SEPTEMBER 3 , 1967
8
9
IO
Figure 5. Hourly concentration variation of CO and CO, over 24-hour interval
in particular was almost always present and for the representative samples listed in Table I had an average concentration of 0.8 p.p.b. and a range from nil to 1.9 p.p.b. Ethanol formed a composite peak with methanol, and together the two materials occurred sporadically a t a level of about 1 p.p.b. The concentrations of ethane and ethylene, detected as a composite peak, averaged 0.05 p.p.b. and ranged from nil to 0.9 p.p.b. These two compounds are also produced by biological mechanisms, and it seems likely that these concentrations represent a local natural source. Others of the detected organic compounds are commonly found in natural gas, and this may be the source of such materials as butane, which ranged from nil to 0.20 p.p.b. and averaged 0.06 p.p.b. Other common natural gas constituents which occurred less frequently in our samples are pentane and benzene. When these materials occurred, their concentrations generally were in the range of 0.1 t o 0.3 p.p.b. Methane and Condensation Nuclei Concentration. The concentration of methane in the Arctic atmosphere is similar to that measured in other clean atmosphere areas (Junge, 1963). The mean methane concentration was 1.59 p.p.m., as compared to the estimated worldwide methane concentration in the range of 1.2 to 1.5 p.p.m. The presence of natural gas wells and pipelines in the Barrow area could contribute methane through leaks and seepage. Vegetation decay proccesses are also common sources of methane. The condensation nuclei concentrations were found to be extremely low and indicative of an atmosphere unpolluted by local combustion sources.
Acknowledgment The field work at Naval Arctic Research Laboratory was ably supported by Max Brewer, NARL, and his staff. Literature Cited Altshuller, A. P., Ortman, G . C., Saltzmann, B. E., J . Air Pollution Control Assoc. 16, 87 (1966). Britton, Max E., “Vegetation of the Arctic Tundra,” Oregon State University Press, Corvallis, Ore., 1966. Fieser, L. S., Fieser, M., “Organic Chemistry,” Heath, Boston, 1966. Hoff, J. E., Kapsalopoulou, A. J., J. Gas Chromatog. 2, 296 (1964). Junge, C. E., “Air Chemistry and Radioactivity,” p. 95, Academic Press, New York, 1963. Kelley, John J., Jr., “An Analysis of Carbon Dioxide in the Arctic Atmosphere at Point Barrow, Alaska,” University of Washington, O N R Contract No. 477(24)NR 307-252, May, 1964. Kelley, John J., Jr., University of Washington, NARL Station, unpublished data, Aug.-Sept., 1967. Nolan, P. J., Pollak, L. W., Proc. Roy. Zrish Acad. Sci. 1 (Sec. A, No. 2), Hodges, Figgis, and Co., Dublin, 1946. Robbins, R. C., Borg, K . M., Robinson, E., J . Air Pollution Control Assoc. 18, 106 (1968). Robinson, E., Robbins, R. C., “Carbon Monoxide Content of Glacial Ice and the Natural Atmosphere,” project data, SRI Project PRU-6508, 1967. Wilks, S. S., Science 129,964-6 (1959). Williams, E. H., Anal. Chem. 37, 1723 (1965). Receiced for reciew August 36, 1968. Accepted December 9, 1968. This research wus supported by the Ofice of Naral Research, Geography Branch, under Contract No. NOOO14-67 C-0515.
Volume 3, Number 3, March 1969 257
-”
..
- ..
.
