JOURNALOF GEOPHYSICALRESEARCH VOLUME 66, No. 1 JANUARY 1961

Helium, Argon, and Carbonin SomeNatural Gases

R. E. ZARTMAN AND G. J. WASSERBURG

The Cali/ornia Institute o• Technology

AND

J. H. REYNOLDS

University o.f Cali/ornia Berkeley, California

Abstract. Thirty-nine samplesof natural gasesrepresenting varied chemicalcompositions and geologicaloccurrences were analyzed for their helium, radiogenicargon, and atmosphericargon contents.The total range in the (He/A)r•d ratio was found to be 1.6 to 130 with most samples having values between 6 and 25. This range of values is essentiallyequal to the productionratio from the uranium, thorium, and potassiumin average igneousrocks and a wide variety of sedi- ments. This indicates that all of these natural gaseshave obtained their radiogenic gasesfrom rather averagerock types. This is true in spite of the fact that the gasesrange in helium content from 37 to 62,000 ppm. A theoretical discussionof the origin of helium and argon in natural gasesis given. It can be shownfrom the ratio of nitrogen to atmosphericargon that most of the nitrogen in these gases cannot comefrom the entrapment of air. From a considerationof the concentrationof atmospheric argon in natural gasesit is possibleto estimate the proportion of gaseousand aqueous phases assumingdiffusive equilibrium. The isotopiccomposition of the carbonin the methane of these gaseswas found to be very light. It was shownthat for coexistingCH 4-CO 2pairs the carbondioxide was always isotopicallyheavier.

INTRODUCTION reservoirgreatly enrichedin U and/or Th, or the accumulation from a rather normal The purposeof this study was to investigate the relationship between the abundances of rock reservoir.Lastly, there existsthe possibility helium and argon in natural gasesof different that the helium representssome more primordial compositionsand environmentsand to examine gases trapped in the earth during its early the isotopic compositionof the carbon in these history and subsequentlypartly releasedinto gases.Since the discoveryof terrestrial helium stratigraphicand structural traps. These various in natural gasesby Cady and McFarland [1906], possibilitieshave been recognizedby some of a vigoroussearch has beenmade for helium-rich the earliest investigators in the field [Rogers, gases,and a considerablenumber of total gas 1921]. It has been found [Faul, Gott, Manger, analyses that include helium determinations Mytton, and Sakakura,1954; Sakakura, Lindberg, exist in the literature. These have been made and Faul, 1959] that severalhelium wells, i.e., using both volumetric and mass spectrometric wells containingover 0.5 per cent helium, are techniques[Rogers, 1921; Andersonand Hinson, radioactive owing to a high radon content and 1951; Boone,1958]. appear to be associatedwith some uraniferous The origin of high helium natural gases,some petroleum residues. Other high helium wells, of which have helium contents as great as however, are devoid of such radioactivity. In 10 per cent, has beenthe subjectof considerable addition, many well gaseshave a high radon speculation. The He4 contained in them is contentand only smallconcentrations of helium presumablythe productof the radioactivedecay [Satterlyand McLennan, 1918]. The composition of U •'35,U•'% and Th 2•'-,and their intermediate of helium-rich gaseshas been found to be quite daughter products. A fundamentalquestion variable,although in somegas fields a correlation ariseswhether these quantities of helium repre- has been suggestedbetween helium and nitrogen sent thc accumulatcd dccay products of a content. 277 278 ZARTMAN, WASSERBURG, AND REYNOLDS

The isotopiccomposition of helium from gas estimates of the contribution of 'atmospheric wells has been investigatedby Aldrich and Ni• gases. [1948]who report valuesof He•/He4 -- 10-7. EXPERIMENTAL TECHNIQUES The results of these authors represent the only publisheddata on He• content in terrestrial We have studied a variety of natural gases gases.The origin of the He• in thesegases has coveringa wide compositionaland environmental been consideredby severalworkers. Hill [1941] range in order to attempt some understanding suggestedthat this isotopeof helium could be of their origin. In this study we have restricted producedin rocksby the reactionLi6(n, a)H8-• our effortsto well gasesand have not investigated He• -]- /•-. More recently, Morrison and Pine volcanic or hot spring emanations.Thirty-nine [1955]have discussedthe relative productionof samples were analyzed for their helium and I-Ie• and I-Ie• in rocks. As shown by Wetherill argon contents and for isotopic composition. [1953, 1954], the principal sourcesof neutrons Partial gas analysesfor other major constituents are the reaction O•s(a, n)Ne • and spontaneous were also made. The isotopiccomposition of the fissionof U TM.By consideringthe variousnuclear total gas carbon, methane carbon, and carbon reactionswhich competefor neutrons,Morrison dioxide carbon were determined. The results of and Pine concludethat the He•/He• ratio this investigation are presentedin Tables 1 and observedby Aldrich and Nier in natural gases 2, together with pertinent well data for each is most reasonablythe productof theseprocesses sample. The accompanyingtotal gas analyses in materials which contain neither uranium nor representdata suppliedus by the participating thorium, or both, in very great concentration. petroleumand natural gascompanies, partial gas Their argumentswould not, however,preclude analyses performed in our laboratory, and a the originof this heliumto be fromfinely divided combination of these two sources. With the uranium minerals disseminated in a rather exception of three samples, the H•S contents normal rock. This would permit uranium con- were less than 0.01 per cent. For samples 14, centrationsof up to a few tenths of 1 per cent. 34, and 35, the H•.S contents were 0.04, 0.09, The radioactive decay of K •ø gives rise to and 0.12 per cent, respectively. the possibilityof high argon-containingnatural With the exceptionof samples29, 30, 31, and gases.Since potassium is a principalelement in 32, the well gaseswere collectedusing standard most crustal rocks as comparedwith uranium high-pressure,stainless steel gas cylinders with and thorium, a study of the ratio of radiogenic valves on both ends. The cylinders were con- heliumto radiogenicargon, (He•/A'ø)•, gives nected to the gas sourceand purged of air by more direct information on the possiblesource passingwell gas through them under positive of helium gas wells.• pressurefor several minutes. The outlet valve Someresults using this approachwere reported was then closed,and the pressurein the cylinder by Wasserburg,Czamanske, Faul, and Hayden was allowed to reach a satisfactoryvalue (20- [1957] for some helium wells in the Texas 3000 psig), and then the valve to the sourcewas Panhandle. These workers showed that the closed. All of these sample vesselswere at a argonin four heliumwells was about 70 per cent pressureof considerablyover I atmospherewhen radiogenicand that the ratio (He4/A•9•.awas they wereused for analysis.Samples 29-32 from about 10. This ratio was well within the values the Texas Panhandlegas field were obtained by to be expectedfrom the present-dayproduction Henry Faul in 1954. They were collected in rates of He• and A •ø in normal igneousrocks. glasscylinders that had openingsat both ends. They therefore concludedthat these helium After allowingthe gasto passthrough the cylin- wellswere probably formed from the accumulated dersat slightpositive pressure for a few minutes, radiogenicgases found in rockswith a ratio of both ends of the vesselwere sealedoff. In every U/K typical of normaligneous rocks, and that instancethe helium, argon, and carbon analyses in no way could these gasesbe the result of were done on the samesample. accumulation from an enriched uranium reservoir. The gas cylinderswere joined onto a vacuum For thesesamples it wasshown that the A•/A •6 line (Fig. 1) with metal-to-glasscouplings and ratio was the sameas that found in atmospheric the systemwas evacuatedwith all the furnaces argon and it was thereforepossible to make heated and outgassedbefore each run. After the HELIUM, ARGON, AND CARBON 279

*<---To hicjh vacuum

Manometer

Cold trap

• t•ASample tubeM•c•Leodcjaucjefurnace Toe ==•..•fur

Fig. 1. Vacuumapparatus used for helium,argon, and carbondetermination and extraction. systemwas foundto be vacuumtight, somegas waschecked by runningsamples of knownhelium was releasedinto a part of the systemof known concentrationand by comparingthe results of volume and the pressuremeasured on a manom- duplicate analysesin which the combustiontime eter. For most samples a tracer of A ss was and the amountsof gas were varied by a factor introduced and the gases mixed. The sample of 4. The purity was checkedby observingthe size rangedfrom 2 to 100 cc STP. For total gas Tesla discharge color. No detectable loss of carbon analyses, all gas was then pumped helium by diffusionthrough the fused quartz directlyinto the combustionsystem. For methane combustiontubes was observed.According to carbonanalyses, liquid N, was placedon a cold the work of Norton [1953] on the diffusion of trap and the noncond•nsiblefraction was helium through fused quartz at various tem- continuously pumped into the combustion peratures,it was expectedthat less than I per systemuntil no gas phaseremained. The com- cent of the helium would be lost by diffusion bustion procedure used was similar to that during the combustionprocedure. Blank runs described by Craig [1953]. A CuO furnace, carried out under conditions similar to those in made of fusedquartz, was heatedto 900-950øC, which a sample was being analyzed showedno and the gases were cycled through it until appreciableintroduction of inert gasesinto the complete combustion to CO•. was obtained. line. During this combustionperiod, dry ice baths The argon contentwas determinedby (1) the were placedon a cold trap to collectthe water differencebetween the volumetricallydetermined produced.After combustion,the dry ice baths total noble gas residueand the helium content, were replaced with liquid nitrogen and the (2) a volumetric determinationof the condensible noncondensibles--includingN2, He, and A--were fractionafter a secondpurification (before which transferred by a Toepler pump into another the helium was pumped off and the Ti furnace sectionof the line (of known volume) containing outgassed),and (3) isotope dilution using A 88 a Ti furnace,1VfcLeod gages, and a sampletube tracers of known content as described by filled with activated charcoal. The Ti furnace Wasserburgand Hayden [1955]. The volumetri- was heated to approximately950øC and the gas cally determined argon values are generally sample purified. The amount of this noble gas higher than the isotope dilution value. The residuewas then measured,using a McLeod gage. first method yields argon contentsup to 40 per Liquid N, was next applied to the charcoaltrap cent higher than the isotope dilution value, and the condensible fraction quantitatively whereasthe secondmethod showsdiscrepancies absorbed. The pressure was again measured of up to 15 per cent. The resultsobtained by the and attributed to helium. The argon sample first method were highly reproducibleover a tube was sealed off from the line and mass long period of time. They were, nevertheless, spectrometricallyanalyzed. The helium yield frequently in error, presumably owing to the 280 ZARTMAN, WASSERBURG, AND REYNOLDS

ß ' u'D•00c.o r-,,

ooooooo .•.•.•

•oooo o•oo.•'•'•-•• •'•••• ••••••oooo.••• ••• HELIUM, ARGON, AND CARBON 281 282 ZARTMAN, WASSERBURG,AND REYNOLDS TABLE 1. Continued B. Analytical data

Lithology of No. producingzone CH4% C•.H6% N•.% CO•.% He ppm A ppm

i ss 98.16 0.13 1.60 0.11 47.5 57.4 2 ss 95.43 2.37 1.57 0.09 41.2 52.2 3 ss 95.12 3.96 0.86 0.06 37 67 4 ss 68.91 0.01 30.97 0.09 96.6 140 5 ss 96.20 0.04 3.76 0.00 101 88.2 6 ss 90.70 0.11 9.15 0.03 37.6 53.1 7 ss 97.32 0.01 2.65 0.02 85 125 8 ss 84.5 10:5 4.03 1.03 101 32 9 ss 90.5 5.3 2.40 1.82 63 39 10 ss 86.7 11.0 • 2.30 140 118 11 ss 53.5 6.6 36.6 26000 1400 12 ss 82.7 14.8 1.0 1.52 152 26.5 13 ss 75.1 23.2 0.87 0.77 151 42.0 14 lms 56.1 21.4 22.2 0.1 1350 360 15 ss 96.52 1.06 1.48 0.90 359 81.9 16 sh and ss 0.0 0.0 0.12 99.9 44.5 28.0 17 sh and ss 0.0 0.0 0.15 99.8 46.6 29.2 18 ss 82.58 15.36 1.98 0.05 348 57.3 19 ss 82.4 14.8 2.6 0.17 480 66.8 20 lms 86.3 10.0 3.4 0.25 69.4 117 21 ss 91.10 7.87 0.94 0.07 172 46.1 22 lms 38.0 10.6 42.5 2.1 62200 5630 23 ss 86.2 8.0 (1) 5.6 1640 77.1 24 ss 65.3 6.7 24.9 •)' 22600 1080 25 ss 97.5 2.0 (1) 0.50 ... 203 6.8 26 ss 95.1 2.8 (1) 1.9 ... 1575 35.8 27 ss 75.7 23.6 (1) 0.64 8O5 14.9 28 ssan d lms 76.5 15.9 6.5 1720 376 29 dolo 73.8 9.6 ( 1) 16.6 9370 877 30 dolo 84.0 6.4 8.95 •)' 4180 470 31 dolo 90.3•- . .. 9.7 ... 4480 482 32 dolo 90.42 9.6 4170 461 as dolo 72.2 •i:• 15.2 616 7OOO 710 34 as 0.0 0.0 0.60 99.3 232 79.5 35 ss 0.0 0.0 0.55 99.3 187 65.4 36 ss 86.2 9.4 2.08 2.29 75 27 37 ss 88.4 10.3 1.22 0.10 158 73.4 38 ss 94.9 5.4 1.8 0.85 757 66.8 39 ss 89.8 7.0 3.09 0.85 152 17.5 40 lms reef 72.6 24.2 2.5 0.6 593 117 41 lms reef 71.1 7.9 3.6 5.0 1670 418

IncludesCO2; 2 Includes C2H6 and higher carbons, and CO2. presenceof contaminatinggases. A second The carbondioxide in the natural gaseswas purificationas describedin (2) gavevolumetric removedby placingliquid nitrogenon a cold valuesin relativelygood agreement with (3). trap and pumping off the noncondensibles All samplesexcept 3, 7, 8, 9, 10,and 36 havebeen throughanother cold trap in order to retain run by isotopicdilution, and, exceptfor these any CO•.that mightbe lost from the first trap. samples,the valuesfor argoncontent listed in The remaininggases, which include CO• and Tablei havebeen determined by isotope dilution. hydrocarbonsless volatile than CH4, werethen After the noncondensibleswere pumped out transferredinto a reactionvessel containing a ofthe part of the line containing the CuO furnace, saturated Ba(OH)•. solution with 80 per cent the liquidnitrogen baths that holdH•.O and C02 phosphoricacid in a sidearm. The CO•.was then were again replacedby dry ice baths and the convertedto the carbonate,and the hydrocarbons C02was allowed to sublimeinto this part of the were pumpedoff. The carbonatewas reconverted line.The CO•.was then transferred into a sample to CO•.by reactionwith phosphoricacid and the tube for isotopicanalysis. CO•.transferred to a vesselwith liquid nitrogen. HELIUM, ARGON, AND CARBON 283 TABLE 1. Continued B. Analytical data

Arad Aair No. ppm ppm • % He/A•.• A4O/A• A•s/A• (N:/A,•,) f,,, V./V,

1 3.69 53.7 6.4 12.9 316 4. 3 0.219 -4- 0.018 298 0.68 25 2 3.9 48.3 7.5 10.6 320 4- 3 325 0.69 24 3 4.4 63 6.6 8.4 3174- 3 0.202':j:' 0.015 137 0.46 62 4 15.7 124 11.2 6.15 334 4- 2 2500 0.49 55 5 4.64 83.6 5.3 21.8 312 4- 6 0.196'•' 0.007 450 0.50 53 6 9.47 43.6 17.8 3.97 361 4- 2 0.195 4- 0.015 2100 0.56 42 7 9.3 116 7.4 9.2 320 4- 6 0.194 4- 0.007 228 0.51 51 8 11.2 21 35.0 9.0 456 4- 14 0.201 4- 0.009 1920 0.68 25 9 4.8 35 12.3 13.1 337 4- 2 0.195 4- 0.007 706 0.74 19 10 13.6 104 11.5 10.3 336 4- 2 0.195 4- 0.010 221 0.44 67 11 1260 140 90.0 20.6 2818 4- 140 . .. 2610 -0.31 12 15.9 10.6 60.0 9.56 736 4- 8 ... 943 0.83 ii' 13 27.2 14.8 64.8 5.55 841 4- 12 ... 588 O. 74 19 14 270 90 75.0 5.00 1185 4- 12 ... 2470 O. 37 90 15 48.2 33.7 58.8 7.46 720 4- 6 . .. 439 0.54 45 16 27.6 0.4 98.4 1.61 22500 4- 1100 . . . 3000 1.00 0 17 29.0 0.2 99.3 1.61 34000 4- 4000 . . . 7500 1.00 0 18 28.2 29.1 49.2 12.3 587 4- 3 . .. 680 o. 71 22 19 37.9 28.9 56.7 12.7 684 4- 3 ... 900 o. 65 29 20 28.1 89 24.0 2.47 389 4- 2 ... 382 o. 43 70 21 13.4 32.7 29.1 12.8 417 4- 2 ... 287 0.84 lO 22 5580 50 99.0 11.1 29100 4- 4000 . . . 8500 o. 36 94 23 62.6 14.5 81.3 26.2 1570 4- 20 . . . 3860 o. 92 5 24 969 111 89.6 23.3 2720 4- 140 ... 2240 0.60 35 25 4.2 2.6 61.8 48 775 4- 30 . . . 1920 0.97 1.6 26 17.7 18.1 49.5 89 587 4- 5 . . . 1050 0.99 0.6 27 6.0 8.9 40.2 134 494 4- 5 . .. 720 0.97 1.6 28 250 126 66.5 6.88 883 4- 10 515 --0.22 29 697 180 79.3 13.4 14354- 14 0.197':•' 0.003 o.oo 30 375 95 79.7 11.2 1465 4- 15 0.201 4- 0.014 942 o. 82 12 31 359 123 74.5 12.5 1160 4- 12 0.186 4- 0.023 789 0.77 16 32 336 125 72.9 12.4 1100 4- 11 0.198 4- 0.004 768 0.77 16 33 518 192 72.9 13.5 1195 4- 11 ... 792 34 75.2 4.3 94.5 3.09 5300 4- 120 ... 1395 6'•8 'i'1 35 60.7 4.7 92.7 3.08 4150 4- 95 1170 0.98 1.1 36 10.0 17 37.0 7.5 4724- 15 0.203':•'0.010 1224 0.82 12 37 12.5 60.9 17.0 12.6 355 4- 3 0.188 4- 0.008 200 o. 76 17 38 59.1 7.7 88.4 12.8 2560 4- 130 0.204 4- 0.024 2340 O. 97 1.6 39 12.2 5.3 69.7 12.5 960 4- 24 0.178 4- 0.018 5830 0.88 7 40 73.4 43.6 62.8 8.09 796 4- 10 ... 573 0.35 96 41 381 37 91.1 4.38 3316 4- 150 ... 973 0.29 130

The accompanyingwater was then held with a analyzed; the argon isotopic compositionis dry icebath, andthe purifiedCOs transferred into given in Table I alongwith the mean deviation a samplevessel and massanalyzed. of the analysis.A88/A 86 ratios are given for In someof the carbondioxide well samples, sampleson whicha run wasmade without using hydrocarbonswere not detected,and therefore a tracer. The total carbon and methane carbon the CO: was purifiedby simplypumping off isotopic analysesare reproducibleto within the noncondensiblesat liquid nitrogen tempera- 0.2 per mil, and the CO: carbonanalyses are ture. The Farnham Dome samplescontained reproducibleto within I per mil. The isotopic considerableamounts of I-I•S, and this was compositionof the argon was determined on removedby passingthe gas repeatedlyover mass spectrometers at the Californi• Institute heatedsilver filings. of Technologyand the Universityof California All of the heliumanalyses and thoseargon at Berkeley. Agreementwas good on several analysesthat were determinedby isotopic samplesrun by both laboratories.The back- dilutionare most probablyaccurate to within groundwas always much lower than the sample 10 per cent. The heliumwas not isotopically peak heights. 284 ZARTMAN, WASSERBURG, AND REYNOLDS

The carbon was analyzed at the California sediments,air dissolvedin water associatedwith Institute of Technology on a 6-inch mass the sediments,gases brought in at some later spectrometerin Dr. Epstein's laboratory. All time by circulating ground water, and con- sampleson which carbonisotopic analyses were tamination during sampling and analysis are made were run as carbon dioxide accordingto all possiblesources. By meansof blank runs and the proceduredescribed by McKinney, McCrea, repeat analyses, it was possibleto show that Epstein, Allen, and Urey [1950]. The samples essentiallynone of the atmosphericargon was were analyzed relative to Caltech Working dueto contaminationfrom laboratoryprocedures. StandardII, and then convertedto corresponding Samples were run with atmospheric argon values relative to the Chicago Standard PDB. contentsas low as 0.2 ppm without any difficulty. Repeat analyses of samples at different times •ITROGEN AND ATMOSPHERIC ARGON yieldedidentical A4ø/A s6 ratios. Inspectionof Mass spectrometricanalyses of the extracted the A4o/As6 ratios shows that thesevalues tend argon samplesshowed that while the A•0/A• to be rather constant within certain gas fields. ratios rangedfrom 312 to 34,000, the A•s/A• Althoughsome of the samplesmay possiblyhave ratio was practically constant and was, to been contaminated during sample collection, within experimentalerror, equal to the value of it is our conclusionthat, for most cases,the 0.187 found in atmosphericargon. atmosphericfractions represent the accumulated Each samplewas found to have an A•ø/A• air argon associatedwith these gases during ratio greater than 295.6, the value for air argon their natural evolution. While the composition [Nier, 1950].These gases appear to be mixturesof and abundanceof atmosphericargon may have radiogenicargon (A•a) and atmosphericargon changedthrough geologic time, we have assumed (A•). The fraction, e, of the total argon that is .the present-day values. It is conceivablethat radiogenicis given by the expression measurementsof the A40/Aa6 ratio may prove useful as a tracer in identifying and studying A• - 1- 296.8 gas reservoircharacteristics. e- Ato•a• (A4ø/ASO)sd-1.2 (1) The natural gasesanalyzed contninedamounts where (Aaø/A•)• is the ratio in the sample.• of nitrogenvarying from 0.1 to 42.5 per cent. These parameters are presented in Table 1. Gases composedof essentially 100 per cent It is not possibleto say under what conditions nitrogen have been reported in the literature, the atmospheric argon was introduced into but we havenot beenable to obtainsuch samples. the samples.Air bubblestrapped in the original The origin of nitrogen in natural gaseshas been often discussed[Hoering, 1957; Scalan, thesis, • The formulas used •o calculate the amounts of University of Arkansas;Rayleigh, 1939; Rogers, radiogenic (Aaø)• and to•al sample (Aa0)• when a 1921; and Zobell, 1952] in terms of (1) the tracer, t, was used are: introductionof atmosphericnitrogen, (2) the releaseof nitrogenby the bacterialdecomposition = _ [\A \A ],\A

where the subscriptst, n, and y refer to the appro- priate ratio in the tracer, in normal atmospheric argon, and in the sample-tracermixture, respec- and tively. HELIUM, ARGON, AND CARBON 285

of nitrogen-bearingcompounds, (3) the release disintegrationof uranium and thorium, and the of nitrogenby the inorganicchemical breakdown radiogenicfraction of the argon by the electron- of organic compounds,and (4) the liberation of capture decay of K •ø, we can make certain inorganic nitrogen from igneous (and meta- theoretical calculations to determine the values to morphic) rocks. be expectedfor the radiogenichelium and argon One of the processescalled upon most fre- abundancesand their ratio (R -- (I-Ie/A),•a). quently to explain the presenceof nitrogen in In order to discuss the abundances of helium and well gasesis the incorporationof air in the pore argon observedin natural gases,it is necessary spaceof variousreservoir rocks. I[ air is included to consider (1) the variation of He'- and A •ø- in the sediments,they will contain not, only productionwith time, (2) the natural distribution but other atmospheric gases as well. Some of uranium, thorium, and potassium in rocks workers have assumedthat when gaseshave a and their distribution among the mineral phases (N2/A) ratio approximatelyequal to that in present, (3) the efficiencywith which the helium air, the nitrogenis due to atmosphericcontamina- and argon can escapefrom crystal lattices and tion. It was pointedout by Wasserburg,Czaman- becomeavailable for accumulation,and (4) the ske, Faul, and Hayden [1957] for certain high processof gas migration and accumulation into helium gases from the Texas Panhandle that gas reservoirs.The first part of the discussion whilethe total (N2/A) ratiosof thesegases were will be devoted principally to estimates of the approximately equal to the atmosphericvalue, (He/A)r,• ratio. The problem of absolute most of the argon was of radiogenicorigin, and concentrations and amounts of these noble the actual (N•/A•ir) ratio was about 850. gaseswill be treated later. They concluded, therefore, that most of the Becausethe half-livesof the uranium, thorium, nitrogen could not be due to trapped air. Fol- and potassiumisotopes are of the order of 10• lowing this approach,we have made a similar years or more, the rates of production of helium calculationof the (N2/A,i r) ratio for our sam- and argon are rather constant for the past ples. The results are given in Table 1. The several hundred million years. For such a time ratio is observedto range between 137 and 8500. period we obtain the relationships The atmosphericratio of nitrogen to argon is 84, and the ratio for the dissolvedgases in water .20 U -{- 0.29 Th)• X 10-• in equilibrium with the atmosphere at 20øC N^ -----3.99 K • X 10-• is 38. Althoughit may be reasonableto postulate simple mechanismsof gas solution and effer- where Nu. and N• are the cc STP of helium vescence, which could account for a factor of and argon, respectively, produced by radio- 2 or 3 changein the atmospheric(N,/A, i r) active decay during a time interval of • years• ratio, most of the gassamples show much higher and U, Th, and I• are the weightsof uranium, nitrogen enrichments. Therefore, in most in- thorium, and potassium in grams. Dividing stances only a small fraction of the nitrogen these equations we see that the ratio R of the content of thesegases reasonably can be attrib- accumulated radiogenic gases is given approx- uted to incorporatedair. imately by Aswill be shownlater, the ratioof (He/A)raa in all samplesinvestigated in this report had R '• [(U/K) X 10a(3.0-{- 0.72(Th/U))] (3) values of 1.6-130. Such a range in this ratio is which is independent of the time. It can be believedto be characteristicof all natural gases, seen that R is not very sensitiveto changesin and therefore,if a gasshows a (He/A) ratio of the Th/U ratio as a shift of from 0 to 4 in this much less than this, it may be assumed,in lieu quantity only changesR by a factor of 2. With of an argon isotopicanalysis, that the argon is the exceptionof severalt;horium-rich minerals, mainly nonradiogenicand has a corresponding the Th/U ratio is rarelygreater than 10. Using amount of atmosphericnitrogen associated with it. values of 3.5 ppm, 10 ppm, and 2.6 per cent for the average rock concentration of uranium, I-IELIUM-ARGON RA•IO thorium, and potassium,respectively, we obtain If essentiallyall of the helium containedin R---• 7. This is in rough agreement with the natural gasesis producedby the radioactive averageof the obscrvcdresults (see Tabl%l). 286 ZARTMAN, WASSERBURG, AND REYNOLDS

ixiO•ø 8

6

I•109 2 I 3 8

6

4 •yrs

8 IxlO 8

ixlO7 I I I I I I 1 0 I 2 3 4 5 6 7 R: (He/A•)rod

Fig. 2. Variation of R -- (He/A),,a with time. (1) The ratio R of the productionrates at a time r yearsago. (2) The ratio R of the total gasesproduced in the intervalfrom a time 4.5 X 109 yearsago until r yearsago. (3) The ratio R of the total gasesproduced in the intervalfrom r years ago to the present.

In order to consider the effects of time on R concentrationsin an average igneous rock as over time intervals longer than a few hundred estimated from selected values in the literature. million years, we have calculatedthis quantity Inspectionof Figure 2 showsthat the time for the followinginstances' dependencyof R is rather slight. This ratio changesfrom 2.0 to 6.8 in the extreme cases 1. The ratio R of the productionrates at a time considered.Except for the dependencyon the r years ago. Th/U ratio, similar curves of this shape are 2. The ratio R of the total gasesproduced in generatedby any U/K ratio. They are simply the interval from a time 4.5 X 10' years ago shifted along the abseissain proportionto the until r years ago. U/K ratio. As indicatedabove, the rangein R 3. The ratio R of the total gasesproduced in is a factor of 3.4. If the helium and argon con- the interval from r years ago to the present. rained in traps in the upper sedimentarycrust These curvesare shownin Figure 2. The pre- are composedof locally derived radiogenic viously stated valuesof uranium,thorium, and material,we wouldexpect the (HO/A4ø),•aratio potassiumwere used to representthe present to be close to the instantaneous value. If these HELIUM, ARGON, AND CARBON 287 gasesare composedchiefly of material in the it is necessaryto considerin more detail the processof upward transport from the deeper abundances of the radioactive elements in crust and mantle, we might well be dealingwith differentrocks. A compilationof the abundances a (He4/A4ø)r•dratio somewherebetween the in various lithologies is presented in Table 3. instantaneous value and the cumulative ones. Wide variationsin concentrationsand U/K and The validity of employingsuch an 'average' Th/U ratios are apparent. For unusual rock rock as used above is indeed questionable,and types (Kolm from Sweden), or where the rock is the host for a mineral deposit (Colorado TABLE 2. Carbon isotopic composition of the Plateau uranium deposits),the ratios may vary total gas, CH4, and C02 containedin somenatural over many powers of 10. Of the normal rock gases.Apparent temperaturescalculated from Craig [1953] for the pair CH4-C02 are also given. All types tabulated, it is seenthat the purer carbon- carbonisotopic data are given as per mil difference ates, although not rich in uranium, have very in the Cx•/Cx• ratio between the sample and the low potassiumcontents yielding ratios as much Chicagostandard, PDB. as one hundred times greater than granites. On the otherhand, an evaporiteproduced from Calculated uncontaminatedsea water would yield a U/K NO. •' Total •CH, •CO, T,øC ratio one hundred times less, and, in addition, may have a Th/U ratio closeto zero.Exclusive i --36.0 ...... -ss.s ...... of carbonates,however, the normal rock types 3 --35.1 ...... exhibita rather constantU/K ratio...... The possiblerange in the U/K ratio is about ...... 5 powersof 10 in extreme rock types. If some ...... 7 --40.2 natural processesoperate over a sufficiently s large scale,they will tend to averageover such 9 --37.2 --39.7 +8.1 71 variations. For example, carbonate rocks are 10 --40.1 --43.5 frequentlyintercalated with shalestending within 11 ... --40.3 --•(•'9 a given stratigraphicunit to cancelthe differences 12 ... --44.1 --7.9 139 13 ... --46.4 --21.9 242 obtainingin the pure lithologies'a 50-50 mixture 14 . .. --41.6 of a pure limestoneand a shale will yield a 15 ... --31.9 --•'7 rather normalU/K ratio. In general,it will be 16 ...... --3.9 ... exceedingly difficult to relate the radiogenic 17 ... --4.1 ... 18 ... --•16 ...... helium and argon in any reservoir to a par- 19 ... --44.1 ticular rock source,since these gases undoubtedly 20 ... -•o.s -i•:4 will have migrated over some distances and 21 --41.0 throughdifferent rock types.In certain instances -s.s it may be possibleto showthat thesegases could .a . -a9.s ...... havebeen produced within their presentreservoir, 25 ...... but this will not be a uniqueattribute. Consider- 26 --49.8 ...... ing the wide variations possiblein R, owing to 27 --•'2 --47.0 ... variationsof the U/K ratio, it is remarkable 28 --41.7 --i•14 ... 29 --:•):9 --42.0 ...... that the measured values lie within such a 30 --38.8 --40.0 ...... restricted range. It is our interpretation that 31 --38.5 ...... this is because of the large- and small-scale 32 --39.6 ...... averaging inherent in the processesof accumu- 33 --39.2 ...... 34 ...... --,•:0 ... lation with the total sourcehaving a 'normal' 35 --5.9 U/K ratio. 36 --:•:2 --:•):5 --15.0 One of the important factors concerningthe -as.0 -4o. He and A to be investigatedis the escapeof 38 --39.4 --41.5 +i•'8 ',i• these gasesfrom the minerals in which they 39 --37.5 --38.5 --8.8 185 40 ... --45.4 originated. That A•0-K •0 and helium age-dating ... methods work at all attests to the retention of at least someof thesegases in the sourceminerals. 288 ZARTMAN, WASSERBURG, AND REYNOLDS ItELIUM, ARGON, AND CARBON 289

Much work has been done by Keevil [1941] and TABLE 5. Henry's Law Constants for Hurley and Goodman[1941] on the lossof He Several Selected Gases from individual minerals and total rock, and correspondingwork by Wasserburg,Hayden, and K, = •/•' Jensen [1956] and Goldich, Baadsgaard,Nier, Solvent and Hoffman [1957]yields someinformation on Fresh Water Marine Water the argon retention ability of several common potassiumminerals. Some data on I-Ie and A Gas 15øC 50øC 15øC 50øC lossesfrom mineralsand rocks are prcsentedin Table 4. Since,in general,the observedhelium N 2 56 87 73 114 and argonlosses are roughlyproportional to the A 25 40 33 53 helium and argon contents,respectively, of the Hc 113 Ill ...... mineral,the (He4/A4ø),aaratio of the 'available' CH 4 27 46 ...... gas will be equalto the (He4/A4ø) production ratio in the total rock multiplied by some con- stant,which will be takento be time independent. has retained 80 per cent of its argon and 40 per The value of this constant will be the fractional cent of its helium is later completely outgassed helium loss,•H., divided by the fractional argon by metamorphism or fusion, the gas thereby loss•A. As an approximationwe might expecta evolved might be preferentially enriched in freshigneous rock to lose3/5 its Ite4 and 1/5 argon by a factor of 2. Thus, from an average its A4ø into interstitial pore spaces.Weathering igneousrock we might expect extreme 'available' and aliageneticprocesses can do little to increase (I-Ie•/A•ø)r=aratios of about3 and 20, depending the alreadyhigh yield for Ite 4 from the rock,but on whether the gasesrepresented the value at- they offerample opportunity for breakingdown tained by compleh•outgassing of a rock which potassium-bearingminerals and releasingmuch had formerly preferentiallylost helium or by the retainedargon. Thus, we mightexpect sediments low temperature diffusion of helium and argon to favor the outgassingof argon more highly out of a young rock. If the sourcerock differs than an equivalentfresh igneous rock. The rela- considerablyin its K, U, and Th abundances tive rates of escapeof helium and argon from from an average igneous rock, the resultant rocks at elevated temperatures and pressures available(I-Iea/A•ø)r=a ratio will, of course,reflect are poorly known. However,a more complete this. Gases originating in pure carbonate rocks loss of both gasesmight be expectedunder may have high values of R. conditionsof metamorphism.If a rock which A number of gas transport mechanismscould producea fractionation of argon and helium. If, during the time of gas accumulation,the helium TABLE 4. Helium and Argon Retentivity in and argon are transportedby solutionin conhate Some Common Rocks and Minerals* water,fresh water, or petroleum,the (He4/A40),,a ratio that we actually observein a natural gas Fractional Retentivity samplewill be influencedby the relative solu- Mineral or Rock Helium Argon bilities of thesegases in the transportingmedium. If diffusive equilibrium is attained, we would Quartz 0.33 expect the followingrelationship to hold: • Feldspar 0.25 0.75 Mica 0.50 1.00 Pyroxene 0.75 R'-- Ku.KA Ro = K,Ro (4) Magnetite 1.00 Hornblende 1.00 where R ø and R• are the atomic ratios of the "Granite 0.40 helium to argon in solution and in the gas Diabase 0.60 reservoirrespectively, and Kn. and K,• are the Henry's law constanH for helium and argon, * Referencesincluded Keevil, 1941; Hurley and respectively. The Henry's law constants for Goodman,1941; Wasserburg, Hayden, and Jensen, 1956; and Goldich,Baadsgaard, Nier, and Hoff- severalgases at various•emperatures are given man, 1957. in Table 5. Rakestrawand Erareel[1938] showed 290 ZARTMAN, WASSERBURG, AND REYNOLDS

that argonis only about 80 per centas solublein of certain potassium evapori•e salts such as marine water of normal salinity as it is in fresh sylvite (K --• 50 per cent) and somevery pure water. Similarwork by z[kerlof[1935] on the carbonaterocks (K <( 0.01 per cent), the potas- effect of salinity on helium solubility suggests sium content of most rocks generally showsless about the same behavior for this gas. The than an order of magnitudevariation. This maior solubilityof argonand heliumin crudepetroleum rock-forming element makes up from 1 to 4 is unknown, as is the importance of noble gas per cent of many sedimentaryand igneousrocks. tr•ansportby this means.in water solutionsfor However, uranium and thorium, which generally temperaturesof 50-80øC, K • ___•2.8, and, there- occur as trace elements, are subject to wide fore, the gas phase would be enrichedin helium variations in abundance,and the possibility of relative to argon as comparedwith the solution. local enrichmentsmust be considered.Indeed, Thus, if much more gas is containedin solution one of the early theoriesproposed for the occur- than occurs in the gas phase, the equilibrium rence of high helium wells attributed the high ratio of He/A in the gasreservoir would be 2.8 helium contentto an underlyinguranium deposit times greater than the 'available' ratio. A more [Rogers,1921]. If an order of magnitudeor larger rigoroustreatment of this problemis given later enrichmentin the helium contentof a natural gas in the paper. is affectedby this process,the gas should show The differencesin solubilityof variousnatural an abnormallyhigh (He/A)r•d ratio. The large gas componentscould in principle, through a increase in the abundance of a major rock- multistageprocess of solutionand effervescence, forming element such as potassiumas would be cause considerable variations in R. It would be needed to maintain the observed (He/A)r•d possible under ideal conditions to produce a ratios could not occur in common rocks. That range in R of over 20 by employingonly three such an enrichment in radiogenic helium over stagesof distillation.It is not possibleat present, radiogenicargon doesnot exist in the caseof the however, to say how effective multistage distil- Texas Panhandlegas field was shownby Wasser- lation actually is in achieving variations in R. burg, Czamanske, Faul, and Hayden [1957]. That many stagesof distillation do not operate The resultsof the presentinvesfigafio• also is indicated by the limited range observedin R. tend to disprovethe idea of high enrichmentsin If compositionalvariations are ignored, it is helium due to abnormal uranium or thorium seenthat the effectsof the other parameterswill concentrationsfor a number of high helium gas permit over an orderof magnitudevariation in R. fields. Thus, the value of R actually observedin natural The problemof recognizingthe contributionof gas may range between extremes of at least gasesfrom the mantle or lower crust is extremely 3-50 without any need for assuming com- difficult. This is particularly true for helium positionaldifferences. All of the samplesanalyzed and argon, since the earth is a highly differen- have values of R that fall within the range of tiated body and the productionof theseelements 1.6 to 130. Most of thes8results agree qui•e well is dependenton the U, Th, and K concentrations. with this model, assumingonly minor variations The ratio R in modern material of chrondritic in the U/K ratio. It is, of course,possible to compositionis about 1; the ratio R in an average attribute the observed variations in R to com- igneousrock 4.5 billion years ago would be 2.0 positional differencesinstead of diffusion and as compared with the present value of 6.8. solubility effects. It is not possibleat present Someof the lowestvalues of (He•/A•o)r• were to say preciselywhich factors account for the in CO• wells that are associatedwith igneous variation. activity. These values are in the direction Damon and Kulp [1958] have measuredthe effected by great age or chondrific production ratio of excessradiogenic helium to argon in rates. These results, of course, are only sug- beryls and cordierires. They obtained ratios gestiveinasmuch as they could be producedby ranging from 0.5 to 130 with an averagevalue a variety of mechanisms. of 20. Our results on well gasesare thus rather It should be noted that the argon from the similar to the values fpund for these trapped CO• wells is extremely radiogenic.If this repre- magmatic gases. sents juvenile argon, it indicatesthat very little It must be pointedout that, with the exception A • is associatedwith it. This is compatiblewith HELIUM, ARGON, AND CARBON 291 the resultsderived from the solubility model for event, the important time factor would be the air argon. The latter data indicate that no length of the time interval over which gas significant amount of atmospherictype argon migrationtook place. occurs in these gases above the amount that Once the gaseshave been removed from the would be present from original equilibration crystal lattices they become available for with air. We infer from this that no deep-seated migration either by solution or by gaseoustrans- gases are contributing significant amounts of fer. Under equilibrium conditions,the distribu- argon of this composition. tion of any gas between a gaseousand liquid phase is related approximately by Henry's law. NOBLE GAS ABUNDANCES IN NATURAL GASES It is, of course, questionablethat equilibrium The helium content of the gas samples in- conditionsprevail over large distancesbetween vestigated varies between 37 and 62,200 ppm the accumulated gases and interstitial pore and the radiogenicargon content varies between fluids. It will, however,be assumedin the follow- 3.7 and 5580 ppm. Sinceany attempt to explain ing discussionthat diffusiveequilibrium obtains the occurrenceof the rare gasesin natural gases for the rare gasesand the consequencesof such must account for the absolute amounts and a model will be investigated. concentrations as well as for the ratio of radio- Equilibrium model. Goryunov and Kozlov genichelium to radiogenicargon, factors influenc- [1940] have pointed out the importance of ing helium and argon abundances will now be solubility phenomenain natural gas accumula- discussed. tion, and part of the followingdiscussion parallels Whereasthe (He/A)r• ratio is only weakly their work. time-dependent over times comparable to the Let us consideran equilibrium reservoirmodel age of the earth, the actual productionof these in which the rock has a porosityp, and suppose gasesis strongly time-dependent.The radiogenic the pore spaceto be occupiedby both an aqueous helium and argon content of a natural gas phase(s) and a gasphase (g), all under a hydro- reservoir is not necessarilyproportional to the static pressureof P atmospheres.Let the volumes age of the source rock, but is, rather, a com- occupied by the gas and aqueous phase be plicated function of the accumulation history V, and V', respectively,and the concentrations of the gas. It is possiblethat much of the radio- of gasspecies i in eachof thesephases be C•, and genic gases are incorporated into the natural C•,, in units of standard cc per cm8. The total gas by a sweeping-upeffect during the time of amount of speciesi in the pore system is then migration from the sourceto the reservoir rock. CdV' • C•'V •, and the fraction fd of this In such an event, the He and A content of the speciesin the gas phaseis rocks at the time that they were traversed by C•• V • the accumulatinggases would be an important factor. If there was little noble gas escape = c, v + c,'v' (5) before gas migration, the length of time between Supposethat each speciessatisfies a Henry's rock formation and petroleum accumulation law relationshipof the form C•, - K•C• • where would determine the concentration of He and A K• is the Henry's law constant for speciesi. in the pore space.Studies of a number of oil and We then have gas fields have shown that this time between sourcerock depositionand petroleum migration may vary from between tens of millions and hundreds of millions of years. The outgassing of very old basement rocks through meta- morphism would be a possiblesource of high =1/(1-]-K,V'••) (6) concentrationsof these radiogenic gases. Such Since KH.---• 110 and KA ----- 40 for aqueous gasesmay leak into a sedimentarysection through solutionsat 50øC,fH. • and fA• will be closeto 1 various fractures or faults. It may also be true when more than about 5 per cent of the pore for somecases that the noble gaseshave escaped spaceis occupiedby a gaseousphase. continuouslyfrom the rocks during all times In the units used, Cd is (assumingideality) except the period of gas accumulation.In this numerically equal to the partial pressure in 292 ZARTMAN, WASSERBURG, AND REYNOLDS

atmospheresof speciesi. The condition that a to be due to the releasefrom connateor ground pure gas phase i exists when the hydrostatic waters which were originally saturated with pressureis P is P,/K• -- C••. It followsthat for argon under atmospheric conditions. If the any given concentrationC• • there is a maximum system is closedsubsequent to burial, we have depth at which a gas phaseof pure i may exist. If we consider an infinitesimal helium bubble at C•'øV'ø = C.'V' + C.•V • 1000-foot depth under a hydrostatic pressureof 30 atmospheres pressure, this will require (s) Cue• --__-0.3 cc STP/cc pore vol. If we assume the porosity to be as low as 10% this will cor- wherethe subscripta representsair argon. C,•,o is respondto 3 )< 10-• cc STP of He per cc of rock. the initial concentrationof air argon in solution As will be shown later, such a value is obtainable and V ,0 is the initial volume of the solution. only under extreme conditions.It is, of course, If we assume that the volumes of the initial obvious that if, in a certain environment, and final water bodies are equal, we obtain radiogenichelium had associatedwith it another componentnot subsequentlyremoved, we would -'7V = K.•C,••ø-- C,•" (9) not find a pure helium gas. If we let C• equal the •nean concentrationof In a diffusiveequilibrium model, the systemis gasspecies i in the total pore spaceof the system, assignedvalues of V ' and Vg. This is, of course, then, from (6) we have unrealistic, since the natural system will not have sharply defined boundaries. The ratio c, = - - (7) V•/V • as calculatedwill thereforeapply for some From (7) we see that if all of the gas speciesi effective volume over which equilibrium is occursin the gas phase the mean concentration attained. Under atmosphericconditions at 15øC will simply equal the concentrationin the gaseous and in equilibrium with ocean water of normal phase. If, however, because of the rather in- salinity, we have C,,'ø • 3.0 X 10-• cc STP/cc soluble nature of some gases,we have most of H•.O [Rankama and Sahama, 1950]. The value the gas dissolvedin an aqueousphase with only of C• is determinedby the measuredconcentra- a very small gas bubble in equilibrium with it tion of air argon in the natural gas at the well (f?----- 0), we gain a factor of K• in C•gover the pressure.Substituting this expressionfor V,/V • mean pore space concentration. Thus, an in (6), and using K• = 53 (for 50øC and normal infinitesimal bubble of helium in equilibrium marine salinity) we have for radiogenic argon with a liquid at 50øCwill have a He concentration 1 of 110 times higher than the new concentration. IAg $ As will be shown below, most reservoir gases appear to have significantamounts of gas in a K•. V" dissolvedphase. The previous model has treated only a gas = 1 -- K•.C,,,o•___1 -- 630•• (10) with an aqueousphase. For gasesassociated with a liquid petroleum phase, the equilibrium rela- and for other arbitrary species tionship involving this phase in addition to an 1 I, • = • (11) aqueousand a gaseousphase must be considered. C• K,• This will obviously increase the number of variables in the model caluclations. The effect of l + K•(K•.C•O_C•) water salinity must also be taken into account We see that for the assumed equilibrium in a more precisecalculation. model it is possibleto calculate the fraction of In addition to the radiogenicnoble gas content speciesi which is in the gas phase. This may of natural gases,argon having the compositionof be applied to other gasesas well as to radiogenic present-dayatmospheric argon is found to be He and A. In Table i are given the calculated present in all of the gas samples.As discussed values of V'/V ,• and f• for the fraction of previously, the origin of this air argon is un- radiogenicargon occurringin the gaseousphase. certain. For purposesof discussion,we assumeit Because of the many obvious uncertainties ttELIUM, ARGON, AND CARBON 293

regardingthe assumptionsinvolved, it is doubtful All the parametersgiven in (14) are expressed that the calculationsare strictly applicableto as the effective values for the system under real natural gas reservoirs.Nonetheless, such investigation. Substitution of reasonable esti- considerationsare usefulas a meansof comparing mates for these parameters yields radiogenic the behavior of natural systemsto the simple heliumand argonconcentrations in natural gases idealized model. Usually the calculatedvalues quite consistentwith actually observedvalues. of fAo and V'/V• appearto be quite reasonable. For example, let us calculate the radiogenic Only two samples(11 and 24) give impossible helium and argon content of two extreme reser- values of fA•, and these discrepanciesare not voir cases.Let us assumea rock density of 2.5 extreme. and an effective U, Th, and K concentrationin If there were originally entrappedair bubbles the rocks of 3.5 ppm, 10 ppm, and 2.6 per cent, in the pore space,or if V' < V 'ø was due to respectively. In the one instance, let us set hydrarich reactions in diagenesis,then the v -- 5 X l0 s yr., •jUo= •j, --• 1, fUog •-' fig•-•- 0, calculatedvalue of Vø/V•, in (9) will be larger P = 10 atmospheres(147 psia), and p = I per than the true value. The calculated values of cent. This correspondsto an environment in f• will be too small, owing to such effects. which the total radiogenic noble gas content Next, let us look at the effect of porosity, p, producedby one-half billion years of decay is on concentration. If we assume the mean noble completelyreleased into rock of 1 per cent mean gas concentrationin the pore space, C•, to be porosity having a hydrostatic pressureof 10 esserrtiatt ' - tmospheres:GeologicM!•this could be brought centration producedin the mineral phaseper cc about by the continuousrelease of radiogenic of rock, we have gases into overlying sediments of one-half billion-yearage, or by the completeretention of C,= •-•N, (12) such gases in basement rocks over this time P interval, followed by some metamorphic event whereC•, as before,has the unitsof cc STP/cc which then releasedthe gas. This former possi- of porespace, N• hasthe unitsof cc STP/cc of bility would appear unlikely where no sedimen- system, and • is the rock degassingfactor. tary cover of sufficientage was presentto trap The concentration of the noble gases in the the continuouslyreleased gas. The gasis virtually pore spaceis inverselyproportional to the poros- all dissolvedin an aqueousphase, and thus the ity. N• for radiogenichelium and argonis given concentrationin the gas phase is increasedover by (2a) and (2b), respectively,calculated per cc the mean pore space concentration by the of rock. The molecular abundancein ppm of numerical value of the Henry's law constant. speciesi in the gas phaseis I'•, These conditions yield Fuø •--- 106 ppm and F•___ 5 X 104ppm. For contrast, let us considera case in which F,- p X 106 (13) r = 5 X 107years, •Ue•'--•,•-•- 0.5, fuoø•f•.o •-•- 1, where P is the pressure of the reservoir in P = 100 atmospheres(1470 psia), and p = 5 atmospheres.This equation assumesthe gases per cent. This correspondsto a situation in to be perfect. For hydrostatic pressureP •--- which one-half of the radiogenicnoble gas pro- 0.030h,where h is reservoirdepth in feet below duction over a 50-million-yearperiod is released the surface. into rock of 5 per cent mean porosity having a Combining (2), (7)' (12), and (13), we have hydrostatic pressureof 100 atmospheres.These conditionsyield I•Ho-- 10 ppm and Fx = 1 ppm. •H.(0.120 U --I--0.029 Th)r If, instead,the gaswere mainly dissolved(f--, 0), rile P.p the value of FHo would become a factor of 100 larger. ß(Kuo- [Ka•- 1]]a•) (14a) An average reservoir might have effective •^(3.99 X 10-6K)r valuesof •' = 10• years,•Uo = 4/5, • = 1/2, FA -- P.p fH,o • 0.9, f,•o •___0.75, P --' 50 atmospheres (735 psia), and p = 3 per cent. Assumingthe ß(K.•- [K.•- 1]]•) (14b) sameU, Th, and K abundancesand rock density 294 ZARTMAN, WASSERBURG, AND REYNOLDS as above, we have FHo = 1040 ppm and FA ---- of the reservoirrock is 2 per cent, the effective 86 ppm with R---• 12. In this case we have porosity of the entire sourcerock system is 80 per cent and 50 per cent of the accumulated much lower. This might be expectedif most of radiogenic helium and argon, respectively, the helium were derived from a very tight resulting from one hundred million years of basementcomplex. It is difficultto tell which decay being releasedinto rock of 3 per cent are governing factors in this case. At present, mean porosity. The values of f? correspondto we surmise that this gas has probably been havingV./Vo_____. 10. producedby the favorableconjunction of several In order to make an estimate of the volume of these effects. It should be emphasizedthat of rock swept out during the formation of a it is rather difficult to account for this factor g•s reservoir,let us considerthe helium-producing of 10; if an additionalfactor of 10 wererequired, zone of the Rattlesnake Gas Field, San Juan it wouldbe impossibleto obtainthis reasonably County, New Mexico. W. M. Deaton, chief in terms of the presentmodel. helium consultant for the Bureau of Mines If we comparethe San Juan high helium gas Helium Activity (personalcommunication) esti- with that from the Texas Panhandle, we see mates the total original volume of the reservoir that the formergas not only containsmore than gas to be about 2.4 X 109cubic feet at 15.0 psia a fivefold higher abundancein helium, but also and 60øF. About half of the total gas has been a 35 times greaterhelium concentration,due to removed,and the field is not being producedat the higher pressureof the San Juan gas. the presenttime. The gas, which has a helium It shouldbe pointedout that a gasthat existed content of 7.6 per cent, was producedfrom the at rather shallow depths with a particular Leadville-Ouray (Mississippian-Devonian)form- abundanceof helium and subsequentlytrans- ations and had an initial pressure of about portedas a' closedsystem to a greatdepth will 3000 psia. Thus the reservoiris approximately retain the samehelium abundance,but it will 1.2 X 107cubic feet, and assumingp: 2 per cent have a muchhigher helium concentration. Such and V'//V g• 94 (approximatevalues of near-by a gaswill, of course,not be in equilibriumwith Navajo C-1), we see that this correspondsto a the surroundingaqueous reservoir and will volume of V'/p of 5.6 X 10•ø cubic feet, or ultimatelyreturn to a lowerhelium concentration. 0.4 cubic mile. Assuming an average uranium There is no evidencesupporting such a trans- and thorium abundance in the rock •Ho --- 1, portationhistory for the San Juangas. and an effective v of 3 )• 10s years, we calculate We have seen that under equilibrium con- from (14) that Puo---• 7000 ppm. This is lower ditions the concentrationof a slightly soluble than the observed value by a factor of 10. gassuch as He or A in the gaseousphase is of Let us now look at somepossible explanations for the order of 10s times its concentration in solution. this discrepancy.Since the gashas a (He/A)r•d The uniform release of these gases from a ratio which is characteristic of common rocks, homogeneoussource rock into a systemwhich we cannot reasonablyassume that either the containsboth poreliquid and gaswould, in the uranium or thorium concentrations, or both, absenceof completediffusive equilibrium, tend in this sourceare abnormally high, although a to make C•g < K•C•'. Under such conditions, factor of 2 increasemay well occur. Although the concentrationof speciesi in the gas phase the gas reservoiris in Pennsylvanianstrata, it is would be lower than expected by diffusive possiblethat the radiogenicfraction of the gas equilibrium.Except for minor effectsbrought was chiefly derived by the outgassingof much about by the temperaturedependency of the older basement rock containing a correspond- Henry'slaw constant,it is difficultto envision ingly higher helium concentration.Also, the a natural situation which would tend to make possibilitythat our equilibriumsolubility model C•,' > K•C• o. With reference to radiogenic is incorrect would allow for a twofold increase helium,only if a sourcematerial is selectively in the helium concentrationin the gas phase contributingthe helium to a gas phase at a above that calculated. A favorable combination rate of KHo times faster than it is feeding a of these three effects could just allow for the liquidphase can C•ø•be greaterthan observed helium concentrationin the gas. In Radon. The existence of radon in many addition,it is possiblethat whereasthe porosity naturalgases has sometimesbeen used to infer HELIUM, ARGON, AND CARBON 295 the presenceof a concentrationof uraniumin will considergases as incompressible,and take the neighborhoodof suchoccurrences. Faul, Gott, the rate of productionQ in units of cc volumeat Manger, Mytton, and Sakakura [1954]have re- well pressureper unit time. ported on the Rn contentof somehelium-rich For the caseof the uniform productionof Rn natural gases.These workers conclude that it is within a cylinderof radius R centerabout the uncertain whether the high helium content of origin and no productionoutside of this region, certain gasesis related to the presenceof Rn. the value of C at the well is This problemis of interest in consideringthe possibilitythat a significantportion of the He in somenatural gasescan be the productof U, C(0)= q/X1 - exp Q I (16) Th decayin their presentreservoir. It is clear that C(0) is insensitiveto changesin Sakakura, Lindberg, and Faul [1959] have the parametersat distancesmuch greater than reportedthe radon contentin the gasesfrom r-- (- Q/X•-Hp)i, i.e., distancesfor which four wells as a function of the cumulative the time of travel to the well is equal to the production,star•ing from a situationwhere the mean life. This distance is probably not in wells had been shut down for 2 or 3 weeks. excess of a few hundred feet for most wells. The net gas flow required before the Rn con- The highestRn concentrationin the Panhandle centration reached a constant value was small helium wells reported by Faul, Gott, Manger, and of such a value as to indicate that the Mytton, and Sakakura [1954] is 500 micromicro- principalsource of the activity was not in the curiesper liter at STP. Usinga well headpressure well hole itself but immediatelyadjacent. These of 16 atmospheres we obtain for a uniform workershave presenteda theoreticaltreatment sourceq = XC-----0.24 Rn decay/sec.cc of pore of the problemof the gastransport of Rn. They space.Assuming a porosity of 10 per cent, this conclude that for the Texas Panhandle gases corresponds to an emanating uranium con- the Rn is due to a uranium concentration of centrationof 10-8 g U/g rock.If but I per cent between0.4 and 9.0 ppm,assuming an emanating of the emanation escapesto the pores, this power of 10 per cent. correspondsto a uranium concentrationof only In the following,we will presenta simplified 100 ppm. For a well in the San Juan basin treatment of the steadystate transportproblem containing6 per cent He, Faul, Gott, Manger, that will sufficefor the purposeat hand. Mytton,and Sakakura[1954] report a Rn activity The equation governing the steady state of only 5 micromicrocuries/litersSTP. concentrationof Rn in a fluid phase for the In some of the casesreported, the Rn level caseof cylindricalsymmetry is givenby was sufficientto accountfor the helium present, assumingsteady productionfor 10s years. In = --kpC(r) -•- pq(r) (15) others, it could contribute only 0.1 per cent r dr of the He. Since more He may escapefrom the Here C(r) is the concentrationof Rn in the reservoir rocks than is indicated by the Rn fluid phaseat distancer, J(r) is the outward concentration, it is quite possible that some radial volume flux of fluid with the dimensions gasesobtain their helium after their finale_ntrap- of velocity,p is the porosity,q(r) is the rate of ment. In any case,the Rn data do not support generationof Rn per cc of pore space, and the case for the generation of these helium •, is the decayconstant for Rn. If the fluid is containing gasesfrom high uranium concentra- incompressible,J(r) = Q/2•-Hr, with -Q/H tions. being the volume of fluid yielded by the well Helium distributionfunction. There are about per unit heightof the producinghorizon. 3400helium analyses for natural gasesas reported If the productionrate of Rn is everywhere by the U.S. Bureau of Mines [Andersonand constant,we seethat C = q/•. The Rn activity Hinson, 1951; Boone,1958]. The gasesanalyzed is then constant and equal to the production comefrom a wide variety of geologicalsituations rate. If the fluid is compressible,the concentra- and, excluding the very high helium wells, tion is a more complicatedfunction of position shouldbe a representativesample of natural gas due to the fact that at lower pressuresthe fluid accumulations.Because of the large number of occupiesa greatervolume. For simplicity,we analyses,it shouldbe possibleto get a rather 296 ZARTMAN, WASSERBURG, AND REYNOLDS

i•iO-2

6

IxlO-3 I , ,I I I I I I I ! I I 1 I I I 0.01 01 i 5 I0 20 40 60 80 90 95 99 99.9 99.99 A(x)

Fig. 3. Log probabilityplot of the helium abundancein natural g•ses.Statistics for abund- ancesless than 0.8 per cent and estimatesfor high heliumproduction are taken from U.S. Bureau of Mines data. good descriptionof their frequencycurve. The productionhas an averageHe concentrationof data used were 3000 analysesof samplescon- 1.5 per cent [U.S. Bureau of Mines, 1959]. raining less than 0.8 per cent He. Becausean Using these data, a frequencyhistogram was excessivenumber of samples were analyzed constructed.The cumulative data, plotted on from known He-producing areas, we have log probability paper, are shown in Figure 3. discardedall analysesreporting over 0.8 per cent It is evident that a lognormal distribution He. Using estimates of the total annual gas representsthe data fairly well. The percentties productionand theirhelium content, we obtained were determined [Aitchisonand Brown, 1957]. the result that 3 per cent of the total gases usingthe straightline drawn throughthe points. contained helium in a concentration of over The following values were determined for this 0.8 per cent. This fraction of the annual gas curve: HELIUM, ARGON, AND CARBON 29?

mean -- 0.2170% gases is among the lowest observed in any analyzedsamples. The atmosphericargon content median ----0.0610•o •r -- 1.59 of these gasesranges between 44 and 124 ppm; mode- 0.0049•o /• - 2.80 the nitrogenvaries from lessthan I per cent to over 30 per cent. Geologicevidence indicates Figure 4 illustrates the frequency curve and that the reservoirrocks may be in communication usesthese parameters for the frequencyfunction with ground water circulation. These Tertiary and Cretaceous sediments crop out a short distance to the east along the western foothills dA(x)dx -- 2•ro'xI exp-- [(ln[ .•/•x-- (17)of the Sierra Nevada. The young age of the sedimentsmay account for the low radiogenic This curvemay be comparedwith the histogram. noblegas contentof thesenatural gases,and the The histogram does not appear to have possibilityof open systemconditions or hetero- bimodal characteristics. This observation further geneity in the sourcematerial is suggestedby suggeststhat the high He gasesdo not (in a the large variationsin the (He/A),• and statisticalsense) represent any specialmechanism, (N•/A, ir) ratios. Chemically,the Sacramento but rather representlow-probability events on Valley gasesare extremely dry, with methane the tail of a continuous-probabilitycurve. making up essentially all the hydrocarbon fraction and nitrogen varying between 1 and SAMPLES 2n no.oo•f. mh..+...11f.hi.• nitrogen cannotbe nr Samples1-7 are gasesfrom Eoceneand Creta- direct atmospheric origin has been discussed ceoussands of the SacramentoValley, California. previously.Sample 6 from the Marysville-Butte They have a wide range in (He/A),• and field was taken from gas closelyassociated with (N•/A•i,) ratios, and their e is rather low. The Pleistocene volcanic extrusives. radiogenichelium and argon content of these Samples8-10 and 36-39 are gasesoccurring

26

24

22

2O

0 - I -2 -5 -4 log X Fig. 4. Frequencycurve and histogramshowing the distributionof helium in natural gases. 298 ZARTMAN, WASSERBURG, AND REYNOLDS in rocks of Cretaceousto Eocene age from has selectivelyenriched the gasphase in helium. northwestern Colorado and southwestern The low value of Vo/Vg may in reality point Wyoming. Although these gasesalso show wide to the removalor absenceof air argonfrom the variationsin (N•/A•ir) and e, their (He/A)r•a system. If the high value of R did result from ratio remains fairly constant despite over an a distillationmechanism operating on an origin- order of magnitude variation in radiogenic ally normal (He/A)•,a ratio, there shouldexist helium and argon content. Whereas the chief somewherea complimentary phase having a factor producing the spread observed in the correspondinglylow value of R. It is alsopossible (N•/Asi•) ratio for the SacramentoValley gases that the sourcesof the western Appalachian was a highly variable nitrogen content, the gasesare characterizedby a low K/U ratio. Green River gases,which demonstratean even Whereassuch a ratio might be expectedto occur greater range in (N•./Asi,) ratios, have more in impure limestones,the fact that most of the uniform nitrogen content, but over an order of Appalachiangases occur in stratigraphictraps, magnitudevariation in air argon content. Thus, involving sandstonelenses surrounded by im- thesegases may representeither (1) reservoirsof perviousshales, seems to excludethis possibility. highlyvariable ¾ø/¾•, (2) reservoirsint• which In no way does this region demonstrate a air argon from sources other than saturated (He/A)•a ratio to be expectedby a simple seawater was introduced,(3) reservoirsin which degassingof a chondriticmantle. varying degreesof diffusivedisequilibrium exist, Samples 11 from the Sexsonfield, Kansas, or (4) sampleshaving incorporatedair argon 22 from a wildcat well in San Juan County, New owing to varying levels of contamination. Mexico, 24 from the Keyes field in Oklahoma, Samples12 and 13 were selectedt• represent and 29-32 from the West Panhandlefield, Texas, gasesfrom a present-day geosynclinalenviron- were selected as representative high helium ment. natural gases. They range between 0.5 and Samples 18-20 are from Lea County, New over 6 per cent helium with (He/A),,a ratios Mexico. The first two gasesare producedfrom ranging from 11.1 to 23.3. Such values of R are Permian sandstones.Sample 18 comes from a in no way compatible with a large increasein reservoir directly overlying the reservoir of the concentration of uranium and thorium over sample 19 and is stratigraphically separated potassiumas comparedwith averagerocks. As from it by about 1000 feet of section.The other stated elsewhere,it is the authors'opinion that gasoccurs in a Permiancarbonate. The similarity high helium wells result from the favorable in the (He/A)•,a ratio, the (N•/A,i•) ratio, and interaction of several processesoperating on e for thesesamples from the sandstonereservoirs normal rock types having averageU, Th, and may be a reflectionof a similar source.The gas K contents.Some of thesehigh heliumreservoirs, from the limestx)nereservoir appears to be such as those of the Texas Panhandle field, distinctly different in these parameters from appear to be closely related to buried Pre- the other two gases. cambrian granite and granite wash, which may Sample23 fromwestern New York and samples have been the sourceof the helium [Cornerand 25-27 from Pennsylvania represent Paleozoic Crum, 1935]. In other areas, such as San Juan gasesfrom the Appalachianpetroleum province. County, New Mexico, no such correlationwith Thesegases contain the highest(He/A)•a ratios basement rock is obvious tHinson, 1947]. In found in any of the samples;they range from many cases,however, the high heliumreservoirs 20.2 to 134. Sucha high ratio is quite surprising are stratigraphically low in areas containing in these gasesbecause of their relatively great severalreservoir horizons. This wouldbe expected age, low ¾ø/¾%and proximity t• orogenic if helium, having been liberated from the base- activity, which may have been more effective ment, rose through the overlying sediments than averagein releasingradiogenic argon. All until it became incorporated into the first of thesefeatures point toward a low (He/A)r•d reservoirit encountered.Numerous exceptions to ratio. the occurrenceof high helium gases in low On the other hand, it is possiblethat some stratigraphichorizons exist. Rogers[1921] found distillation mechanismhas been particularly that in the mid-continentregion it may be that operative in the productionof these gasesand certainintermediate horizons, or eventhe highest HELIUM, ARGON, AND CARBON 299 producingzone, contain the largestpercentage of Sample 40 from Texas and sample41 from helium. It is still not possibleto give a rigid Alberta, Canada, were includedto represent explanation of the processesby which certain gasesfrom carbonatereef environments.They reservoirsbecame highly enriched in helium. show no indication of an abnormally high These gasescontain between 50 and 200 ppm (He/A)•,aratio as mightbe expectedfrom pure of atmosphericargon. A comparisonbetween carbonatesources. Sample 41 contains12.2 per the A4o/A86 ratio as given by Wasserburg,cent H,S. Czamanske,Faul, and Hayden [1957] and by this paper (sample31, Table 2) showsa much CARBON ISOTOPES greater enrichment in radiogenic argon for our analysis.This is most likely due to atmos- Isotopic analyseswere made on some of the pheric contaminationin the older analysis. All carboncompounds contained in the gases.The of the Texas Panhandle gasesnow have similar results are shown in Table 3. The isotopic A4o/A86 ratios, and this is believedto be a compositionof the carbon is given in terms of regional characteristic.All of the high helium its delta ($) valuerelative to the Chicagostand- gaseshave (Ns/A,i,) ratiosof an orderof mag- ard, PDB [Craig, 1953]. nitude or morehigher than the atmosphericvalue. Most of the Californiasamples were run only Sample 22 was observedto have a minimum as total gas;however, since they generallycon- possiblevalue of (A•ø/A•s),•dof 1.9 X 105. tainedless than 0.1 per centCOs, and very little +• total '•'•-'•'^-' ' ' This value is iarger than the maximum value predicted by Gerling, Levskii, and Afanasyeva nearly equal to the methane carbon. Several of [1956] by a factor of 2.4 and, therefore, adds the TexasPanhandle and the Keyes,Oklahoma, additional evidence[Wasserburg and Bieri, 1958; highheli• gaseswere also run onlyas total gas. Signerand Nier, 1959] againstthe presenceof a A comparisonwi•h samples29 and 30 from the long-lived isomer of K 3s. Texas Panhandle area, in w•ah both •tal Samples 16, 17, 34, and 35 were chosen to carbonand methanecarbon were run, indicates include carbon dioxide gasesin the survey. The that about a 1-2 per • differencemay be Bueyerosfield gases, which are over 99.8 per cent expectedbetween total gas and methane carbon CO,, showthe lowest (He/A)r,d ratios of any for this suite, with the methane having the gasanalyzed. They alsocontain less than 0.5 ppm lighter carbon. The remainder of the natural atmosphericargon, and have ds of 0.98-0.99. gaseswere analyzed for methane carbon, and The Farnham Dome gasescontain approximately in some casesan additional analysisof total 99.3 per cent CO,. They also have quite low gas carbon was made. The carbon dioxide well (He/A)r,d ratios and high values of •. Both gaseswere analyzed for CO• carbononly. Carbon these gas fields occur in areas containing car- dioxidewas analyzed from all naturalgas samples bonate rock and Tertiary basaltic intrusives contatung more than 0.2 per cent CO• •th and lavas. Geologic evidence, as well as carbon the exceptionof sample24. AH of the methane isotopic data which will be discussedlater in analysesgive delta values between --57.6* and this report, suggeststhat the COs was derived --29.2 per mil. Both these extreme values are from the decompositionof carbonaterocks during from wells within the SacramentoValley gas metamorphism by the basalt. Under such fields of California. Usually individual gas- conditions,it is possiblethat most of the original producing dist•c•s possessa much narrower gases and liquids in the near-by sediments range in methane carbon composition.This is were driven off before the carbon dioxide ac- exemplified by samples 8-10 and 36-39 from cumulation.This would explain the high purity the Green River area and samples29-33 from of the COs and the extremely low atmospheric the Texas Panhandle. The methane carbon from argon content of the gases. Since the molten the Green River suite lies between --38.5 and basalt might causea relatively higher releaseof radiogenicargon, compared to radiogenichelium, * Although this analysisfor sample 5 refers to than results from low temperature diffusion, total gas carbon,it is from a gas essentiallyfree of COs and other hydrocarbons;therefore, it un- we could expect such an environment to yield doubtedly is within a few tenths of a per mil of low valuesof (He/A),•d. the methane value. 300 ZARTMAN, WASSERBURG, AND REYNOLDS

--43.5 per mil, and the total carbon from the contact metamorphism. The results from the Texas Panhandle gases suggestsan even nar- two CO•. fields studied in this report are in rower range. agreementwith the work of Lang. These results are similar to the observations The isotopiccomposition of the oxygenfrom made by Silverman and Epstein [1958] and three of the carbon dioxide wells was also Wasserburg,Gzamanske, Faul, and Hayden [1957], determined. Mitchell No. 4, Farnham Dome who found that methane has the lightest carbon No. 2, and Farnham Dome Equity gave values found in nature. The exact mechanismby which of (]co, relative to PDB of--23.0, --20.2, and methane is producedis not known at present. --20.3 per rail respectively.Such values might It is probablethat somebreakdown of organic be expected either for CO2 produced by the matter, either by bacterialor inorganicprocesses, high temperature decomposition of marine results in the liberation of methane. It was noted carbonate or possibly by CO•. in equilibrium by Silvermanand Epstein [1958] that natural with certain reservoir waters at appropriate gas carbon is generallylighter than associated temperatures.In the first case, if fractionation petroleum,and that this in turn is lighter than of only a few per mil occurredbetween the car- the organisms from which it is supposedly bonate and CO2 during such high temperature derived. Rosenreidand Silverman [1959] have processesas contact metamorphism, we would found an unusually high fractionationbetween obtain values of 5co• close to those of marine methanol and the methane which is produced carbonate, i.e., --20 to --10 per cent [Clayton from it by anaerobicbacterial decomposition. and Epstein, 1958; Engel, Clayton, and Epstein, The methane is about 8 per cent enriched in 1958] recomputedrelative to the acid extracted C • relative to the methanol. Further investiga- PDB standard. In the secondcase, we see that tion of the origin of methaneis requiredbefore a although the (] value is too low to represent morespecific discussion of its C•3/C• ratio can CO• in equilibrium with normal marine water be made. at room temperaturessuch values might result An analysiswas made of a sample of gas from equilibrationwith light watersor at elevated collectedby Louis Gordon from a gas-seepover temperatures. [Epstein and Mayeda, 1953]. the head of the north fork of Scripp's Canyon At least traces of carbon dioxide are observed off the coast of southern California near La Jolla. in nearly all natural gases,and gasesranging in The delta value for this sample was --44.2; compositionup to almost pure CO• have been it falls in the same general range as the well found. Someliterature [Lang, 1959; Miller, 1937] methanes.I•. O. Emery (personalcommunica- has been publishedon the origin of high carbon tion) considersthis gas to be a seep from a dioxide gases; however, no detailed work has subsurface source. been done on the CO•. in ordinary petroleum In thosegases where total methaneand CO• gases.Here, the CO• generallymakes up between carbonwere analyzed, it waspossible to calculate a few hundredthsof a per cent and a few per approximateranges of the • value for the higher cent of the total gas. hydrocarbons.Because small uncertainties in the The 5 valuesfor all the CO2 analysesare seen known • values are greatly increasedin such a to lie between --21.9 and •-12.8. In every case, calculation,it can only be concludedthat the the CO•. carbon is considerablyheavier than • value of the higher hydrocarbonsis generally that of the coexisting methane or total gas. 5-15 per mil greater than the methane. This is a remarkableregularity which must have If we consider the carbon dioxide gases, we important genetic significance. see that they are often associatedwith strati- If we assumethat the CH,-CO.,. pairs were in graphic sections also containing carbonate isotopic equilibrium within a gas phase under rock and igneousintrusives and lavas. Lang certain conditions,and that subsequentto that [1959]found that the C•3/C•2 ratiosof several time these values were 'quenchedin,' we may CO2 gaseslay within the range of carbon in calculateeffective temperatures. Such tempera- marine limestones. This is in agreement with tures were computed using the fractionation the theory that the CO• comesfrom the non- factor as calculated by Craig [1953]. These equilibrium dissociation of carbonate during results are given in Table 3. Since the kinetics HELIUM, ARGON, AND CARBON 301

of this exchange are completely unknown, we the genesis of the carbon dioxide found in are unable to say under what conditions we petroleums. would expect equilibrium to occur and be CONCLUSIONS frozen in. The possibleranges in temperature which we might obtain, assumingequilibrium All the natural gasesstudied exhibit values of to have occurredat sometime, are room tempera- R which indicate that they are of a common ture and the maximum temperaturewhich these family and have obtained their radiogenicgases gasesmay have experienced.Some of the cal- from rather averagerock types. The variations culatedtemperatures range up to severalhundred in abundance of radiogenichelium and argon degrees centigrade. This is considerably in in thesegases are due to the effectsof leakage excessof the actual well temperatures. Since and entrapment, solubility, porosity, and age it is generally agreed that petroleums do not of the source rocks. Major differencesin the result through high temperature processesof abundancesof He cannotbe due to large varia- formation, we concludethat someof thesegases tions (over a factor of 50) in the abundanceof U lie well outside the range to be expectedfrom and Th in the source rocks. equilibrium considerations. For many casesstudied, only a smallfraction of Although it is thus doubtful that complete the N,. presentmay be attributed to the incorpor- isotopic equilibrium prevails, it is impossibleto ation of air. say whether the CIt4-COs pairs represent a The methane was found to have extremely partial attainment of isotopic equilibrium. it is light carbon as suggestedby previous workers. significant that in every case of unreasonable In addition,it wasfound that the carbonisotopic temperature,the t]co, value appearsto be lower compositionof COs was always heavier than than the expected equilibrium value. That is, in the coexistingCHq. This could not be reason- if isotopic equilibrium were to be attained in ably attributed to completeisotopic equilibrium those gases, t]co. would have to become more betweenthese gaseous species. positive. Thus, assumingonly equilibrium pro- It is possible that the concentration of at- cessesto be operative between the COs and the mosphericargon in a gas reservoir may be CH4, the observedt• values would representthe correlatable with the amount of reservoir water maximumvalue which they couldhave originally with which it was equilibrated. had. Although nine of the samples listed in Table 3 have values of t•co. greater than --10 Acknowledgments.This work could not have been carried out without the generousassistance of per mil and, therefore, may have possibly the many peoplewho aidedus in obtainingsamples. arisenby either the equilibriumor disequilibrium In particular, we would like to thank the members breakdownof marine carbonaterock, the other of the following organizationswhich supplied us seven samples for which we have data give with material: The U.S. Bureau of Mines Helium Activity Station, The Mountain Fuel Supply Co., $ values too negative to representsuch a source. Skelly Oil Company, Carbonic Chemicals Corpo- If we assume that all the carbon dioxides once ration, Carbon Dioxide and Chemical Company, had $ values at least as negative as the lightest Humble Oil and RefiningCompany, United Natural sample now observed (•- --20 per mil) and Gas Company, Iroquois Gas Corporation, The that the presentspread resulted from later partial Sylvania Corporation, Standard of California and affliates, and The Ohio Oil Company and affiliates, equilibration, we must look for other sourcesof and Continental Oil Company and affiliates. carbon capable of yielding such values. We would like to thank Dr. S. Epstein for grant- When COs occurs in much smaller concentra- ing us the liberal use of his laboratory and for his tions than CH•, any changetoward equilibrium continuedinterest in this problem. We also wish to thank Dr. Sol Silverman and Dr. K. Chave who in the isotopiccomposition of the systemwould contributed severalvaluable suggestions.Mr. C. W. causea muchgreater change in the C•a/C•s ratio Mink assistedin the mass spectrometric work at of the COs than of the CHq. Thus, while such a Berkeley. processwill affect the original$c•, value only Contribution No. 978 of the California Institute weakly, even partial attainment of equilibrium of Technologywas supportedby a National Science Foundation Grant, and in part by a grant in aid will tend to obscureany original value of •co,. from the California ResearchCorporation and by a At present,it is not possibleto explainuniquely grant from the Atomic Energy Commission. 302 ZARTMAN, WASSERBURG, AND REYNOLDS

REFERENCES ages and the isotopic composition of argon from meteorites,Astrophys. J., 127, 224-236, 1958. Adams,J. A. S., J. E. Richardson,and C. C. Temple- Gerling, E. K., L. K. Levskii, and L. I. Afanasyeva, ton, Determination of thorium and uranium in On the discovery of A 38in potassium-containing sedimentaryrocks by two independentmethods, Geochim.et Cosmochim.Acta, 13, 270-279, 1958. minerals, Dokl. Akad. Nauk SSSR 109, 813-815, .•]•kerlof,GSsta, The solubilitiesof noblegases in 1956. aqueoussalt solutionsat 25øC, J. A m. Chem.Soc., Goldich, S.S., H. Baadsgaard, A. O. Nier, and J. H. Hoffman, The reproducibility of A4ø/K4ø 57, 1196-1201, 1935. age determinations, Trans. Am. Geophys.Union, Aldrich, L. T., and A. O. Nier, The occurrenceof hellium-3 in natural sources of helium, Phys. 38, 547-551, 1957. Rev., 74, 1590-1594, 1948. Goryunov, M. S. and A. L. Kozlov, Voprosygeo- Aitchison and Brown, The Lognormal Distribution, khumii gelienosnykhgazov i usloviia nakopleniia v University Press, Cambridge, 176 pp., 1957. semnoi kote., Russian (State Sci.--Tech. Pub. Co. Oil and Solid Fuel Lit.), Leningrad-Moscow, Anderson,C. C., and H. H. Hinson, Helium-bearing 1940. natural gasesof the United States, Bureau Mines Bull., 436, 141 pp., 1951. Hamaguchi, H., G. W. Reed, and A. Turkevich, Barahoy,V. L., A. B. Ronov,and K. G. Kunashova, Uranium and barium in stone meteorites, Geo- Geochemistryof thorium and uranium in clays chim. et Cosmochim.Acta, 12, 337-347, 1957. and carbonate rocks of the Russian platform, Hill, R. D., Production of helium-3, Phys. Rev., 59, Geokhim.,Izd. Akad. Nauk SSSR, 3, 3-8, 1956. 103, 1941. Bate, G. L., J. R. Huizenga and H. A. Potratz, Hinson, H. H., Reservoir characteristicsof Rattle- Thorium in stone meteorites by neutron acti- snake oil and gas field, San Juan County, New vation analysis,Geochim. et Cosrnochim.Acta, 16, Mexico, Bull. Am. Assoc. Petrol. Geol., 31, 731-771, 1947. 88-100, 1959. Bell, K. G., Uranium in precipitatesand evaporites, Hoering, T. C., Nat. Acad. Sci., Nat. Res. Council U.S. Geol. Survey Pro[ess. Paper 300, 381-386, Pub., 572, Nuclear ScienceSeries Report No. 23, 1956. 161-170, 1957. Boone, W. J., Jr., Helium-bearingnatural gasesof Hurley, P.M., and C. Goodman,Helium retention the United States,Bur. Mines Bull., 576, 117 pp., in common rock minerals, Bull. Geol. Soc. Am., 1958. 52, 545-560, 1941. Cady, H. P., and D. F. McFarland, Helium in Joly, J., The amount of thorium in sedimentary Kansas natural gas, Science,24, 344, 1906. rocks, II, Arenaceous and argillaceous rocks, Clayton, R. N. and S. Epstein, The relationship Phil. Mag., 20, 353, 1910. between O•s/O•6 ratios in coexisting quartz, car- Keevil, N. B., Helium retentivitiesof minerals,Am. bonate, and iron oxidesfrom various geological Minerol., 26, 403-404, 1941. deposits,J. Geol.,66, 352-373, 1958. Lang, W. B., The origin of some natural carbon dioxidegases, J. Geophys.Research, 64, 127-131, Cotner, Victor, and Crum, H. E., Geologyand oc- 1959. currenceof natural gasin Amarillo district, Texas, Geologyof Natural Gas, Am. Assoc.Petrol. Geol., McCrea, J. M., Ph.D. Thesis, University of Chi- 385-415, 1935. cago, 1949: On the isotopicchemistry of carbon- Craig, Harmon, The geochemistryof the stable ates and a paleotemperature scale, J. Chem. carbonisotopes, Geochim. et Cosmochim.Acta, 3, Phys., 18, 849-857, 1950. 53-92, 1953. McKinney, C. R., J. M. McCrea, S. Epstein, H. A. Damon, P. E., and J. L. Kulp, Excesshelium and Alien, and H. C. Urey, "Improvements in mass argonin beryl and other minerals,Am. Mineral., spectrometersfor the measurementof small dif- ferencesin isotope abundance ratios," Rev. Sci. 43, 433-459, 1958. Eklund, Josef,Urantillg•ngar och Energif6rs6rjning, Instr., 21, 1950, p. 724-730. Kosmos(Swed.) 24, 74, 1946. Miller, J. C., Carbon dioxide accumulationsin geo- Engel, A. E. J., R. N. Clayton, and S. Epstein, logic structures, Am. Inst. Mining Met. Engrs. Variations in isotopiccomposition of oxygenand Tech. Pub., 841, 28 pp., 1937. carbon in Leadville limestone (Mississippian, Morrison, P., and J. Pine, Radiogenicorigin of the Colorado) and in its hydrothermal and meta- helium isotopesin rock, Ann. New York Acad. morphicphases, J. Geol.,66, 374-393, 1958. Sci., 62, 69-92, 1955. Epstein, S., and T. Mayeda, Variation of O•8 con- Murray, E.G., and J. A. S. Adams, Thorium, tent of waters from natural sources,Geochim. et uranium and potassium in some sandstones, Cosmochim.Acta, 4, 213-224, 1953. Geochim.et Cosmochim.Acta, 13, 260-269, 1958. Evans, R. D., and C. Goodman,Radioactivity of Nier, A. O., A redetermination of the relative rocks, Bull. Geol.Soc. Am., 52, 459-490, 1941. abundancesof the isotopesof carbon, nitrogen, Faul, Henry, G. B. Gott, G. E. Manger, J. W. oxygen, argon, and potassium, Phys. Rev., 77, Mytton and A. Y. Sakakura,Radon and helium 789-793, 1950. in natural gas, Compt. rend. 19e Congr. Geol. Nockolds, S. R., Chemical compositionsof some Intern. Alger. Sec. 9, 339-348, 1954. igneousrocks, Bull. Geol. Soc. Am., 65, 1007- Geiss,Johannes, and D.C. Hess,Argon-potassium 1032, 1954. HELIUM, ARGON, AND CARBON 303

Norton, F. J., Helium diffusion through glass, or. dances of neon and argon extracted from radio- Am. Cer. Soc., 36, 90-96, 1953. active minerals, Phys. Rev., 96, 679-683, 1954. Palache, C., H. Berman, and C. Frondel, The Sys- Wetherill, G. W., Spontaneousfission yields from tem of Mineralogy, Vol. I, John Wiley and Sons, uranium and thorium, Phys. Rev., 92, 907-912, Inc., New York, 834 pp., 1944. 1953. Palache, C., H. Berman, and C. Frondel, The Sys- Zobell, C. E., Part played by bacteria in petroleum tem of Mineralogy, Vol. II, John Wiley and Sons, formation, or. Sed. Pert., 22, 42-49, 1952. Inc., New York, 1124 pp., 1951. Rakestraw, N. W., and V. M. Emmel, The solu- APPENDIX bility of nitrogen and argon in sea water, or.Phys. Chem.,•œ, 1211-1215, 1938. Potassium and uranium in carbonates. Of the Rankama, Kalervo, and T. G. Sahama, Geochem- common rocks, limestones may exhibit wide. istry, University of Chicago Press, Chicago, 912 variations in the content of trace elements such pp., 1950. Rayleigh, Lord, Nitrogen, argon and neon in the as potassium and uranium. For this reason, earth's crust with applications to cosmology, various carbonates of known uranium content Proc. Roy. Soc., London, 170, p. 451, 1939. were analyzed for potassium.The measurements Rogers, G. S., Helium-bearing natural gas, U.S. were made with a Perkin-Elmer flame photo- Geol. SurveyProfess. Paper, 121, 113 pp., 1921. Rosenfeld, W. D., and S. R. Silverman, Carbon meter. Samples of the carbonatesweighing isotope fractionation in bacterial production of 1.02 gram weredissolved in HC1 and evaporated methane, Science,130, 1658-1659, 1959. to dryness.They were put into aqueoussolution Sakakura, A. Y., Carolyn Lindberg, and Henry with i ml of 12 N HC1 and diluted to 250 ml Faul, Equation of continuity in geology with volume. The solutions were• run on the flame applicationstothe transport 0f r•oactive gas, photometer and the readings compared with U.S. G. S. Bull., 1052-I, 287-305, 1959. Satterly, J., and J. C. McLennan, The radioactivity those obtained on identically prepared CaC03 of the natural gasesof Canada, Trans. Roy. Soc. solutions of known potassium content. The Canada, sec. III, 12, 153-160, 1918. results were highly reproducible,and at con- Senftle, F. E., and N. B. Keevil, Thorium-uranium centrationsabove 100 ppm the data are probably ratios in the theory of genesis of lead ores, accurate to within 10 per cent of the values Trans. Am. Geophys.Union, 28, 732-738, 1947. reported.The data are givenin Table 6. Column2 Signer, P., and A. O. Nier, An upper limit for radiogenic A •s in potassium minerals, Geochim. of this table givesthe per cent of the CaCOa that et Cosmochim.Acta, 16, 302-303, 1959. is aragonite. Silverman, S. R., and S. Epstein, Carbon isotopic Samples of the Bikini a•oll cores had been compositionof petroleums and other sedimentary analyzedpreviously for uranium by Dr. M. S. organic materials, Bull. Am. Assoc.Petrol. Geol., Coopsof the Radiation Laboratory, Livermore, •2, 998-1012, 1958. Tatsumoto and Goldberg, Geochimet Cosmochim. California. The analyses were done using a Acta, 17, 201-208, 1959. neutron activation technique in which the Tomkeieff, S. L., The geochemistry of uranium, Np TMwas counted.Dr. Coopskindly provided ScienceProgress, 3•, 696, 1946. us with these samples. U.S. Bureau of Mines, Minerals Yearbook,1958, Someof the uranium analysesare the results Vol. II, Fuels, U.S. GovernmentPrinting Office, Washington, 484 pp. 1959. reportedby Tatsumotoand Goldberg [1959] using a Urey, H. C., H. Lowenstam,S. Epstein, and C. R. colorimetricmethod. Dr. Goldberggenerously McKinney, Measurements of paleotemperatures, providedus with someof thesesamples, and in Geol.Soc. Am. Bull., 62, 399-416, 1951. addition made other uranium analysesreported Wasserburg,G. J., G. Czamanske, H. Faul, and here. Two samples of marbles analyzed for' R. J. Hayden, Nat. Acad. Sci., Nat. Res. Council uranium by Dr. B. Doe are also included. Pub. 572, Nuclear ScienceSeries Report No. 23, 156-158, 1957. The analytical method used in each case is Wasserburg,G. J., and R. Bieri, The A 3scontent given in the table. of two potassiumminerals, • Geochim. et Cosmo- It is evident that in the carbonate rocks chim. Acta, I5, 157-159, 1958. reportedhere the potassiumcontent is widely Wasserburg,G. J., and R. J. Hayden, A40-K40 da- variable. The Muschel Kalk had the highest ting, Geochim.et Cosmichim.Acta, 7, 51-60, 1955. value of 4,600 ppm and the Algal ls. and Lead- Wasserburg,G. J., R. J. Hayden, and K. J. Jensen, A4ø-K•ø dating of igneousrocks and sediments, ville ls. the lowest values, 6 ppm. The high Geochim.et Cosmochim.Acta, 10, 153-165, 1956. potassium contents seem to be commonly Wetherill, G. W., Variations in the isotopic abun- associated with sizable insoluble residues. The 304 ZARTMAN, WASSERBURG, AND REYNOLDS marblesGMW-2 and F6-B are rich in phlogopite. content sincethe uranium valuesremain roughly It was found that the dissolutionprocedure was the same for these materials. vigorousenough to dissolveabout one-half this The marbles show a range in K content mica. The values reported for samples with similar to that found in limestones. All of the potassic noncarbonatefractions will therefore marbles contained some dolomite. Sample tend to be intermediate between the values of GMW-2 contains phlogopite, the larger part the whole rock and the pure carbonate fraction. being in the +80 sieve fraction. This effect is The weight per cent of insoluble residuesis clearlyseen in the resultson the total rock and given in column 3. the --40 q-80 sieve fraction. Of the two Franklin The K/U ratio in purelimestones such as the marbles,F6-B containssome phlogopite. Sample Spergen fm. and the Leadville Is. may be as F6-A was taken in the mine and is from a low as 25. This ratio is lower than that in average lithology which contains some coarse biotite igneous rocks by a factor of 400, and would crystals.No biotite was observedin the sample yield a value of R of 3 X l0 s. This differenceis analyzed. due to the great decrease in the potassium Sample fd 7-3 is a white dolomitic marble

TABLE 6. Analytical Data of Potassiumand Uranium in Carbonates

% Wgt. % Samples Aragonite Residue ppm K ppm U K/U

LIMESTONES Buckhorn 59 10.2 321 Buckhorn residue 2220 Cretaceous Chalk 0 1.4 198 1.082 188 ChicagoFormation Dol. 0.7 131 0.012 13100 BasalCoquina 46 0.6 112 0.872 147 Pr 2--Algal Isl. 0 6 1.412 4.25 Muschel Kalk 0 17.1 4600 Muschel residue 2900 Oolites(0-03) 98 0. ! 87 3.4' 25.6 GMW-2 (-40q-80), •narble 0 4.1 1000 GMW-2, total hand spec.crushed marble 0 7.1 2100 fd 7-3, marble 4,3.2 -<5 0.0834 < 60 fd 7-2, marble < 0.3 235 0.6404 390 F6-A Franklin Marble 0 <0.1 40 0.122 333 F6-B Franklin Marble 0 1.6 384 Leadville Limestone 32 0 0.4 58 0.282 207 Leadville Limestone 82 0 0.4 6 0.242 25 Leadville Limestone 128 0 0.3 20 SpergenFormation 0 0.4 25 1.02• 24.5 Calcite I 0 <0.1 _<5 <0.01 • Calcite II 0 <0.1 _<5 Calcite III 0 <0.1 9

MODERN SiII,;LLS 1336 99 < 0.1 1t}2 1337 100 74 1338 100 38 1339 100 54 1345 99 < o. 1 79 1348 100 <0.1 35 s-07 (B) 100 153 0.013' 11.77 X 10 a S-11 100 0.3 62 0.012' 5.17 X l0 s S-13 99 0.4 51 0.067 • 0.76 X l0 s S-14 100 39 0.022 • 1.77 X 10 • S-15 44 0.032 • 1.38 X l0 s S-17 0.6 154 0.041 • 3.76 X 10 • HELIUM, ARGON, AND CARBON 305

containing65 per cent dolomite and 35 per cent inorganically precipitated CaCO•, such as the diopside.The residueof this samplewas found calcitespar, has virtually no potassium;however, to contain 61 ppm of potassium.Sample fd 7-2 the oolite samplegave a result comparablewith is a dolomitic marble containing 0.26 per cent the shells and corals. These data suggestthat graphite. These samples were given to us by carbonatesprecipitated from seawa•er will have Dr. B. Doe, who determined the uranium somepotassium in chemicalcombination. On the contents in the course of another investigation. other hand, it is possiblethat some potassic Samplesof calcitespar were run for comparison. clays may be containedin all these precipitates. It is evident that thesecarbonates have extremely Somewell preservedfossil shells were analyzed low levels of potassium as suggestedby A. E. for K. SamplesI and II were large complete Engel (personalcommunication). specimens which were carefully cleaned to In order to determine whether any potassium eliminate any contamination. Sample III was is present in the original shell materials which taken from a glauconite-rich matrix and a constitute limestones as distinct from con- small amount of contamination with this mineral tributions from clay materials and detrital could easily cause this high result, while it particles, analyses were made of modem and appears that under optimal circumstancesthe fossil shells and corals. The concentration of potassiumcontent of shellsmay be preservedin potassiumin both modem shellsand corals are fossils.The concentrationlevels are very low roughlythe sameand rangefrom 30 to 150 ppm. but it may be possibleto utilize these materials It follows that these materials contain a signifi- for A4ø-K4ø dating with the most modern tech- cant amount of this element when formed. niques. Assumingcomplete retention for sample Many of the purer limestones have K con- I, this wouldcorrespond to 2 X 10-s cc STP/g centrations lying inside of this range. Some of A •ø. While this is readily measu.rable,the

TABLE 6. Continued

% Wgt. % Samples Aragonite Residue ppm K ppm U K/U

FOSSIL SHELLS I 99 0.2 29 II 100 0.3 27 III Shell Fragments 99 0.7 20i MODERN CORAL Coral I (0-04) 70 31 2.5 • 12.4 Coral II (0-05) 99 0.3 50 3.2' 15.6 Bikini Corals: Tare, depth 4' 51 0.2 35 2.3-3.23 12.5 13' 49 0.3 43 4.2-5.2 11.6 2.2-3.2 34' 89 0.1 60 6.5-8.1 8.2 63' 73 0.2 65 90-1/2' 66 0.1 85 4.7-5.7 15.7 97' 51 0.6 63 3.1-4.0 12.4 4.1-4.9 Roger, depth 4' 66 0.5 46 3.3-4.13 12.4 8' 62 0.5 67 4.3-5.1 14.3 23' 62 0.3 96 3.4-4.2 25.0 44' 97 0.2 43 6.1-7.5 6.3 FOSSIL CORAL 1437 Reef Coral 99 63

• M. Tatsumoto and E. D. Goldberg (Colorimetric) Geochim.et Cosmochim.Acta, 17, 201, 1959. "B. O. Goldberg (Fluorescence)personal communication. 8 M. Coops(Neutron activation) personalcommunication. • B. Doe (Isotope dilution) Thesis,Calif. Inst. of Technology,1960. 300 ZARTMAN, WASSERBURG, AND REYNOLDS problemof correctionfor atmosphericargon would found to be from 400 to 3000 ppm. This is a be formidable. much higherlevel than found in shells.If it is The shellsappear to have a uranium concen- assumedthat this potassiumis in the apatite tration about 100 times smaller than the corals. crystals rather than in fluids in the marrow, The K/U ratios for the shellsare much closer it may be possibleto date well-preservedfossil to the values for averageigneous rocks than the bone of Cretaceousage. corals which have ratios 1000 times smaller. Acknowledgments.We would like to thank Dr. The ashed bones of 15 different modern H. A. Lowenstam for providing us with many of the samplesof limestonesand shells.The authors vertebrates were also analyzed to investigate are indebted to Mr. Theodore Wen and Mrs. the possibility of dating appropriate fossil Dorothy Settle for their careful work on the potas- materials. The total range in concentrationwas sium determinations.

TABLE 7. Descriptive Data LIMESTONES Buckhorn,Mid-Pennsylvanian, Sulphur, Okla. CretaceousChalk, Campanian,Vigny, France ChicagoFormation, Niagaran, reef core, Thornton, Ill. Basal Coquina,Devonshire Marine Isl., Grape Bay, Bermuda Pr œ,Mid-Eocene, Algal Isl., Puerto Rico MuschelKalk, Mid-Triassic,Rorigliana, Italy Oolites(0-03) BahamaBank about 1 mile S.E. of S. Pt. Cat Cay 79ø15'W, 25ø30.5'N water depth 3 feet (recent) GMW-œ (-•0-•80 Mesh Fraction) Marble, Balducciquarry, Gouverneur,N.Y. GMW-œ, Marble, total hand spec.,Balducci quarry, Gouverneur,N.Y. Fd7-3 Marble, fetid limestone,shore of Sylvia Lake, N.Y. Fd7-œMarble, graphitic,900 level, shaft of No. 3 mine, Balmat, N.Y. F-6A Marble, New JerseyZinc Mine, Franklin, N.J. F-6-B, Franklin Marble, Franklin, N.J. L V-82, BaritesCabin, Colo., zone 2 LeadvilleLimestone scribedby Engeland LV-128,LV-32, Elk SweetCreekWater Colo., Lake,zone Colo.,i z.one i t TheseEngle.samplesMs. in preparationarede- SpergenFormation, Oolitic portion of the SpergenFormation, Mississippian, St. Genevieve,Missouri CalciteI, Iceland Spar Calcite II, Manhatten, Nev. CalciteIII, Spar 611, Hockerville, Okla. SHELLS 1336 Lodakia Orbicularis,Belmont Isl., QueensFree Cave, Ferry Rd., Bermuda 1337 LaevicardiumLaevigatum (recent) (TM•), Patorreef Lagoon,Bermuda 1338 Lodakia Orbicularis(recent), WhaleboneBay, Bermuda 1339 LaevicardiumLaevigatum (Pleist.), Basal Devonshire,Marshall Isl., Bermuda 1345 ConusCalifornianus (recent), Naval BaseE. Santa Barbara, Calif. 1348 ConusCalifornianus (Pleist.), Hilltop Quarry, San Pedro, Calif. S-07 (B) Haliotis Corrugata, La Jolla, Calif. S-11 Littorine Planaxis,La Jolla, Calif. S-13 AcmecaLimatula, La Jolla, Calif. S-1• Olivella Biplicata, La Jolla, Calif. S-15 Acanthina Spirata, La Jolla, Calif. S-17 Tetraclita squamose,La Jolla, Calif. FOSSIL SHELLS FossilI CrassatelitesVadosus, L. Maestrichtian,Ripky Formation,Coon Creek, Tenn. FossilII Trigonia Stantoni, Coon Creek, Tenn. III ShellFragments Mathews Landing Marl., Paleocene,Wilcox Co., Ala. CORAL Coral I (0-04) Key Largo Is. (Diploria labysinthiformis) Coral II (0-05) Key Largo Is. (Montastrea Annulasis) Bikini Coral taken at different depths 1]•37 Ree! Coral Pleistocene,Bermuda (Manuscriptreceived April 12, 1960;revised October 11, 1960.)