The UThPb age of equilibrated L chondrites and a solution to the excess radiogenic Pb problem in chondrites

Earth and Planetary Science Letters, 58 (1982) 75-94 Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

75

[6]

The U-Th-Pb age of equilibrated L chondrites and a solution to the excess radiogenic Pb problem in chondrites D.M. Unruh U.S. Geological Survey, M.S. 963, Box 25046, Federal Center, Denver, CO 80225 (U.S.A.)

Received June 18, 1981 Revised version received October 8, 1981

U, Th, and Pb analyses of whole-rock and troilite separates from seven L chondrites suggest that the excess radiogenic Pb relative to U and the large variations in Pb-Pb model ages commonly observed in chondritic meteorites are largely due to terrestrial Pb contamination induced prior to analyses. Using the Pb isotopic composition of troilite separates to calculate the isotopic composition of the Pb contaminants, the whole-rock data have been corrected for pre-analysis terrestrial Pb contamination. Two approaches have been used: (1) the chondrite-troilite apparent initial Pb isotopic compositions were used to approximate the mixture of indigenous initial Pb and terrestrial Pb in the whole-rock sample, and (2) a single-stage(concordant) model was applied using the assumption that the excess radiogenic Pb in these samples was terrestrial. Data for L5 and L6 chondrites yield a 4551 ---7 My age using the former correction and a 4550-+5 My age using the latter one. Corrected data for one L4 chondrite, Tennasilm, yield a 4552-+ 13 My age which is indistinguishable from that of the L5-L6 chondrites. However, the other L4 chondrite, BjurbiSle,yields a 4590-+6 My age. Th-U-Pb data suggest that this older age may be an artifact of the correction procedure, and that some of the discordancy of the Bjurbrle data is the result of either a recent geologic disturbance to the U-Th-Pb system or to terrestrial U loss. Some aliquots of the L5-L6 chondrites also show small amounts of discordancy (--10%) which are not easily attributable to terrestrial Pb contamination. The data from the L5-L6 chondrites and Tennasilm suggest that there are no more than ~ 15 My differences in the ages of L4-L6 chondrites.

1. Introduction There are two m a j o r p r o b l e m s regarding the U - T h - P b systematics of chondritic meteorites. First, large a p p a r e n t 2°7pb/2°6pb age differences (up to - 5 0 My) have b e e n observed, even a m o n g chondrites of the same class a n d petrologic grade [1-5]. Second, most chondrites appear to have more 2°rpb, 2°7pb a n d 2°8pb t h a n could have been derived from U a n d T h decay over the last -- 4550 M y [1,2,5,6]. T h a t is, when the data are plotted o n a U - P b evolution (concordia) diagram, a s s u m i n g the Pb isotopic c o m p o s i t i o n of Cation D i a b l o troilite [2] as primordial, the data plot above the c o n c o r d i a curve. Three general i n t e r p r e t a t i o n s (or

c o m b i n a t i o n s thereof) have b e e n proposed to a c c o u n t for the a p p a r e n t excess radiogenic Pb a n d a p p a r e n t age differences: (1) complex (two or more stages) U - P b e v o l u t i o n a r y histories, (2) variable p r i m o r d i a l (initial) Pb isotopic compositions, a n d (3) terrestrial Pb c o n t a m i n a t i o n in excess of that a c c o u n t e d for b y analytical Pb b l a n k s [1-5]. The P b - P b m o d e l age differences m a y reflect real age differences, c o r r e s p o n d i n g either to p r i m a r y differences in the times of f o r m a t i o n of the meteorite p a r e n t bodies or to either partial or complete resetting of the U - P b clock d u r i n g early thermal or i m p a c t m e t a m o r p h i s m (e.g. [2,5]). The m o d e l age differences could also reflect recent ( ~ 10 8 - 1 0 9 years) partial reequilibration of U - T h -

0012-821X/82/0000-0000/$02.75 © 1982 Elsevier Scientific Publishing Company

76 Pb. In this instance the model ages may have no rigorous significance. Evidence for recent disturbances to the Rb-Sr and K-Ar systems has been well documented. Many shocked L chondrites show moderate to large amounts of radiogenic 4°Ar loss at ~ 300-500 My ago [7-9]. These young ages are attributed to a major impact event (or events) and perhaps to the catastrophic break-up of the Lgroup parent body. Studies of shocked L-group chondrites (e.g. [10,11]) also indicate that Rb-Sr systems of these meteorites have been disturbed within the last ~ 1 0 9 years. Although periods of impact or thermal metamorphism on the chondrite parent bodies could conceivably explain some of the variation in Pb-Pb model ages, such an interpretation will not easily explain the apparent net enrichment of radiogenic Pb relative to U that is observed in most chondrites. In other words, a redistribution of U and Pb during metamorphism should not produce a net excess of Pb. In fact, because Pb is a fairly volatile element, one might expect any open-system behavior to produce Pb loss (e.g. [12,13]). Therefore, before any strict significance can be attached to the Pb-Pb model ages, the significance of the excess radiogenic Pb must be evaluated. The apparent 2°Tpb-2°rpb model age differences and excess radiogenic Pb could result from the wrong choice for the initial or primordial Pb isotopic composition [ 1,4,14-16]. There are, however, several good arguments against such an interpretation. First, there is at least one meteorite from each of the ordinary chondrite groups which is compatible with single-stage evolution from a Caiton Diablo troilite-type initial Pb. Second, the initial 2°~pb/2°4pb ratios in the meteorite parent bodies would have had to vary by a factor of 2 or more to account for the discordancy of some chondrite U-Pb data. Furthermore, because 2°7pb/2°rpb model ages show apparent variations of up to 50 My even within a single chondrite group and petrologic grade [1-4], the heterogeneities in initial Pb would have to have been present on a very small scale. Owing to the volatility and geochemical incompatability of Pb in most meteoritic minerals, it would seem improbable that large heterogenieties, even if initially present, could have been preserved during the thermal metamor-

phism that presumably produced the equilibrated chondrites (e.g. [17, and references therein]). The possibility of pre-analysis terrestrial Pb contamination has been acknowledged in almost all U-Pb studies of meteorites. Manhrs and All~gre [5] have proposed that virtually all of the excess radiogenic Pb in chondrites is terrestrial. Detailed studies of individual meteorites [ 18-20] have shown that terrestrial Pb contamination alone would not account for the discordancy of some meteorite data, but the possibility of some terrestrial contamination could not be ruled out. This work represents an attempt to determine the isotopic composition and source of the nonradiogenic Pb component in several L chondrites by studying the U-Th-Pb systematics of troilite separates and whole-rock splits from these meteorites. Troilite in iron meteorites may contain several ppm Pb but virtually no U (e.g. [2,21]). The only reported U-Pb analyses of chondritic troilite are those for the Barwell L5-6 [18] and Riehardton H5 [22] chondrites. The troilite in these meteorites contained small amounts of U (1 ppb) and higher, but variable, amounts of Pb (~ 20-30 ppb). Consequently, barring Pb addition from an outside source, the Pb isotopic composition measured today should be little changed from that at the time the parent body formed. Therefore, troilite is probably the most likely phase from which to determine the isotopic character of the nonradiogenic Pb component.

2. Experimental procedures 2.1. Samples and sample preparation All of the meteorites analyzed were observed falls with terrestrial residence times ranging from 115 years (Knyahinya) to 16 years (Barwell). Samples used for whole-rock analyses consisted of 50- to 100-mg-sized chips take from the interiors of pieces either cut (Harleton) or broken (the rest) from larger masses. In order to further minimize the potential effects of weathering and/or terrestrial contamination, chips were selected to be as free as possible from oxidation. However, traces of iron-staining (generally confined to the metallic

77 phases) were present in all samples. This sampling procedure may produce a sampling bias toward metal-depleted "whole-rock" analyses. Thus, the U, Th, and Pb abundances reported here may be somewhat higher than those in a truly representative sample. The chips were washed for 10 minutes in doubly-distilled 95% ethanol in an ultrasonic bath and dried at room temperature in a clean-air work station. The chips were then crushed to finer than 0.3 mm using a stainless steel mortar and pestle. Disposable stainless steel inserts were used in the mortar to minimize cross-contamination. The alc o h o l - w a s h i n g solutions were occasionally analyzed, and no evidence for preferential leaching of U or Th from the samples was found, although some Pb with terrestrial-like isotopic ratios was removed by the washing procedure. Troilite separates, with one exception, were obtained from partially crushed material by handpicking under a microscope on a clean-air bench. The troilite separate from Modoc was obtained using heavy liquids and an electromagnetic separator. The troilite separates appeared to be > 95% free from inclusions of other minerals, although the somewhat higher than expected U abundances in these separates (Table 3) suggest that they may not have been as pure as visual estimates indi. cated.

2.2. Analytical procedures Chemical procedures used in this study were similar to those given by Unruh et al. [18]. U, Th, and Pb concentrations were obtained using a 2°sPb-235U-23°Th-enriched tracer during the early phase of this study and a 2°spb-233U-236U-23°Th enriched tracer in the later phase. The latter tracer enables us to determine not only the U, Th, and Pb abundances, but the fractionation-corrected U isotopic compositions as well. The new tracer is quite p u r e (238U/233U ~ 1.4 × 10-4; 232Th/23°Th - 6 . 2 × 10 -4, 2°6pb/2°sPb-4.1 × 10 -4) so that spike corrections to the isotopic composition data were generally insignificant. Silicate samples were decomposed using an HFH N O 3 mixture in a T F E teflon bomb at ~ 120°C for 2 days during the early phase of this study, and

an H F - H N O 3 mixture in disposable PFA teflon screw-cap containers at ~ 60°C for 3-5 days during the later phase. Troilite samples were decomposed in teflon beakers with 2N HBr. Pb was separated from other elements using anion exchange in 1.0N HBr medium. A two-column procedure was employed. The semi-pure Pb fraction obtained from a 200-#1 column was loaded onto a 30-#1 column for final purification. When the sample size was ~< 20 mg, the large column was not used and the small column was used twice. Pb blanks were generally 0.2-0.3 ng (10-9g) for whole-rock analyses and 0.10-0.2 ng for troilite analyses. There was one exception in which an ~0.5-ng blank was obtained (see Table 1). The higher blank was traced to a contaminated H N O 3storage bottle. Blank uncertainties were obtained from duplicate or triplicate blank analyses performed concurrently with 1-5 sample analyses, and encompass the total range observed for a set of samples. A potential analytical problem with Th exists regarding equilibration of spike and sample Th. Perchloric acid was not used during decomposition of the samples in order to minimize the Pb blank. Consequently, it is possible that insoluble fluorides containing Th could have formed and equilibration between sample and spike was not attained. This problem was observed by _Tatsumoto et al. [14] and is also discussed with regard to some of the samples analyzed in section 4.2. U and Th were separated from other elements using anion exchange in 6.5N H N O 3 medium. U and Th blanks were 1-5 pg (10 -t2 g) and 2-10 pg, respectively, for whole-rock analyses and 1-3 pg and 2-5 pg, respectively, for troilite analyses. However, as pointed out in other studies [18,28], Th-contaminated filaments used for mass spectrometry can produce erroneously high Th values in low-Th samples. The zone-refined Re ribbon used for center filaments in U-Th mass spectrometry was washed in 6N HC1, and random segments were examined in the mass spectrometer for U and Th contamination. The U and Th signals from these filaments were less than 5 × 10-18A and 2 × 10 - 17A, respectively at 2050°C, and both were less than 2 × 10-1SA (less than 10 counts/s) at 2000°C, the normal running temperature. The

78

lowest U-Th samples provided a signal of > 2 × 10-16A for 238U and 232Th, so the background signals do not affect the calculated concentrations beyond the assigned uncertainties. Isotopic ratios were obtained using an NBStype, two-stage mass spectrometer. Pb was run using the single-filament, H3PO4-silica gel method, and U and Th were run by the conventional triple-filament technique. Data were obtained either by pulse-counting or from a Faraday cage and vibrating reed electromer, depending upon the intensity of the ion beam. For more complete details, see Tatsumoto and Unruh [29]. Uncertainties in the reduced data (Tables 1-6) and the correlation coefficients between the 2°7pb/2°4pb and 2°6pb/2°4pb data (Table4) were obtained from equations equivalent to those given by Ludwig [30].

3. Results

3.1. U, Th, and Pb abundances

U, Th, and Pb abundances in the L chondrites are shown in Table 1. Included are data from our previous study of the Barwell L5-6 chondrite [ 18]. U and Th concentrations range from 5.1 to 16.6 ppb and 32.7 to 49.5 ppb, respectively. 232Th/238U ratios range from 2.7 to 6.6. The U concentrations and the T h / U ratios show the greatest variation among the L6 chondrites. The data exhibit a crude trend of decreasing T h / U with increasing U and Th abundances, thus suggesting the presence of a high-U, low-Th/U component in these samples. Uranium is known to be heterogeneously distributed in equilibrated chondrites and concentrated in minor phases such as chlorapatite and whit-

TABLE 1 U, Th, and Pb abundances in selected L chondrites Meteorite

Class

Sample (mg)

Blank a 2o6Pb

Total b 206Pb

Concentration U

Bjurb6le

-1 -2

L4

61.99 87.75

0.36-+0.07 0.40-*-0.03

9.98 -+0,02 15.29 --+0.03

11.50---0.06 11.53---+0.06

Tennasilm

-1 -2 -3

L4

95.76 45.50 54.60

0.404-0.03 0.19---0.03 0.23-+0.04

32.09 -+0.04 16.98 -+0.02 19.84 -+0.02

11.27---0.06 11.20---0.07 11.77--+0.07

Knyahinya

-1 -2 -3

L5

61.04 71.09 66.15

0.64--+0.12 0.27 ± 0.03 0.23-+0.04

4.581 -+0.015 5.107 -+ 0.006 4.520---0.006

13. I 1---0.09 13.74 -+-0.09 11.38---0.06

Barwell

-I b -2

L5-6

100.19 93.40

0.34--+0.09 0.34 -+ 0.09

5.811 -+0.026 6.235 ± 0.011

9.35-+0.07 9.84 ± 0.05

Bruderheim - 1 -2

L6

107.26 89.50

0.36 -+ 0.07 0.27 -+ 0.03

8.939-+ 0.051 7.263 -+ 0.013

13.93 -+ 0.07 14.98 -+0.09

Harleton

-1 -2 -3

L6

95.98 75.29 72.84

0.36-+0.07 0.27-+0.03 0.23-+0.04

6.481 -----0.021 4.186-+0.009 4.644-+0.007

6.22-+0.04 5.51 -+-0.03 5.11 ±0.03

Modoc

-I -2 -3

L6

64.55 92.35 77.65

0.64±0.12 0.27-+0.03 0.23 -4-0.04

6.867--+0.032 10.330-+0.031 11.432 -+ 0.022

12.59-+0.08 14.61 -+0.08 16.56 ± 0.09

a In picomoles. Total 2o6Pb=sample+blank. b " W R I " and "WR2" from Unruh et al. [18]. U isotopic composition from M. Tatsumoto, personal communication (1980).

79

lockite (e.g. [31]). Furthermore, these U-enriched phosphate minerals have highly variable T h / U ratios [31]. Both total Pb and 204pb abundances are shown in Table 1. The 204pb contents are more indicative of the true (primary) variations because the other Pb isotopes are comprised of both primordial Pb and radiogenic Pb from U or Th radioactive decay. 2°4pb abundances, and consequently the 238U/204pb ratios, vary by more than a factor of 50. Furthermore, there is an overall trend of decreasing 2°4pb abundances from type 4 to types 5 and 6. Pb studies of L3 chondrites confirm and magnify this trend [3,4,26]. 204Pb abundances in the limited number of L3 chondrites analyzed range from 1.8 to 110 ppb. Thus, the total range of 2°4pb concentrations in L chondrites spans more than 3 orders of magnitude. However, within the

L3 chondrites there is a 60-fold variation in the 204Pb concentrations [26]. The magnitude and trend of the 2°4pb variations are consistent with those seen in other heavy volatiles, TI and Bi [32-34], and in the noble gases (e.g. [35, and references therein]). Hg is the only heavy volatile that does not appear to follow the trend [6,36]. Uranium isotopic compositions were found to be, within error, the same as the terrestrial value of 238U/235U = 137.9, thus confirming the observations of Chen and Wasserburg [37]. The somewhat lower 238U/235U ratio (135.2--+0.2), which was originally measured in Barwell [18], was found to be in error. The error presumably resulted either from 235U contamination from a previously used isotopic tracer or to background contributions to the 235U peak during mass spectrometry. In any case, the complete rebuilding of the vacuum

Atomic ratios 238U//204pb

232 Th//23s U

238U//235 U

2.01 -----0.02 2.19 ± 0 . 0 1

4.88± 4.49±

0.05 0.03

3.75±0.04 3.80±0.03

136.9±0.8

3 1 0 . 7 ± 1.5 340.6-+0.9 334.0±0.8

5.10 -+0.01 5.50 ± 0 . 0 2 5.42 ± 0 . 0 2

1 . 8 8 ± 0.01 1 . 7 4 ± 0.01 1.85-+ 0.01

3.58±0.03 3.62±0.03 3.58±0.03

1 3 7 . 0 ± 1.0 136.6--+2.2

4 3 . 7 ± 1.0 45.8-+0.4 39.5--+-0.2

35.4±0.7 38.4±0.2 37.0±0.2

0.081 ± 0 . 0 1 4 0.120±0.004 0.122±0.005

3.44±0.08 3.45±0.04 3.59±0.03

137.4±1.0 137.9±0.6

38.9±0.5 -

33.8±0.4 39.0 ± 0.4

0.149-+0.007 0.230 -+ 0.007

53.2 ± 38.8 ±

2.6 1.2

4.30±0.07 -

137.2±0.6

36.6±0.9 44.9±0.3

46.2±0.4 44.3±0.2

0.238---+0.006 0.205±0.003

49.8 ± 63.0 ±

1.3 1.1

2.71±0.06 3.10--+0.03

137.6±0.8

37.5±0.2 32.7 ± 0.3 32.8±0.2

48.1 ± 0 . 4 33.3 -+ 0.2 46.6±0.3

0.422±0.006 0.263 ± 0.003 0.412±0.004

1 2 . 5 3 ± 0.18 1 7 . 7 8 ± 0.24 1 0 . 5 4 ± 0.13

6.23±0.05 6.12±0.06 6.63±0.05

1 3 7 . 3 ± 1.2 137.6±0.7

42.9-+ 1.8 48.5±0.9 49.5±0.4

64.8±0.9 70.1 ± 0 . 3 98.2±0.3

0.514±0.014 0.518-+0.004 0.881 -+0.005

2 0 . 8 3 ± 0.59 1 9 . 0 6 ± 0.16 16.01 ± 0.12

3.52-+0.15 3.43±0.06 3.09±0.03

137.5±0.8 137.5±0.4

Th

Pb

2~pb

41.7±0.3 42.4-+0.2

134.5±1.0 147.7-+0.4

39.1 ± 0 . 2 39.2±0.2 40.8±0.2

137 ±25 97.6 ± 3.5 78.8 ± 3.4

80

pumping system in the mass spectrometer, and total elimination of the 235U tracer from the laboratory seem to have generally eliminated the problem, although the combined data in Table 1 may reflect an ~ 0.2% bias toward lower 238U/235U ratios. All 2°7pb/235U ages are calculated using the terrestrial 23SU/235U ratio of 137.9.

3.2. Pb isotopic compositions and U-Pb systematics The Pb isotopic compositions of the bulk chondrites are shown in Table 2. As one might expect from the U and Pb abundances, Pb in the IA chondrites is considerably less radiogenic (2°6pb/2°4pb ~ 13.2-15.8) than that in the L5 and L6 chondrites (2°6pb/2°4pb ~ 30-160). The Pb isotopic characters of the individual meteorites are reasonably reproducible, considering that analyses

were performed on separate chips rather than powder splits. The Pb isotopic data are plotted on a 2°7pb/2°npb vs. 2°6pb/2°4pb diagram in Fig. 1. The data define a linear array, the slope of which corresponds to an age of 4521 ± 10 My (2o, [38]). The error includes the scatter of the points about the best-fit line. As one can see from the large value of S / ( N - 2) ~ 34 (S = sums of the squares of the residuals; N = number of samples, [51]), the data are quite badly scattered. Thus, the apparent 4521 My age is of questionable significance. The inset in Fig. 1 shows the percent deviation of the 2°7pb/2°4pb data from the best-fit line as a function of 2°6pb/2°4pb. The slope of the line is controlled by the more-precise Harleton and Modoc data such that the best-fit line (horizontal line in the inset) completely misses most of the radiogenic

TABLE 2

Pb isotopic compositions in selected L chondrites Meteorite

Class

Raw data 2°6pb/2°4pb

207pb/2°6pb

208pb/2°6pb

Corrected for blank and mass fractionation 206pb/2°4 pb

Bjurbble

-1 L4 -2

15.673±0.040 15.845±0.018

0.9062-+0.0022 0.9012±0.0009

2.246±0.003 2.226-+0.003

15.615+ 0.045 15.807± 0.021

Tennasilm

-1 L4 -2 -3

13.249±0.026 13.757-+0.008 13.531±0.015

0.9541 -+0.0009 0.9386±0.0002 0.9460-+0.0005

2.488-+ 0.003 2.432-+0.002 2.456-+0.001

13.225 ± 0.032 13.740± 0.015 13.519-+ 0.019

Knyahinya

-1 L5 -2 -3

78.6 92.5 86.7

±0.8 ± 1.6 -+ 0.6

0.6709±0.0010 0.6645-+0.0007 0.6655-+ 0.0005

1.141±0.004 1.108-+0.002 1.135-+0.002

161 115.9 107.6

-+19 ± 4.1 ± 4.5

Barwell

-I 'L5-6 -2

63.4 50.8

-+0.2 ±0.2

0.6862-+0.0013 0.7028-+0.0010

1.337-+0.004 1.309±0.002

74.8 56.7

± 3.3 -+ 1.6

Bruderheim

-I L6 -2

61.7 70.3

±0.4 -+ 0. I

0.6847-+0.0012 0.6725 ± 0.0003

1.135±0.006 1.077-----0.002

68.6 78.8

± ±

Harleton

-1 L6 -2 -3

29.70 ± 0 . 1 2 29.38 ±0.06 29.05 ±0.05

0.7543-+0.0010 0.7584±0.0007 0.7614--+0.0008

1.867±0.008 1.888±0.001 1.920±0.002

30.79 -+ 0.23 30.51 ± 0.13 29.95 ± 0.16

Modoc

-1 L6 -2 -3

35.35 ±0.25 41.56 ± 0 . 2 0 32.84 ±0.07

0.7378±0.0012 0.7177--+-0.0015 0.7412-+0.0004

1.488-+0.003 1.380--+0.004 1.526±0.001

38.86 -+ 0.72 42.92 ± 0.26 33.42 ± 0.11

1.5 1.2

a Mass fractionation 0.1 -'-0.03%/M.U.; blank isotopic composition: a = 19.0-+0.10, ~ = 15.65-+0.05, ~,= 38.5-+0.3. b Relative to Cation Diablo troilite Pb [2]. c Data from Unruh et al. [I 8].

a

81

data. When the data are given equal weighting and the regression calculation is performed, a 4552 --- 7 My age is obtained (solid curve in the inset). Based on previous U-Pb studies, this latter age would seem more reasonable, but this apparent isochron passes distinctly ( ~ 2%) below C a t o n Diablo troilite Pb (solid circle in the inset). Consequently any apparent age derived from these data may not have rigorous significance. The U-Pb data are plotted on a U-Pb concordia diagram [39] in Fig. 2. The data are corrected for the Pb isotopic composition of Cation Diablo troilite Pb [2] as the assumed initial Pb isotopic composition. The trend defined by the data intersects the concordia curve at 4547 --- 12 My and - 4 7 0 -+ 233 My. Data for the L5-L6 chondrites alone yield 4 5 4 7 -+ 15 My and - 5 1 0 ± 3 6 7 My intercepts. Giving the data equal-weighting yields apparent

ages which are indistinguishable from these. The upper intercept is in marginal agreement with the apparent Pb-Pb age (Fig. 1) and a recent disturbance to the U-Pb system is suggested by the near-O My lower intercept. There are three factors which preclude such a simple interpretation. First, all of the data plot above the concordia curve, showing from ~ 5 to 150% excess radiogenic Pb. If the recent event caused large-scale Pb redistribution on the parent body, then one would expect the data either to be linearly distributed either about the concordia curve or below the curve as a result of recent Pb loss (e.g. [12,13,40]). Second, the data are grossly scattered with S / ( N - 2) = 65 for all data and 85 for the L5-L6 chondrites alone. A two-stage model such as the one discussed in detail by Gale et al. [20] would predict a precise linear array on the

M o d e l ages ( M y ) h ~-o7P b / 2 o 6 Pb

2o8pb/2O6 Pb

2o7pb/2O4 P b

2o8pb/2O4 P b

2o7pb/2O6 P b

2o7p b / 2 0 8 P b

0.9098±0.0023 0.9042±0.0011

2.254±0.003 2.236±0.003

14.206 + 0.055 1 4 . 2 9 2 ± 0.026

35.193± 35.351±

0.115 0.067

4556±13 4544± 7

4579+24 4589+14

0 . 9 5 6 7 ± 0.001 I 0.9409±0.0005 0 . 9 4 8 4 ± 0.0007

2.499 ± 0.004 2.441 ± 0 . 0 0 2 2.466 ± 0.002

12.653 ± 0 . 0 3 4 1 2 . 9 2 8 -+- 0.015 12.815 ± 0.020

3 3 . 0 4 7 ± 0.095 33.542± 0.049 33.321 ± 0 . 0 5 4

4 5 1 2 ± 12 4494± 6 4507± 6

4472±23 4459±15 4461 ÷ 16

0.6469±0.0048 0.6570±0.0014 0.6577±0.0016

0.997±0.025 1.063±0.007 1.089±0.009

104 76.1 70.8

0.6779+0.0027 0.6961±0.0022

1.297--+0.012 1.270±0.010

50.7 39.5

± ±

0.6791+0.0019 0.6674+0.0009

1.099±0.010 1.042±0.005

46.6 52.6

± ±

0.7519+0.0019 0.7550±0.0011 0.7589±0.0011

1.859±0,010 1.883±0.002 1.918--+0.002

23.16 ± 23.04 + 22.72 ±

0.7298±0.0025 0.7158±0.0020 0.7403±0.0006

1.437±0.010 1.367±0.005 1.515-+0.002

161 123.1 117.2

+29 ± 4.4 + 5.0

4553 + 5 4 5 5 0 -+- 3 4543± 2

4528±38 4503+16 4531+11

2,2 1.1

97.0 72.0

± 4.4 ± 2.1

4549± 4545±

4583±25 -

1.0 0.8

75.3 82.1

--+ 1.8 ± 1.3

4 5 3 7 -+- 4 4529 + 1

4342--+34 4540--+12

0.19 0.14 0.13

57.25 ± 0.55 57.45 + 0.26 57.46 + 0.32

4504± 4510 + 4514 +

5 4 4

4736+15 4690±15 4763+11

28.36 --+ 0.54 30.72 -+ 0.21 24.74 ± 0.08

55.8 ± 1.1 58.67 --+ 0.43 50,65 ÷ 0.18

4535± 4527± 4506±

6 5 2

4500±57 4493±25 4324±12

± 12 ± 2.7 ± 3.0

4 3

82 14B 12EI 5 - S ../ ........................................ ~...........................:

Ia

~

-~

o

20p 4~

loo

200pb I

88

2oo

BA-lpa~BR_1 B A - 2 _ / " BR-2

4B

T= 4521 ± 18 My

p l B~DI~2NI~_;2

SLOPE - B. EBBS ± . BB42 INTERCEPT- 4. B3 ; . Bfl S/GN-2)- 34

211 T m

~

4~

81]

1BO

129

286ph12~4pb

20a

Fig. 1.2°Tpb/2°4pb vs. 2°6pb/Z°dpbdiagram for L chondrites. The best-fit line through the data yields a 4521± 10 My apparent age. Uncertainties in this and other diagrams are 2(; (a priori)×S/(N-2) [38]. The large amount of scatter suggests the apparent age may not be valid. Abbreviations used are TE= Tennasilm, BJ = Bjurbble, HA = Harleton, MO = Modoc, BA=Barwell, BR=Bruderheim, KN=Knyahinya. The inset shows the percent deviation of the 2°Tpb/2°4pb data as a function of 2°6pb/Z°4Pb. The curve is a 4552 My apparent isochron calculated for equally-weighted data points. CD is Cation Diablo troilite Pb.

concordia diagram ( S / ( N -- 2) ~ 1), provided that the data are corrected for the proper initial Pb isotopic composition. Third, a simple two-stage model (formation at ~ 4550 My and a disturbance within the last ~ 500 My) would predict an increase in the single-stage 2°Tpb/2°rpb model ages with increasing discordancy above concordia. Comparing Fig. 2 and Table 2, one can see that, in general, the opposite is true. Furthermore, this two-stage model would predict good agreement (within -----10 My for 2°rpb/238U~0.7-1.5) between the 2°7pb/2°rpb and 2°7pb/E°8pb (actually 2°7pb/E°Spb X232Th/235U) model ages. The 2°7pb/E°Spb model ages (Table2) are less precise than the 2°7pb/2°6pb ages because the uncertainty in these ages includes the uncertainty in the T h / U ratio. However, the data still indicate that a simple two-stage model cannot be applied to these meteorites. Therefore, these meteorites were either formed, metamorphosed, or shocked at distinctly different times; or variable amounts of Pb, isotopically distinct from that in Cation Diablo troilite, have been added at some time during the meteorites' histories. It is this latter possibility that we have chosen to investigate in detail ([18,25,38], this work). 3.3. U - T h - P b analyses o f troilite separates

TE-2~"

2.4 ?.,B

L5-L5

./*"

INTERCEPTS AT 454"7 ± 15 AND -51 ~1 Z. 3 6 7 "y

H^-~,I¢/I¢/TE-I

,^,_2..7,y-2

,.2

8" B1 % ' ~ . Y / . /"

,l,.. f ta

/ I f TE-3

f"/'-

. 40

.

a

I~:LR::::S AT

. ~

.

.

120 207pb/235U

.

.

. 16~

. 2gg

Fig. 2. U-Pb concordia diagram for L chondrites. The best-fit line intersects the concordiacurve at 4547± 12 and -470±233 My. L5-L6 chondrites alone yield 4547± 15 and -510±367 My intercepts. The large amount of scatter and the negative lower intercept suggests the 4547 My age may not be valid. The data are corrected for Cation Diablo troilite Pb as the initial Pb [2].

In order to examine further the excess radiogenic Pb problem, troilite separates have also been analyzed in hope of finding the least-radiogenic Pb component in these meteorites. The U, Th, and Pb concentrations of these separates are shown in Table3 and the Pb isotopic compositions in Table 4. U and Th concentrations of the troilite separates, though variable and lower than those in the whole-rock, are generally higher than expected. These abundances and the low T h / U ratios relative to the bulk meteorites probably reflect the presence of small amounts of mineral impurities enriched in U relative to Th. Pb abundances in the troilite are also quite variable and by comparing Fig. 2 and Table 3 one can see that there is a crude correlation between the amount of excess radiogenic Pb in the whole-rocks and the Pb concentrations in the troilite.

TABLE 3 U, Th, and Pb abundances in troilite separates from L chondrites Sample

Bjurb61e Termasilm Knyahinya BarweU a Bruderheim Harleton Modoc

Class

L4 L4 L5 L5-6 L6 L6 L6

Sample (rag)

15.90 5.04 10.34 17.93 10.46 5.15 33.59

Atomic Ratios

Concentration (ppb)

Pb blank (ng)

0.121 -+0.008 0.151-4- 0.016 0.95 -+ 0.014 0.20 ±0.06 0.20 ±0.02 0.20 -+0.02 0.095 -+ 0.014

U

Th

Pb

232Th//238 U

238U//204 pb

0.65-+0.07 3.49 + 0.22 0.30 -+ 0.06 0.85 ±0.07 1.47-+0.09 2.30-+0.34 0.99 ± 0.02

1.00-+0.05 1.76 ± 0.23 0.41 -+ 0.21 0.65-+0.05 1.58+0.18 2.44-+0.52 1.32 -+ 0.04

55.9-+0.5 224 -+4 6.5-+0.7 199 ± 3 43.6-+0.7 150 --+3 106 -+ 1

1.59-+0.19 0.62 -+ 0.08 1.4-+0.8 0.80-+0.09 1.11-+0.14 1.09-+ 0.29 1.38-+0.05

0.67 -+ 0.07 0.77 ÷0.05 3.0-+0.5 0.27 4- 0.02 2.24-+0.17 1.00±0.15 0.60 -4-0.02

a Tr 2 from table 2 of Unruh et al. [18].

TABLE 4 Blank-corrected and apparent initial Pb isotopic compositions in troilite separates from L chondrites Sample

BjurbSle Termasilm Knyahinya Barwell c Bruderheim Harleton Modoc

Type

L4 L4 L5 L5-6 L6 L6 L6

Apparent initial Pb isotopic compositions a

Corrected for blank and mass fractionation

b

2o6pb/204 pb

207pb/204pb

208pb/204 pb

pb

2o6pb/2°4 Pb

207Pb/2°4 Pb

208pb/204 Pb

P

16.90 ± 0.04 12.06 -+ 0.07 19.30 -+ 0.28 18.44±0.05 19.84-+0.16 19.24 ± 0.09 18.94-+ 0.03

14.63 -+0.04 I 1.92 -+ 0.08 16.35 ± 0.34 15.67±0.05 16.47-+0.18 16.04-+ 0. I 1 t 5.99 -+ 0.05

36.74 -+0.12 32.03 -+0.22 38.45 -+ 0.80 37.85--+-0.12 39.50±0.52 39.02 ± 0.26 38.73 ± 0.13

0.739 0.889 0.707 0.910 0.733 0.723 0.710

16.19±0.09 11.27 --+0.08 16.35±0.35 18.16±0.06 17.55±0.24 18.21 ±0.18 18.32±0.03

14.18±0.07 I 1.43 -----0.09 14.45±0.48 15.49±0.05 15.04-+0.21 15.41 ±0.14 15.61 ±0.05

36.42±0.12 31.93 ± 0.22 37.38±0.97 37.79±0.12 38.87--0.52 38.75--+0.26 38.52--+0.13

0.874 0.873 0.856 0.940 0.806 0.850 0.715

a Calculated from the 238U/2°4pb and 232Th/2°4pb ratios (Table 3) and an assumed age of 4550 My. b Correlation coefficient between 2otpb/2°4pb and 2°Tpb/2°4pb [30]. c T r 2 from table 3 of Unruh et al. [18].

84 4. Discussion • TROILITE o WI-IOLE-RDE:X

t-- 12

J TEP.RI[STRIhLFPo BA MO

'rE1~

i11

8

lfl

12

14

IB

18

206pb/2f14pb

Fig. 3. 2°7pb/2°4Pb vs. 2°rpb/2°4pb plot of apparent initial ratios from troilite separates. Troilite separates are shown by closed symbols, and whole-rocks by smaller open symbols. CD is Cation Diablo troilite Pb, the stippled field shows the isotopic compositions of Pb blanks in several U.S. and European cities and is referred to as "normal" terrestrial Pb. The troilite data all fall within the area of possible mixtures of normal terrestrial Pb (broken lines). The solid line is a 4550 My reference line. Both the blank-corrected and U- and Th-decaycorrected (apparent initial) Pb isotopic compositions are shown in Table 4. The apparent initial Pb isotopic compositions are plotted on a 2°Tpb/2°4pb vs. 2°rpb/2°4pb diagram in Fig. 3. The terrestrial Pb field consists of reported blank isotopic compositions ([3,5,18,19,28,41], and this work) in several U.S. and E u r o p e a n cities. This field should represent the most probable range of isotopic compositions of industrial or pollutant Pb contaminants, but of course not all possible terrestrial Pb contaminants. Also shown for reference are Cation Diablo troilite Pb (CD) and a 4550 M y reference line (solid line). All of the troilite data plot within the area b o u n d e d by the terrestrial Pb field and Cation Diablo trolite Pb (dashed lines). Furthermore, the data from those samples that are most discordant in the lead-excess direction in Fig. 2 and show the highest troilite Pb abundances plot near or within the terrestrial field. The troilite Pb data plotted on a 2ospb/204pb vs. 206pb/204 Pb diagram (not shown) exhibit the same features. Consequently, the most likely interpretation is that the Pb in these separates, and therefore the excess Pb in the wholerocks, is indeed terrestrial.

If it is accepted that these meteorites are contaminated with terrestrial Pb in excess of that accounted for by analytical blanks, then the a m o u n t and isotopic composition of the contaminant must be determined before any rigorous significance can be attached to the apparent ages (Figs. 1 and 2). Two approaches to this problem have been employed [25]. Both involve using the Pb isotopic compositions of the troilite separates to approximate the extent and isotopic composition of the Pb contaminant.

4.1. Troilite "initial" Pb Given that the terrestrial Pb contamination observed in the troilite separates indicates contamination of the bulk meteorites, then the proper non-radiogenic Pb correction to the whole-rock data will consist of a mixture of terrestrial Pb and true initial Pb. The range of possible corrections is given by a mixing line in Fig. 3 through the

1.2

1, fl

f_

fl.B

ft. 4

ft. 2

ft. 4

ft. 6

238U/285pb

ft. 8

1.8

1.2

Fig. 4. 2°7pb/2°6Pb vs. 23SU/2°6Pb modified U-Pb concordia diagram [50] for Harleton. Open circles denote the three whole-

rock analyses of Harleton uncorrected for initial Pb. Partially filled circles (lines B) show the location of the data when corrected for Cation Diablo troilite (CD) initial Pb and the closed circles (lines C) represent corrections for the troilite apparent initial Pb (Tr I). Line A represents undisturbed and uncontaminated 4550 My old samples uncorrected for initial Pb. Correction of data that lie on line A for initial Pb produces a single point (A1). The horizontal lines depict the mean model ages obtained after correction for initial Pb.

85 chondrite troilite apparent initial Pb and Cation Diablo troilite Pb, the assumed initial Pb. If the individual whole-rocks and troilite separates contain the same relative proportions of contaminant Pb and indigenous initial Pb, then use of the troilite apparent initial Pb isotopic compositions as the non-radiogenic Pb corrections to the wholerocks is approximately valid. Clearly, this assumption will not hold for Bjurbtle (Fig. 3) because the troilite initial Pb isotopic composition is more radiogenic than that in the whole-rock. Therefore, correction for the troilite Pb would produce negative ages. The correction procedure is shown graphically, using Harleton as an example, on a modified U-Pb concordia diagram [50] in Fig. 4. Samples which were undisturbed and uncontaminated since their formation 4550 My ago, and which evolved initially from Pb with the isotopic composition of Cation Diablo troilite (CD) would plot on the segment of line A between CD and A1 when no initial Pb correction is made, and at a single point (A 1 ) when corrected for Cation Diablo troilite Pb. The Harleton data, uncorrected for initial Pb, are shown by the open circles and lie distinctly off reference line A. Correction for Cation Diablo

_

t \ % o

a. 0 <]-1 70

1, 2

HA

'

'

MO 8o

. oo

L~

BAI-2 ~ ' " ~ ~

I''~~J~HoBRI.

78

7"

TE~$3 . ' ' / KN3 7

,oo

20rpb123SU

|

/"

B8

TEl

LS-L6

KN2

INTERCEPTSAT

08

188

118

121~

207pb/235 u

Fig. 5. U-Pb concordia plot for chondrite troilite-corrected data. The L5-L6 data define an array that intersects the concordia curve at 4551±7 and 376±460 My. The data are still somewhat scattered, but the amount of scatter has been considerably reduced. The inset shows the relative deviation (%) of the 2°tpb/23Su data from the best-fit line as a function of 2°7pb//235 U.

troilite Pb causes the data to move along lines B to the positions occupied by the half-filled circles. These positions are equivalent to those shown in Fig. 2. The 2°Tpb/Z°tPb model ages in Table 2 are represented by the intersection of a horizontal line through the data with the concordia curve [50]. Correction for the Harleton troilite apparent initial Pb ( T r I in Fig. 4) shifts the points along lines C to the positions shown by the solid circles. The troilite-corrected data (correction C in Fig. 4) for the L5-L6 chondrites and Tennasilm are plotted on a U-Pb concordia diagram in Fig. 5. All of the data from the L5-L6 chondrites now plot within ~--+ 12% of the concordia curve. The trend defined by these data intersects the concordia curve at 4551 --+7 My and 376 --+460 My. The amount of scatter has been reduced by a factor of 2.4 relative to the Cation Diablo-corrected data ( S / ( N - 2) 35) but there is still much more scatter than would be predicted by the assigned uncertainties. Therefore, the ages obtained will have real significance only if the scatter is normally distributed about the best-fit line--that is, if the scatter results from random over- or under-correction of the individual samples for terrestrial Pb. Close examination of the data (inset in Fig. 5) suggests that this indeed may be the case. Those meteorites, Harleton and Modoc, on which the troilite correction had the largest effect ( ~ 50% of the radiogenic 2°tpb) show the largest deviations from the best-fit line. Furthermore, the mean value of the three analyses of each meteorite plot on or near the best-fit line. In any case, the near concordancy exhibited by most of the data suggest that the 4551 +--7 My age represents a better approximation of the true age than does the age obtained by correcting the data for Cation Diablo troilite Pb alone. The 376 +--460 My lower intercept, though imprecisely defined, is in accord with the ~ 500 My disturbances to the K-Ar systems [7-9] and recent disturbances to the Rb-Sr system [10,11] observed in many shocked L chondrites. One should note, however, that because most of the data cluster around the upper intercept (Fig. 5), the lower intercept is very sensitive to any contaminationcorrection bias. Thus the apparent agreement may be fortuitous. Portions of some of the meteorites (those that are nearly concordant in Fig. 5) do not

86

appear to have been greatly affected by the recent disturbance, at least on an ~ 100-mg-sized scale. The results for Barwell are consistent with the apparently undisturbed K - A r system of this meteorite [9]. However, a detailed study of Barwell [18] showed that there was a recent episode of U - T h - P b fractionation among its constituent phases. A semi-independent check of the terrestrial Pb correction procedures can be obtained from the Th-U-Pb data. The 2°Tpb/2°spb ratios of terrestrial Pb contamination relative to Cation Diablo troilite Pb are ~ 0.59-0.67 whereas the same ratio in the meteorites ranges from ~ 0.4 (Harleton) to ~ 0 . 8 (Bruderheim) as a result of the variation in T h / U ratios among the samples (Table 1). 2°Tpb/2°rpb ratios of terrestrial Pb relative to Cation Diablo troilite are ~0.55-0.63. The meteorites have nearly constant ratios of ~ 0.600.62. Therefore, removal of terrestrial Pb contamination generally tends to increase the 2°Tpb/2°rpb age but may either increase or decrease the 2°Tpb/2°apb age. Furthermore, the 2°7pb/2°spb ages of samples w i t h 2 3 2 T h / 2 3 8 U ratios much higher or much lower than 4.0 will be more affected by terrestrial contamination than will the 2°7pb/E°6pb ages.

ll31

/

L~

4.2. Single-stage U-Th-Pb evolution

_ .,/,~--iLt~' HAl

"~

~f.~ ~

Th-U-Pb data corrected for the troihte apparent initial Pb are shown in Fig. 6. The data for the L5-L6 form a cluster about the concordia curve rather than a line. The centroid of this cluster (CEN in Fig. 6) is, within error, concordant and yields an age of 4569---32 My which is in good agreement with the 2°7pb/2°rpb age in Fig. 5. The solid line in Fig. 6 is a 4 5 5 3 - 400 My reference chord from Fig. 5. The Tennasilm data plot far below and to the right of the reference line, indicating that these data may be grossly overcorrected, at least for 2°8pb. Comparing Figs. 5 and 6 one can see that the majority of the individual analyses of L5-L6 chondrites plot on the same side of the reference isochron in both diagrams. The deviation for these samples can therefore be explained wholly in terms of overcorrection or undercorrection for terrestrial Pb. However, data for some samples such as Bruderheim (BR2) plot on opposite sides of the reference isochron in the two figures. These data may reflect a disturbance to the U-Th-Pb system early in the meteorites' histories, but may also reflect recent T h / U fractionation, errors in the Th concentration, or as is apparently the case with Tennasilm, the troilite apparent initial Pb may not adequately approximate the true mixture of indigenous initial Pb and contaminant Pb for all isotopes. Consequently, resolution of the ages of any early events must await detailed studies of individual meteorites.

H^2 ~

~ ~I

CeNtroln

T- 4589 ± 32 My

o

0

L~

TF.3

B.18 / / ~ , TEl 207pb/235 U

Fig. 6. U-Th-Pb concordia plot for chondrite troilite-corrected data. The data form a cluster rather than a straight line. The centroid of the cluster (CEN) is, within error, concordant at 4569-+-32 My. The Tennasilm data (TEl, TE2, and TE3) plot below and far to the right of the reference isochron and appear to be grossly overcorrected for 2°spb.

The second approach to correction of the data for terrestrial Pb contamination involves the assumption of closed-system U-Pb evolution. That is, the U-Pb data are assumed to be precisely concordant, and all deviations from concordancy in the Pb-excess direction are assumed to be solely the result of terrestrial Pb contamination (see Manhrs and Allrgre [5]). These corrections require that the Pb isotopic composition of the contaminant be known. The contaminant Pb isotopic compositions have been calculated for each of the meteorites analyzed by assuming that the apparent initial Pb in the troilite separates is a mixture of Cation Diablo Pb and "normal" terrestrial Pb (the terrestrial Pb field in

87

Fig. 3). The intersection of a line through Cation Diablo and the troilite Pb with the best-fit line through the terrestrial Pb field (2°7pb/2°4pb= 0.0898 × 2°6pb/2°4pb + 13.92) yields the 2°6pb/2°4pb and 2°Tpb/2°4pb ratios of the contaminant. The 2°8pb/2°4pb ratio is calculated from the 2°rpb/2°4Pb ratio and the best-fit approximation to the terrestrial Pb field (2°spb/2°4pb= 0.697 ×2°6pb/2°4pb + 25.3). These equations can only be used in the range of 2°rPb/2°4pb ~ 16-19. Furthermore, these calculations require the assumption of closed-system U-Th-Pb evolution for the troilite as well as the bulk meteorite. Pb with the calculated isotopic composition is then used as a second blank correction such that the U-Pb data become concordant when Cation Diablo troilite Pb is used as the initial Pb. The calculated isotopic compositions of the Pb contaminants are shown in Table 5. The calculated 2°8pb/2°4pb ratios of the contaminants in the three L6 chondrites (and the 2°6pb/2°4pb and 2°7pb/2°4pb ratios in Modoc) are slightly lower than the apparent initial ratios in the troilite--that is, the troilite apparent initial ratios plot slightly above the assumed terrestrial trend (cf. Fig. 3). Nevertheless, the calculated contaminant ratios were used in order to maintain consistency. It should also be pointed out that the 2°spb/2°4pb a n d 2°6pb/E°4pb ratios are not as well-correlated in blank Pb's as a r e 2°7pb/E°4pb and 2°6pb/2°4pb ratios. In the ensuing discussion, the uncertainties

in the terrestrial contaminant were calculated assuming perfect correlation. That is, the uncertainties in the best-fit lines through the terrestrial contaminants trends (e.g. Fig. 3) were not incorporated into the uncertainties in the individual calculated contaminant Pb isotopic compositions. Consequently, the uncertainties in the individual c o n t a m i n a t i o n - c o r r e c t e d 2°7pb//2°6pb and 2°Tpb/2°8pb ages may be slightly underestimated. The terrestrial-Pb-corrected concordant 2°Tpb/2°6pb ages and resulting 2°Tpb/2°spb ages are shown in Table6. The columns "A6c" and "A8c" denote the percent discordance of 2°rpb/238U and 2°8pb/232Th, respectively, after correction. A positive number indicates deviation above the curve and a negative number below. The data have been corrected such that A6c= 0, whenever possible. The maximum correction possible by this method is obtained by removal of all 2°4Pb and proportionate amounts of the other isotopes as terrestrial contamination. Some of the samples (those with A6c>0) still plot - 1 - 1 0 % above the concordia curve even when the maximum correction is applied. Thus, it appears that some of the excess radiogenic Pb cannot be attributed to terrestrial Pb contamination alone [ 18,20]. Given a simple two-stage model (Fig. 5), these discordancies imply that Pb was redistributed over area larger than the samples analyzed during a recent event [ 18,20]. In order to evaluate the ages in Table6, the

TABLE 5 Calculated contaminant Pb isotopic composition

a

Sample

Class

206pb/20,* Pb

207pb/204 Pb

2ospb/2O4 Pb

Bjurbrle Tennasilm Knyahinya Barwell Bruderheim Harleton Modoc

L4 L4 L5 L5-6 L6 L6 L6

18.7 ± 0.1 18.4-'- 0.5 18.1 -+0.8 18.3--+0.1 18.5 -4-0.3 18.5 -+ 0.2 18.220.1

15.60 ± 0.02 15.58 ± 0.05 15.5420.08 15.56--+0.02 15.58 -± 0.03 15.58 ± 0.02 15.56-+0.02

38.3 ± 0.1 38. I ± 0.4 37.9--+0.6 38.1 --+0.1 38.2 ± 0.3 38.2 -4-0.2 38.020.1

Calculated from the troilite apparent initial Pb assuming that these compositions lie on a mixing line between Cation Diablo troilite Pb and "normal" terrestrial Pb. "Normal" terrestrial Pb is expressed by the equations: (I) 2 ° T p b / 2 ° 4 P b = 0 . 0 8 9 8 × 2 ° r p b / 2 ° 4 p b + 13.92 (2) 2°spb/2°4pb =0.697 × 2°rpb/2°4pb + 25.3.

a

--+0.6 ±0.7 -+0.8

+-1.2 -----6

--1.0 -+0.7

0 -+0.9 +0.9±0.7 0 -+0.6

0 -- 1.1 0 -----0.8 +5.4±0.9

0 0

+ 2 . 8 - + 1.3 ÷ 1 . 2 ± 1.0

+ 1 . 9 ± 1.9 0 --'-0.7 +10.8-+0.8

0 0 0

0 0

A6c b

4555 -- 23

4554±56 4555±26 (4393--15)

4533 ± 16

4552-'- 16 4532-+15 4 5 2 4 ± 12

(4403-+39) 4584±14

4578--25 -

4534±20

4540±39 4514-+16 4544-+12

4580 ± 16

4685±35 4585-----26 4571 --26

4642 ± 30

4620-----26 4 6 5 2 -+- 18

2°7pb/2°spb

-0.2-+4.2 -0.7--2 +12.8-+1.0

- 1 -- 1.5 +1.2±1.0 + 9 . 9 ± 1.0

+ 1 6 --3 -5.0±0.9

+ 1 . 0 ± 1.8 -

+3.5±3.1 +4.1±1.2 +11.2±0.9

-2.1 ±2.8 - 4 . 4 - + 1.8 --I.l--+l.9

-3.2--2.0 --6.2-+ 1.2

A8c b

0.47 0.52 0.78

0.40 0.28 0.42

0.21 0.11

0.15 0.23

0.08 0.09 0.12

I.l 1.6 1.4

0.27 0.43

whole-rock

1.4 -

1.9 -

0.51 -

2.6 -

0.07 -

0.40 -

0.59 -

troilite

C o n t a m i n a t i o n 2°4pb (ppb) a

0.042 0 0.100

0.024 0.065 0

0.027 0.090

0 0

0 0.033 0

4.0 3.9 4.0

1.74 1.76

whole-rock

I n d i g e n o u s 204 Pb (ppb) a

~0 -

0.06 -

0.05 -

0.05 -

0.02 -

3.47 -

0.20 -

troilite

a C a l c u l a t e d a s s u m i n g that all excess Pb is terrestrial w i t h Pb isotopic c o m p o s i t i o n s from T a b l e 5. b Excess ( + ) or deficiency ( - ) in radiogenic 2 ° t p b or 2°spb,in %. The d a t a have been corrected such that A 6 c = 0 or until all 2°4pb has b e e n r e m o v e d as terrestrial Pb. c 2o7pb/206pb and 2 ° 7 p b / 2 ° s p b m o d e l ages of these s a m p l e s are in gross d i s a g r e e m e n t a n d were therefore not included in the c a l c u l a t i o n s of m e a n ages.

4546 ± 12

4552-- 5 4540± 5 (4514-+ 2) c

Average

4 5 5 0 ± 14

Modoc

-1 -2 -3

Average

L6

Harleton

4543± 5 4544-- 4 4566± 5

L6

Bruderheim-1 -2

-1 -2 -3

4560-+ 4 ( 4 5 5 1 ± 5) c 4535± 4

L6

Average

4560-- 5 4560± 5

L5-6

Barwell

-1 -2

4552-+ 4

Average

4556± 6 4554± 6 4549-+ 3

L5

Knyahinya

-I -2 -3

4552 ± 13

Average

L4

4 5 6 4 ± 14 4542± 9 4 5 5 9 - - 11

-l -2 -3

Tennasilm

6

4590 ±

4588 -4-15 4591 ± 7

2°7pb/2°tpb

M o d e l ages (My)

Average

L4

Bjurb/51e

-I -2

Class

Sample

C o n t a m i n a t i o n - c o r r e c t e d c o n c o r d a n t 2°7pb//2°tpb ages a n d resulting 2 ° 7 p b / 2 ° s p b ages a n d 2°4pb a b u n d a n c e s

TABLE 6

89 following interpretations should be kept in mind: those samples that show negative A8c and 2°Tpb/2°8pb model ages significantly older than the 2°7pb/2°rpb model ages are generally consistent with overcorrection for terrestrial Pb, and conversely, those samples that show positive A8c and younger 2°Tpb/2°spb ages may be undercorrected. The data for the L5 and L6 chondrites in Table6 generally show fair agreement between both the 2°7pb/2°~pb and 2°7pb/2°spb model ages, and A6c and A8c. The most notable exceptions are Bruderheim-1 and Modoc-3. These samples have the lowest T h / U ratios of all the samples, so the lack of agreement between the 2°Tpb/2°6pb and 2°7pb/2°spb ages could reflect ~ 10% errors in the Th concentrations, resulting from lack of equilibration between sample and spike during decomposition of the samples (e.g. [14]). However, it should be pointed out that these meteorites may have experienced a rather complex history. A detailed U-Pb study of Bruderheim [20] has shown that the U-Pb systems of portions of this meteorite have experienced at least three stages of evolution (specifically, see samples 3, 6, and 8 in Gale et al. [20]). Furthermore, the K-Ar system of this meteorite has also been severely disturbed [9]. The Sm-Nd system of Modoc [49] has suffered some small-scale open-system behavior at ~ 4000 My ago, although the whole-rock Sm-Nd system appears to have remained closed over at last --4550 My. Consequently, the discrepancies for these two samples could be the result of applying a simple model to complex samples. The somewhat young 2°Tpb/2°rpb model age for Modoc-3 could also mean that the age, discrepancy results in part from undercorrection of the data for terrestrial Pb (the sample when properly corrected would show significant recent Pb loss). In any case the model ages of these two samples are not included in the following calculations of mean ages. The mean 2°7pb/2°rpb concordant age of the L5-L6 chondrites is 4550-+ 5 My (20, S / ( N - - l) = 17), and the mean 2°Tpb//2°spb age is 4544 - 15 My ( S / ( N - 1) = 8). The somewhat lower amounts of scatter in both the 2°7pb//E°6pb and 2°Tpb/E°Spb model ages relative to those obtained by the previous procedure suggests that this procedure represents a slightly better approach. However, both

methods produce the same 2°Tpb/2°rpb mean age, and in both instances the mean 2°Tpb/2°spb ages are in agreement with the 2°7pb/2°rpb ages. The large amount of scatter which results, regardless of which correction procedure is used, probably reflects overcorrection or undercorrection of individual samples as evidenced by the fact that the mean ages of individual meteorites (Table 6) show considerably less scatter. It is also possible that the scatter reflects small differences in the ages of these samples, resulting from partial resetting of the U-Th-Pb system during thermal or impact metamorphism. For example, the young 2°Tpb/2°rpb age of Bruderheim-2 (4535 My), yet apparent overcorrection for terrestrial Pb as inferred from the negative A8c and older 207pb/2OSpb age suggest that some non-recent disturbance has affected the U-Th-Pb system of this sample. The 4535-My age of Bruderheim-2 is in excellent agreement with the 4535 ± 6 My whole-rock isochron age of this meteorite reported by Gale et al. [20]. The majority of the samples analyzed by Gale et al. showed considerable amounts (up to 80%) of excess radiogenic Pb and, although it was shown that terrestrial Pb contamination alone could not account for all of the excess radiogenic Pb in some samples, the potential terrestrial contamination problem was not fully evaluated. Also, as previously mentioned, some samples were not consistent with a two-stage model, even allowing for terrestrial Pb contamination. Thus the apparent young age may be the result" of a disturbance to the U-Th-Pb system early in the meteorites' history (~ 4535 My). The 4550 My mean age obtained by both procedures is in excellent agreement with the 4552 My 2°Tpb/2°rpb model ages of two nearly-concordant L5 chondrites [5], and with the 4555 ± 4 My concordant model age of these two meteorites corrected for an assumed Pb-contaminant isotopic composition [5]. An internal U-Th-Pb study of the Barwell chondrite [18] yielded 4564± 5 My and - 1 7 0 - +250 My U-Pb concordia intercepts when the data were corrected for the Barwell troilite apparent initial Pb. The upper intercept is slightly older than the mean concordant age, although the negative lower intercept suggests that the data were

90 slightly overcorrected. In this regard, it is interesting to note that both over and undercorrection of a suite of samples for terrestrial Pb can produce an erroneously young lower intercept if the model age of the contaminant is younger than the true age of the samples (i.e. if the contaminant plots to the right of the reference isochron in Fig. 3). Therefore, the 4564 -+ 5 M y age of Barwell should probably be viewed as an upper limit to the true age. Thus, with the possible exception of Bruderheim, there is no evidence for primary (formation or equilibration) age differences of ~> 15 My among the L5-L6 chondrites analyzed. The data for the L4 chondrites in Table 6 do not permit a simple age interpretation. The old 2°Tpb/2°apb model ages and "normal" discordancy of the 2°8Pb data (negative A8c) suggest that these samples are all overcorrected for terrestrial Pb, although the effect is somewhat less for Tennasilm. The mean concordant 2°7pb/2°6pb age of Tennasilm of 4 5 5 2 +- 13 My is in excellent agreement with that of the L5-L6 chondrites whereas the 2°Tpb/2°6pb age of Bjurb61e is distinctly older (4590 -+ 6 My). Although this age is in agreement with the ~ 4580-My formation age of chondrites [5] inferred from U-Pb analyses of four E chondrites and contamination-corrected concordant model ages of two H4 chondrites, the apparent overcorrection of the data evidenced by the 2°Tpb/2°8pb data suggests that the agreement may be fortuitous. The 2°7pb/E°6pb and 2°7pb//2°8pb ages of the Bjurb61e samples cannot be made to agree regardless of the amount of terrestrial Pb removed. In fact, the closest agreement is obtained with removal of no pre-analysis Pb contamination (Table2, A6c~ +25 to +40%). Similarly, best agreement between the 2°7pb/E°6pb and 2°Tpb//2°spb model ages of Tennasilm is reached at A6c = + 20 to + 50%. In any case, it appears that not all of the excess radiogenic Pb in these two I_.4 chondrites can be attributed to terrestrial Pb contamination. The discordancies may therefore be due in part to (1) recent (terrestrial) U loss, a n d / o r (2) a recent disturbance which caused Pb to be redistributed over areas larger than the samples analyzed. The discrepancies in the concordant ages (Table 6) for both meteorites could be reconciled

by ~ 3-6% U loss, resulting in a 5-10 My reduction in the concordant 2°7pb/2°6pb model ages. Such an interpretation for BjurbSle is not unreasonable in view of the fact that this meteorite fell into seawater [23]. The lack of agreement between the 2°Tpb/2°rpb and 2°Tpb/2°spb model ages of Bjurbrle could also reflect discordancy produced by a recent, but distinctly non-zero (e.g. ~ 108 years) disturbance. Given this possibility, then the model ages of Bjurbrle would of course not be valid ages, but would represent upper limits to the true age of this meteorite. Consequently, although the possibility of an older age for Bjurbrle cannot be ruled out, the better agreement between the 2°Tpb/2°rpb and 2°Tpb/2°spb ages for Tennasilm suggests that the 4552-+ 13 My mean 2°7pb/2°rpb age of this meteorite may be closer to the true age of the L4 chondrites. A preliminary study of three L3 chondrites [26] has shown that portions of these meteorites also contain large amounts of excess radiogenic Pb (A6c~ +50 to +150%). The "old" model ages ( > 4550 My) and Th-U-Pb relationships ([26], and unpublished data from this laboratory) suggest that the excess Pb in these meteorites is not entirely due to terrestrial Pb contamination. An age of 4550-+ 60 My for three L3 chondrites was obtained from the Cation Diablo troilite-corrected U - P b data whereas the 2°Tpb/2°4pb vs. 2°rpb/2°4pb data yielded a 4585-+ 10 My apparent age [26]. Pb isotopic data for magnetic and non-magnetic separates from these meteorites suggest that the 4585 My apparent age is not be a true age, but an artifact of a recent ( ~ 3 × 108 years) disturbance to the U-Th-Pb systems of these meteorites [26]. The combined results suggest that recent but distinctly non-zero age parent body-wide redistribution of Pb relative to U and Th may have taken place. The recent event could correspond to the ~ 500-My event recorded in the K-Ar systems of some L chondrites (e.g. [7-9]). This event produced net enrichment of Pb relative to U in L3 and some L4 chondrites and only slight enrichment or depletion in L5-L6 chondrites on an 100-rag-sized scale. In order to maintain mass balance, then some portion of the parent body would have to have suffered significant Pb-

91

depletion relative to U. Apparently this portion has not been analyzed. Regardless of whether small age differences exist among these meteorites or for that matter which method of correction is used, there is still a significant porblem regarding the "age" of chondrites: the U-Th-Pb, Rb-Sr, and K-Ar systems yield systematically different ages. The ~ 4550 My U-ThPb age obtained here is significantly older than the 4480 ± 30 My Ar-Ar ages obtained for unshocked ordinary chondrites [27], and the 4504 ± 15 My R b - S r whole-rock isochron for chondritic meteorites [45]. However, it should be pointed out, that H chondrites alone yield 4518 ± 26 My Rb-Sr age [46], which is only slightly different from the U-Th-Pb ages presented here. If these discrepancies are due to errors in 87Rb and 4°K decay constants (relative to those for U), this would imply that h87Rb and ~,4°K should be reduced by ~0.6-1% and ~ 1.5%, respectively. On the other hand, it is also possible that the ages reflect different closure temperatures for the three systems. If slow cooling rates of the order of 1-10°C//My are considered [47,48] and if the closure temperature of a system is controlled to a significant extent by the volatility of the parent rather than the daughter isotope, the observed systematic differences might be expected. This is a problem which certainly merits further investigation. 4.3. Terrestrial-corrected Pb abundances

The right-hand portion of Table 6 shows the calculated contaminant and indigenous 2°4pb concentrations in the whole-rocks and troilite separates, again assuming that the U-Pb data are concordant and that the Pb in the troilite is a mixture of terrestrial and Cation Diablo-type Pb. The whole-rocks analyzed appear to contain ~0.08-1.6 ppb of terrestrial 2°4pb. This corresponds to ~ 6 - 1 2 0 ppb total Pb. In view of the 200 ppb and 2 p p m terrestrial Pb found in Nakhla [40] and Moama [42], respectively, and the large amounts of terrestrial Pb (several ppm) found in some meteorite fusion crusts (e.g. [18,22]), even the larger amounts of terrestrial Pb do not seem unreasonable. In addition, the sintering compounds used in diamond blades are known to

contain large amounts of Pb [43]. There are no clear-cut correlations between the amount of terrestrial Pb and sample handling, although some generalizations can be made: (1) samples chipped from larger stones that were protected by fusion crusts (Knyahinya, Barwell, Bruderheim) generally show low amounts of contamination; (2) the only sawn sample, Harleton, shows a large amount of terrestrial contamination. In addition, the most altered sample, Modoc, contains large amounts of terrestrial Pb. The Moama eucrite, which was grossly contaminated with terrestrial Pb [42], was also quite badly oxidized. There are no obvious reasons why Tennasilm should be the most contaminated, provided that this sample was not grossly overcorrected for terrestrial Pb. However, it should be mentioned that the allocation of this sample consisted of only 3 g. Thus it is possible that some portions of the exposed surface were inadvertently incorporated into the chips analyzed. Even the "interior" of this sample was no more than a few millimeters from an exposed surface. The troilite separates generally contain 1-5 times as much terrestrial Pb as the whole-rocks. The reasons why troilite would be enriched in terrestrial Pb relative to the bulk sample are not known. It is, of course, possible that at least some of the contamination was induced during the separation procedures (see also Tflton and Chen [28]). However, the amount of terrestrial Pb in the troilite separates correlates well with the amount of contaminant Pb in the bulk sample even though sample preparation procedures were quite different (see section 2.1). This suggests that the majority of the contamination occurred prior to separation and analysis. The major exception to this correlation is Barwell. Pb concentrations in three troilite separates from this meteorite varied from ~ 20 to 300 ppb [18], and although the isotopic composition of Pb in these three separates was uniform, it was distinctly different from both the terrestrial contaminant found in the fusion crust and the laboratory blank Pb. Furthermore this Pb was not easily leachable in 0.1N HBr, thus suggesting that the Pb was in the crystal lattice and not simply on the surface of the grains. These observa-

92 tions led us [18] to conclude that the excess radiogenic Pb in the Barwell troilite resulted from redistribution of U-Th-Pb during a recent thermal event, rather than terrestrial Pb contamination. In view of the data presented here, it is apparent that this original interpretation was incorrect. However, the source of the terrestrial Pb in the Barwell troilite remains unknown. If this Pb represents a mixture of terrestrial Pb in the fusion crust and laboratory blank Pb, then one would expect to see variations in the Pb isotopic composition that are correlated with the Pb concentration. Therefore, if the contamination is due to sample handling in this laboratory, then this Pb must have had a slightly different isotopic composition from that of a normal blank. The fact that the majority of this Pb was not removed by a dilute acid leach may suggest that the contaminant Pb was located in Feoxide alteration products associated with the troilite. A mean value of 0.029 --+0.024 ppb indigenous 2°4pb for the L5-L6 chondrites is obtained from the mean (unweighted) of the averages of the two or three analyses for each sample. This corresponds to 1.4--+ 1.1 ppb total initial Pb (assuming this Pb has the isotopic composition of Cation Diablo troilite Pb). Consequently, calculated contamination-corrected "2°6pb/2°4pb ratios for the L5-L6 chondrites are very radiogenic (~> 100-400). Initial Pb concentrations increase drastically from types 5 and 6 to type 4. 204pbin the Mez/5-Madaras L3 chondrite [3,4,26] is enriched by a factor of 4 × l03 relative to the L5-L6 chondrites and by a factor of ~ 60 relative to Bjurb61e. Troilite in these meteorites does not appear to be enriched in indigenous 204pb relative to the bulk samples. Troilite in the L5-L6 chondrites and Tennasilm has approximately the same amount of Pb as the whole-rock, whereas troilite in BjurbiSle contains only about 1/10 as much Pb as the whole-rock. No enrichment in Pb was observed in troilite from the L3 chondrite, Khohar [44]. In view of possible mineral impurities in the troilite separates, especially the Tennasilm troilite (Table 4, and section 3.3), the Pb contents in these separates should probably be viewed as upper limits to the Pb abundances in "pure" troilite.

5. Conclusions (1) U, Th, and Pb analyses of whole-rock and troilite separates from seven L chondrites suggest that the majority of the excess of radiogenic Pb and the wide variations in 2°Tpb/2°tpb model ages c o m m o n l y f o u n d in equilibrated ordinary chondrites are the result of terrestrial Pb contamination in these samples prior to analysis. The meteorites analyzed appear to contain from ~ 5 to 120 ppb terrestrial Pb. However, aliquots of some L5-L6 chondrites contain up to 10% excess radiogenic Pb which cannot be accounted for by terrestrial Pb contamination. Similarly, Th-U-Pb relationships in BjurbiSle suggest that a significant portion of the excess radiogenic Pb in this meteorite may not be the result of Pb contamination, but may be due to terrestrial U loss or a recent geologic disturbance to the U-Th-Pb system. (2) Pure and uncontaminated troilite in L5-L6 chondrites is essentially U, Th, and Pb, free (~> 3 ppb Pb). Furthermore, troilite in L4 chondrites is depleted in Pb by about one order of magnitude relative to the bulk sample. Consequently, the Pb isotopic composition of the troilite is very sensitive to terrestrial Pb contamination. The Pb isotopic compositions in the troilite can be used to determine the isotopic composition of the Pb contaminant, and Th-U-Pb data can be used in conjunction with U-Pb data to determine how much of the excess Pb in a given sample can be accounted for by terrestrial Pb. (3) L5-L6 chondrites yield an age of 4551 - 7 My when the troilite apparent initial Pb isotopic composition is used to approximate the proper mixture of indigenous initial Pb and terrestrial Pb. If a single-stage plus terrestrial contamination model is assumed, the data from the L5-L6 chondrites yield a 4550 --- 5 My age. Two analyses of Bjurbtle yield a 4590---6 My concordant 2°7pb/2°6pb age, but the 2°7pb/2°apb data suggest that the data have been overcorrected for terrestrial Pb. The Tennasilm data yield a 4552 - 13 My mean concordant 2°7pb/2°6pb age, and fairly good agreement between the 2°7pb/2°tpb and 2°7pb/2°8pb ages. Therefore, this age may represent a better approximation to the age of the L4 chondrites.

93 The results from Tennasilm and the L5-L6 chondrites indicate that there are no appreciable age differences a m o n g type 4 to type 6 L chondrites. Assuming that L5-L6 chondrites represent metamorphosed L3 or L4 material, then either metamorphism of the L chondrite parent body occurred within a short time after formation (~< 15 My), or the U-Th-Pb systems of all of the meteorites analyzed were either not reset or reset to the same extent during metamorphism. In a rigorous sense the 4550 My mean age of the L5-L6 chondrites should be viewed as an upper limit to the true equilibration age. However, the extremely low corrected 2°4pb abundances of the L5-L6 chondrites suggest near-complete degassing of this portion of the parent body during metamorphism - - g i v e n the assumption that volatile element depletion in equilibrated ordinary chondrites is indeed the result of metamorphism. Therefore, the U-Th-Pb systems of these meteorites would have been almost totally reset, and the 4550 My upper limit would be very close to the true equilibration age. Alternatively, if the 4590 My apparent age of BjurbiSle is a true age, then the combined data would support Manh6s and All6gre's [5] assertion that metamorphism occurred ~ 30 My after formation of the L-group parent body. Further studies of L3 and L4 chondrites will be required to resolve the significance of these apparent older ages. (4) The corrected U-Th-Pb ages are somewhat at odds with the Rb-Sr (4504 -+ 15, My) and Ar-Ar (4480 -+ 30) ages of chondritic meteorites. The discrepancies may be due to - 1.5% and 1~ errors, (relative to those for U) in the 4°K and 87Rb decay constants, respectively, or may reflect different closure ages for the three systems.

Acknowledgements I wish to thank M. Tatsumoto, Z.E. Peterman and T.R. Wildeman for numerous useful comments regarding this work. Thoughtful reviews by R. Hutchison, G. Manh6s, and G.R. Tilton resulted in considerable improvement to the manuscript. The following individuals generously provided

samples: R.S. Clarke, U.S. National Museum of Natural History (Harleton, Modoc); R.E. Folinsbee, University of Alberta (Bruderheim); R. Hutchison, British Museum (Natural History) (Barwcll, Knyahinya); and E. Olsen, Chicago Field Museum (Bjurb61e, Tennasilm). I also thank K.R. Ludwig for the use of his plotting/regression program, Ed Daugs and Phil Heidt for laboratory maintainance and assistance, April Vuletich for mass spectrometry assistance and Marge Henneck for typing the manuscript. This work was funded by NASA Interagency Transfer T-783H.

References 1 N.H. Gale, J.W. Arden and R. Hutchison, Uranium-lead chronology of chondritic meteorites, Nature 240 (1972) 56-57. 2 M. Tatsumoto, R.J. Knight and C.J. Ali6gre, Time difference in the formation of meteorites as determined from the ratio of lead-207 to lead-206, Science 180 (1973) 12791283. 3 G.R. Tilton, Isotopic lead ages of chondritic meteorites, Earth Planet. Sci. Lett. 19 (1973) 321-329. 4 J.M. Huey and T.P. Kohman, 2°Tpb-2°6pbisochron and age of chondrites, J. Geophys. Res. 78 (1973) 3227-3244. 5 G. Manh~ and C.J. All~gre, Time differences as determined from the ratio of lead-207 to lead-206 in concordant meteorites, Meteoritics 13 (1978) 543-548. 6 G.W. Reed, K. Kigoshi and A.L.-Turkevich, Determinations of Heavy elements in meteorites by activation analysis, Geochim. Cosmochim.Acta 20 (1960) 122'-140. 7 G. Turner, Thermal histories of meteorites by the 39Ar-4°Ar method, in: Meteorite Research, P.M. Millman, ed. Reidel, Dordrecht, 1969)407-417. 8 D.D. Bogard, L. Husain and R.J. Wright, 4°Ar-39Ardating of collisional events in chondrite parent bodies, J. Geophys. Res. 81 (1976) 5664-6578. 9 G. Turner, J.A. Miller and R.L. Grasty, The thermal history of the Bruderheim meteorite, Earth Planet. Sci. Lett. I (1966) 155-157. l0 K. Gopalan and G.W, Wetherill, Rubidium-strontium studies on black hypersthene chondrites: effects of shock and reheating, J. Geophys. Res. 76 (1971) 8484-8492. l 1 J.F. Minster and C.J. All~gre, S7Rb-S7Sr dating of L chondrites: effects of Shock and Brecciation, Meteorites 14 (1979) 235-248. 12 D.M. Unruli, N. Nakamura and M. Tatsumoto, History of the Pasamonte achondrite: Relative susceptibility of the Sm-Nd, Rb-Sr and U-Pb systems to metamorphic effects, Earth Planet. Sci. Lett. 37 (1977) 1-12. 13 J.H. Chen and G.R. Tilton, Isotopic lead investigations on the Allende carbonaceouschondrite, Geochim. Cosmochim. Acta 40 (1976) 635-643.

94 14 M. Tatsumoto, D.M. Unruh and G.A. Desb0rough, U-ThPb and Rb-Sr systematics of Allende and U-Th-Pb systematics of Orgueil, Geochim. Cosmochim. Acta 40 (1976) 617-634. 15 N.H. Gale, J.W. Arden and R. Hutehison, U-Pb studies of the Appley Bridge meteorite, Naturwissenschaften 66 (1979) 419-420. 16 A.G.W. Cameron, Are large time differences in meteorite formation real?, Nature 246 (1973) 30-32. 17 J.T. Wasson, Meteorites (Springer-Vedag, New York, N.Y., 1974) Chapter XVIII, pp. 181-205. 18 D.M. Unruh, R. Hutchison and M. Tatsumoto, U-Th-Pb age of the Barwell chondrite: anatomy of a discordant meteorite, Proc. 10th Lunar Planet. Sci. Conf. (1979) 10111030. 19 M.C.B. Abranches, J.W. Arden and N.H. Gale, Uranium lead abundances and isotopic studies in the chondrites Richardton and Farmington, Earth Planet. Sci. Lett. 46 (1980) 311-322. 20 N.H. Gale, J.W. Arden and M.C.B. Abranches, Uraniumlead age of the Bruderheim L6 chondrite and the 500-Ma shock event in the L-group parent body, Earth Planet. Sci. Lett. 48 (1980) 311-324. 21 V.M. Oversby, Lead, In: Handbook of Elemental Abundances in Meteorites (Gordon and Breach, New York, N.Y., 1971) 499-510. 22 D.M. Unruh, M. Tatsumoto and R. Hutchison, U-Pb study of the Richardton H5 chondrite, Meteoritics 14 (1979) 551-553. 23 M.H. Hey, Catalogue of Meteorites (British Museum (Natural History), London, 1966) 637 pp. 24 R. Hutchison, A.W.R. Bevan and J.M. Hall, Appendix to the Catalogue of Meteorites (British Museum (Natural History) London, 1977) 297 pp. 25 D.M. Unruh and M. Tatsumoto, A uniform U-Pb age for L chondrites and a method for correcting for terrestrial Pb contamination, Meteoritics 15 (1980) 380-381. 26 D.M. Unruh, M. Tatsumoto and R. Hutchison, U-Pb study of L3 chondrites: old ages or young disturbances, in: Lunar and Planetary Science XII (The Lunar and Planetary Institute, Houston, Texas, 1981) 1109-1111. 27 G. Turner, M.C. Earight and P.H. Cadogan, The early history of chondrite parent bodies inferred from 4°Ar-39Ar ages, Proc. 9th Lunar Planet. Sci. Conf. (1978) 989-1025. 28 G.R. Tilton and J.H. Chen, Lead isotope systematics of three Apollo 17 mare basalts, Proc. 10th Lunar Planet. Sci. Conf. (1979) 259-274. 29 M. Tatsumoto and D.M. Unruh, KREEP basalt age: grain by grain U-Th-Pb systematics study of the quartzmonzodiorite clast 15405.88, Proc. 7th Lunar Sci. Conf. (1976) 2107-2129. 30 K.R. Ludwig, Calculation of uncertainities of U-Pb isotope data, Earth Planet. SCi. Lett. 46 (1980) 212-220. 31 G. Crozaz, Uranium and thorium micro distributions in stony meteorites, Geochim. Cosmochim. Acta 43 (1979) 127-136. 32 J.C. Laul, I.Z. Pelley and M.F. Lipschutz, Thallium contents of chondrites, Geochim. Cosmochim. Acta 34 (1970) 909-920. 33 J.C. Laui, D.R. Case, F. Schmidt-Bleek and M.F. Lipschutz, Bismuth contents of chondrites, Geochim. Cosmochim. Acta 34 (1970) 89-103.

34 D.R. Case, J.C. Laul, I.Z. Pelley, M.A. Wechter, F. Schmidt -Bleek and M.F. Lipschutz, Abundance patterns of thirteen trace elements in primitive carbonaceous and unequilibrated ordinary chondrites, Geochim. Cosmochim. Acta 37 (1973) 19-33. 35 D. Heyman, in: Handbook of Elemental Abundances in Meteorites (Gordon and Breach, New York, N.Y., 1971) 29-66. 36 G.W. Reed and S. Jovanovic, Mercury in chondrites, J. Geophys. Res. 72 (1967) 2219-2228. 37 J.H. Chen and G.J. Wasserburg, A search for isotopic anomalies in uranium, Geophys. Res. Lett. 7 (1979) 275278. 38 D.M. Unruh and M. Tatsumoto, Excess Pb on L chondrites: is it terrestrial?, in: Lunar and Planetary Science XI (The Lunar and Planetary Institute, Houston, Texas, 1980) 1184 -1186. 39 G.W. Wetherill, Discordant uranium-lead ages, Trans. Am. Geophys. Union 37 (1956) 320-326. 40 N. Nakamura, D.M. Unruh and M. Tatsumoto, The young magmatic event in the Nakhla achondrite parent body, in: Short Papers of the Fourth International conference, Geochronology, Cosmochronology, Isotope Geology, R.E. Zartman, ed. (U.S. Geol. Surv., Open-File Rep. 78-701 (1978) 305-306. 41 F. Oberli, J.C. Huneke and G.J. Wasserburg, U-Pb and K-At systematics of cataclysm and precataclysm lunar impactites, In: Lunar and Planetary Science X (The Lunar Science Institute, Houston, Texas, 1979) 940-942. 42 J. Hamet, N. Nakamura, D.M. Unruh and M. Tatsumoto, Origin and History of the adcumulate eucrite, Moama, as inferred from REE abundances, Sm-Nd and U-Pb systematics, Proc. 9th Lunar Planet. Sci. Conf. (1978) 11151136. 43 P.D. Nune~ D.M. Unrnh and M. Tatsumoto, U-Th-Pb systematics of Apollo 16 samples 60018, 60025, and 64435; and the continuing problem of terrestrial Pb contamination of Lunar samples, in: Lunar Sample Studies, NASA SP-418 (1977) 61-69. 44 D.S. Woolum, R. Mascitelli, D.S. Burnett and L.S. August, Evidence for the siderophilic behavior of meteoritic bismuth, Proc. 9th Lunar Planet. Sci. Conf. (1978) 1173-1186. 45 J.F. Minster and C.J. All~gre, 87Rb-87Sr dating of L and LL chondrites: effects of shock and brecciation, Meteoritics 13 (1978) 563-564. 46 J.F. Minster and C.J. All~gre, 87Rb/STSr chronology of H chondrites: constraint and speculations on the early evolution of their parent body, Earth Planet. Sci. Lett. 42 (1979) 333-347. 47 P. Pellas and D. StiSrzer, Colling histories of stony meteorites, In: Lunar Science VIII (The Lunar Science Institute, Houston, Texas, 1977) 762-764. 48 J.A. Wood, Chondrites, their metallic minerals, thermal histories and parent bodies, Icarus 6 (1967) 1-49. 49 N. Nakamura and M. Tatsumoto, A 4.0 b.y. impact metamorphism age of the Modoc L6 chondrite determined by the Sm-Nd method, Meteoritics 15 (1980) 334-335. 50 F. Tera and G.J. Wasserburg, U-Th-Pb systematics in three Apollo 14 basalts and the problem of initial Pb in lunar rocks, Earth Planet. Sci. Lett. 14 (1972) 281-304. 51 D. York, Least squares fitting of a straight line with correlated errors, Earth Planet. SCi. Lett. 5 (1969) 320-324.