New Lu–Hf and Pb–Pb age constraints on the earliest animal fossils

New Lu–Hf and Pb–Pb age constraints on the earliest animal fossils

Earth and Planetary Science Letters 201 (2002) 203^212 New Lu^Hf and Pb^Pb age constraints on the earliest animal fossil...

348KB Sizes 1 Downloads 17 Views

Earth and Planetary Science Letters 201 (2002) 203^212

New Lu^Hf and Pb^Pb age constraints on the earliest animal fossils Gry H. Barfod a;b; , Francis Albare'de c , Andrew H. Knoll d , Shuhai Xiao e , Philippe Te¤louk c , Robert Frei a;b , Joel Baker a b

a Danish Lithosphere Centre, ster Voldgade 10, L, 1350 Copenhagen, Denmark Geological Institute, University of Copenhagen, ster Voldgade 10, 1350 Copenhagen K, Denmark c Ecole Normale Supe¤rieure de Lyon, 46 Alle¤e d’Italie, 69364 Lyon Cedex 7, France d Botanical Museum, Harvard University, 26 Oxford Street, Cambridge, MA 02138, USA e Department of Geology, Tulane University, New Orleans, LA 70118, USA

Received 18 December 2001; received in revised form 15 April 2002; accepted 19 April 2002

Abstract The Neoproterozoic Doushantuo Formation, South China, preserves a unique assemblage of early multicellular fossils and overlies rocks, which are thought to have formed during an ice age of global extent. The age of this formation is thus critical for understanding the important biological and climatic events that occurred towards the end of the Proterozoic Eon. Until now, direct dating of sedimentary formations such as the Doushantuo has been difficult and associated with large uncertainties. Here, we show that dating of Doushantuo phosphorites by a novel Lu^Hf dating method and conventional Pb^Pb geochronometry independently yield ages of 584 9 26 Ma and 599.3 9 4.2 Ma, respectively. These ages are in agreement with bio- and chemostratigraphical observations and show that the Doushantuo animal remains predate diverse Ediacaran fossil assemblages, making them the oldest unambiguous remains of metazoans currently known. Furthermore, the Pb^Pb age for the post-glacial Doushantuo rocks suggests that the Neoproterozoic glaciation in China might predate glacial rocks in Eastern North America commonly associated with the younger (Marinoan) of two major Neoproterozoic glaciations. The combination of Lu^ Hf and Pb^Pb dating shows considerable potential for dating other phosphorite successions and future application of these methods could therefore provide further constraints on Proterozoic biological and environmental history. A 2002 Elsevier Science B.V. All rights reserved. Keywords: absolute age; biologic evolution; Doushantuo Formation; Guizhou China; Lu^Hf system; Neoproterozoic; Pb/Pb; phosphate rocks; geochronology; Sinian

1. Introduction

* Corresponding author, at address a. Tel.: +45-3-814-2663; Fax: +45-3-311-0878. E-mail address: [email protected] (G.H. Barfod).

The ages of Phanerozoic sedimentary formations are conventionally estimated from indirect methods based on bio- and chemostratigraphy. Combined, these form a geochronological framework, which is calibrated by U^Pb zircon dates as

0012-821X / 02 / $ ^ see front matter A 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 2 - 8 2 1 X ( 0 2 ) 0 0 6 8 7 - 8

EPSL 6251 17-6-02


G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

well as by K^Ar and Ar^Ar ages obtained from the few volcanic rocks that have been found intercalated with sediments (e.g., [1]). In older rocks, the accuracy of this framework decreases as the fossil record diminishes and the stable isotope reference curves become imprecise, at the same time as volcanic rocks within sedimentary strata are rare [2]. The ages of many sedimentary successions, which record key events in the biological and climatic history of the Proterozoic Earth, are therefore only poorly constrained. The Neoproterozoic Doushantuo Formation, South China, contains exquisitely preserved eggs and embryos produced by animals of unknown taxonomic a⁄nities [3,4], possible sponges [5], and branching camerate tubes thought to have been produced by stem cnidarians or, perhaps, stem eumetazoans [6]. These fossils place important constraints on the hypotheses of early metazoan evolution. The formation lies above glaciogenic rocks of the Nantuo tillite, which are generally correlated to the Marinoan ice age (V590^610 Ma), the younger of the two most prominent Neoproterozoic glaciations. The ‘Snowball Earth’ [7] and less extreme models (e.g., [8]) proposed to explain the Neoproterozoic glaciations also hint that a direct causal relationship may exist between global glaciation and animal diversi¢cation. To test such models, it is critical to resolve the chronology of the Neoproterozoic climatic and biologic events. However, as these events are often represented by formations such as the Doushantuo that lack ash layers, direct dating of sedimentary rocks is required. Success with direct dating of sediments has so far been limited by the often large scatter of Pb isotope data [9] and the generally low parent/ daughter ratios ( 6 0.2) observed for most other isotopic systems (e.g., Rb^Sr, Sm^Nd and Lu^Hf) [10]. Phosphates are typically rich in rare-earth elements and for this reason, the Lu^Hf isotopic system in sedimentary apatite is promising for chronology. Here, we report the ¢rst direct age determinations of Neoproterozoic phosphorites by the Lu^Hf system and compare the results with ages obtained by conventional Pb^Pb dating. Combined, these ages place important constraints on early animal evolution and the Neoproterozoic

glaciation occurring around 600 Ma. Although our data have some limitations due to the complexities inherent to radiometric dating of sedimentary rocks, the consistency in ages derived from two very di¡erent chronometers (Lu^Hf, Pb^Pb) leads us to believe that our method o¡ers a promising way to directly date phosphatic sediments.

2. Samples and methods The type-section of the Doushantuo Formation is located in the Yangtze Gorges area on the Yangtze platform, southeastern China (Fig. 1). Here, the formation consists of carbonates, shales and phosphatic shales disconformably overlying Nantuo glaciogenic rocks. We studied Doushantuo phosphorites, outcropping about 600 km southwest of Yangtze Gorges, in the Guizhou Province at Weng’an and Kaiyang [4,11] (Fig. 1). In these areas, the Doushantuo rocks consist solely of intercalated carbonates and phosphorite units, which locally overlie the Nantuo tillite. In the Weng’an area, the Doushantuo phosphorites contain the exceptional multicellular fossils described by Xiao et al. [3] and Li et al. [5] that are of a quality believed to be preserved only if phosphate precipitation took place within days of cell death [12]. The best exposed sections (40^ 50 m) in this area include two V10 m phosphaterich intervals (locally referred to as the lower and upper ores) separated by a sub-aerial exposure surface [4]. We studied samples from the V15 m Lanmaao and Mofen pro¢les which, due to limited exposure, only include one of these phosphorite intervals. However, the phosphorites in these pro¢les contain fossils relating them to the upper ore [13] (Fig. 1). In the Kaiyang area, approximately 50 km west of Weng’an, Nantuo rocks are overlain by up to 20 m mud and siltstone, possibly belonging to the Doushantuo succession, which again is overlain by Doushantuo carbonates and phosphorites [4]. At Kaiyang, samples were examined from the Yongshaba pro¢le, which comprises about 10 m of exposed strata [13] (Fig. 1). As no fossils are observed, it is not known whether the phospho-

EPSL 6251 17-6-02

G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

Fig. 1. Paleogeographic map showing the location of the Yangtze platform (modi¢ed after [11]). Inset shows the locations on the Yangtze platform of the Doushantuo type-section at Yangtze Gorges and the Doushantuo phosphorites in the Guizhou Province at Weng’an and Kaiyang (stars; modi¢ed after [11]). The thickness of the Nantuo tillite increases from northwest to southeast on the South China Block. Stratigraphic columns of the Doushantuo Formation are from the Lanmaao and Mofen pro¢les at Weng’an and the Yongshaba pro¢le at Kaiyang (simpli¢ed after [13]). Included in the pro¢les are the approximate stratigraphic positions of studied samples.

rites here correlate to the lower or upper phosphorite ore exposed at Weng’an. The stratigraphic levels from which the samples were collected are shown on the stratigraphic columns in Fig. 1. The phosphorite samples were crushed to 6 200 Wm and separated from quartz and, in the case of the Yongshaba samples, from


sul¢des by settling in heavy liquids (bromoform and methylene-iodide). After ¢nal handpicking, yellowish, transparent apatite remained, in which few clay particles could be distinguished. Lu^Hf and Pb were isolated from 0.5 g and 0.1 g apatite fractions, which were dissolved in 6 N and 1 N HCl, respectively. For Lu^Hf separation, spikes of 176 Lu and 180 Hf were added separately to the samples. After sample^spike equilibration, all dissolved sample solutions were centrifuged in order to remove undissolved clay particles. The high abundance of £uorine in phosphates represents a signi¢cant problem due to the strong association between the dissolved Hf and £uorine ions. This was overcome by expelling £uorides in the presence of HClO4 . Lutetium and hafnium extraction was performed on HDEHP columns followed by puri¢cation of the Lu fraction from Fe on anion-exchange columns (Barfod et al., in prep.). For Pb separation, undissolved clay particles were removed by centrifugation from the dissolved apatite solutions prior to Pb separation by standard anion exchange techniques. To estimate the possible e¡ects from the clay on the Lu and Hf isotope systematics, Lu and Hf separation was performed on a second subset of the Yongshaba samples, where the clay was not removed. During the perchlorite evaporation step, the clay particles therefore dissolved, resulting in integrated phosphorite^clay analyses (Yongshaba bulk analyses in future references). For estimation of the Pb isotopic signal from the clay, we analyzed clay fractions, which had been separated from the phosphorites. The clay fractions were dissolved in a concentrated HNO3 ^HF mixture and taken separately through Pb column chemistry. All isotopic Lu and Hf analyses were obtained in static mode on the multiple collector-inductively coupled plasma-mass spectrometry (MCICP-MS), the Plasma 54, in Lyon [14]. The mean 176 Hf/177 Hf ratio determined for the JMC475 Hf standard was 0.28216 9 1 (2 S.D.) after mass fractionation correction by normalization to 179 Hf/177 Hf = 0.7325. External reproducibility for the 176 Lu/177 Hf ratio in Lyon is 6 1% and this estimate was assigned to the 176 Lu/177 Hf ratio of all presented analyses [14]. Pb isotopic analysis of the Mofen and Lan-

EPSL 6251 17-6-02


G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

maao phosphorites was performed by MC-ICPMS at the Danish Lithosphere Centre, whereas the Yongshaba phosphorites and the separated clay fractions were analyzed for Pb on the thermal ionization mass spectrometer (TIMS) at the University of Copenhagen. Correction for mass fractionation was done by use of doped Tl at the MC-ICP-MS and by repetitive analysis of the SRM-981 standard at the TIMS (values of [15]). Reproducibility of the 206 Pb/204 Pb ratio from replicate analyses of Tl-doped SRM-981 Pb standard was 150 ppm (n = 14; 2 S.D.) on the MC-ICP-MS and 450 ppm for the NBS 981 Pb standard (n = 8; 2 S.D.) on the TIMS. For the 176 Hf/177 Hf, 206 Pb/204 Pb and 207 Pb/204 Pb analysis, the external reproducibility of the standard was chosen to represent the errors in isochron calculations, except where samples were measured less precisely. In such cases, the internal standard errors (2 S.D.) observed for the individual samples were used instead. Procedural blanks were 6 200 pg for Hf, 6 3 pg for Lu and 6 75 pg for Pb.

3. Results 3.1. Phosphorite Lu^Hf and Pb isotopic data for the centrifuged apatite solutions are given in Table 1 and illustrated in isochron diagrams in Figs. 2 and 3. In general, the Lanmaao and Mofen samples from the Weng’an area are clearly distinguished from the Yongshaba samples at Kaiyang by their relatively low trace element concentrations, 176 Lu/ 177 Hf ratios and 208 Pb/204 Pb ratios as well as by their larger spread in 206 Pb/204 Pb and 207 Pb/204 Pb ratios (Table 1 and Figs. 2 and 3). Ages for the pro¢les at Weng’an overlap within uncertainty. Regression of the 176 Lu/177 Hf and 176 Hf/177 Hf data results in correlation lines with slopes de¢ning ‘ages’ of 602 9 43 Ma (MSWD = 2.0; Fig. 2a) for the Lanmaao pro¢le and 575 9 81 Ma (MSWD = 25; Fig. 2b) for the Mofen pro¢le. The Pb isotopic data for the Mofen pro¢le plot along an isochron of 599.3 9 4.2 Ma (MSWD = 2.9; Fig. 3a). For the Yongshaba pro¢le at Kaiyang, the Lu^

Hf data form a linear array corresponding to an ‘age’ of 584 9 26 Ma (MSWD = 6.0; Fig. 2c). In contrast, the Pb data for these samples are highly scattered and yield no meaningful age information, which is partly a re£ection of the low spread in 206 Pb/204 Pb and 207 Pb/204 Pb ratios observed for these samples (Fig. 3b). 3.2. Detrital clay Disturbance from clays on the Lu^Hf phosphorite ages is evaluated by comparing Lu^Hf analyses for the centrifuged and non-centrifuged ( = bulk) set of the Yongshaba samples. These are listed in Table 1 and compared in Fig. 4. Evaluation of the in£uence from detrital clays on the Pb^Pb ages for the phosphorites is done from the Pb analyses of clay separated from the Mofen and Yongshaba samples. The Pb isotopic data for the separated clay are listed in Table 2 and included in Fig. 3. Overall, this shows that the clays in the Doushantuo phosphorites are isotopically distinct from their hosts (Figs. 3 and 4). From comparison of the centrifuged and noncentrifuged set of the Yongshaba samples, it is seen that the presence of clay results in less radiogenic 176 Hf/177 Hf ratios, a lower initial 176 Hf/177 Hf ratio intercept as well as a signi¢cantly older apparent age (676 9 40 Ma versus 584 9 26 Ma; Fig. 4). This demonstrates that the clay in the Doushantuo phosphorites represents an older detrital component and that centrifuging the dissolved phosphorite solutions largely eliminated the in£uence from this phase on the Lu^Hf isotopic systematics. However, minor clay interference, possibly related to the relatively stronger dissolution acid required for Lu^Hf determination, cannot be entirely ruled out from Fig. 4. The clay in the Mofen samples has Pb isotope ratios, which would signi¢cantly disturb the wellde¢ned isochron observed for these phosphorites, had it not been removed prior to analysis (Fig. 3a). This is true even if only minor amounts of the clay a¡ected the phosphorite Pb budget. Based on this and the isochron produced by the data for centrifuged phosphorite solutions, we conclude that the Pb analyses of the phosphorites are una¡ected by the removed clay.

EPSL 6251 17-6-02

Table 1 Lu^Hf and Pb isotopic data for the Doushantuo samples Sample No. [Lu]a (ppm)


Lu/177 Hfa


Hf/177 Hfa


ObHf (600 Ma)

Agec Lu^Hf (Ma)

0.145 0.0709 0.0690 0.0332 0.0332

0.247 0.547 0.550 0.398 0.395

0.284901 9 43 0.288229 9 76 0.288441 9 95 0.286611 9 78 0.286611 9 78

310.4 9 1.5 312.8 9 2.7 36.6 9 3.4 310.4 9 2.8 39.1 9 2.8

602 9 43 MSWD = 2.0

0.0824 0.106 0.0358 0.0342 0.160

0.149 0.136 0.465 0.497 0.199

0.284574 9 50 0.284279 9 27 0.287710 9 39 0.288438 9 74 0.285003 9 25

17.3 9 1.8 12.3 9 1.0 1.7 9 1.4 14.9 9 2.6 12.4 9 0.9

575 9 81 MSWD = 25

0.233 0.239 0.314 0.180 0.277 0.264 0.4193 0.4232 0.3333 0.2851 0.2733 0.2765 0.3905 0.3484

1.357 1.005 0.561 0.965 0.595 0.584 0.772 0.755 0.537 0.414 0.665 0.656 0.396 0.454

0.297717 9 19 0.293531 9 39 0.288760 9 17 0.293176 9 28 0.289264 9 31 0.289139 9 22 0.291943 9 12 0.291799 9 20 0.288940 9 18 0.287260 9 26 0.290667 9 13 0.290517 9 15 0.287401 9 11 0.287768 9 28

31.6 9 0.7 38.4 9 1.4 0.6 9 0.6 35.0 9 1.0 4.5 9 1.1 4.7 9 0.8



Pb/204 Pba


Pb/204 Pba



Pb/204 Pba


1.82 2.23 1.43 1.69

38.9470 9 60 38.8262 9 65 127.483 9 20 129.206 9 22

16.8702 9 25 16.8632 9 47 22.1678 9 32 22.2788 9 62

0.9617 0.9375 0.9724 0.9739

37.9860 9 90 37.979 9 13 38.3114 9 90 38.127 9 13


68.078 9 11

18.6131 9 27


37.7111 9 89

584 9 26 2.72 MSWD = 6.0 3.59 3.88 3.68 3.53

40.412 9 98 41.079 9 55 40.545 9 76 38.628 9 65 38.808 9 54

16.874 9 42 16.965 9 24 16.864 9 33 16.826 9 29 16.804 9 25

0.989 0.980 0.985 0.987 0.981

40.31 9 10 39.723 9 58 39.127 9 77 39.223 9 70 38.961 9 59

Agec Pb^Pb (Ma)

599.3 9 4.2 MSWD = 2.9

676 9 40 MSWD = 21

G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

EPSL 6251 17-6-02

Lanmaao pro¢le LMA.1 0.253 LMA.3 0.273 LMA.3 0.268 LMA.4 0.0933 LMA.4 0.0925 Mofen pro¢le MF.1 0.0866 MF.1 0.101 MF.2 0.117 MF.2 0.120 MF.4 0.225 MF.5 Yongshaba pro¢le YSB.11 2.223 YSB.12 1.689 YSB.13 1.240 YSB.14.5 1.222 YSB.15 1.160 YSB.15 1.087 YSB.11b 2.2819 YSB.11b 2.2510 YSB.13b 1.2615 YSB.14b 0.8319 YSB.14.5b 1.2814 YSB.14.5b 1.2783 YSB.15b 1.0896 YSB.15b 1.1141


YSB.b = YSB.bulk analysis. Errors are 1% error for 176 Lu/177 Hf ratios and Lu concentration and 2c errors for 176 Hf/177 Hf, 206 Pb/204 Pb and 207 Pb/204 Pb ratios. b Calculated from 176 Hf/177 HfCHURð0Þ = 0.282772 and 176 Lu/177 HfCHURð0Þ = 0.0332 at 600 Ma. c Calculated from V 176 Lu = 1.876U10311 yr31 [36,37], V 238 U = 1.55125U10310 yr31 and V 235 U = 9.8485U10310 yr31 . Errors on the ages were estimated on the basis of 1% error for 176 Lu/177 Hf ratios and 2c errors for 176 Hf/177 Hf, 206 Pb/204 Pb and 207 Pb/204 Pb ratios. d 206 Pb/204 Pb vs. 207 Pb/204 Pb error correlation (r) calculated according to [38]. a



G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

isochron sensu stricto is de¢ned by a suite of samples with identical initial isotopic composition, which lie on a present-day correlation line with a MSWD around 2.5 or less [16]. According to this, only the Lu^Hf age of 602 9 43 Ma and the Pb^Pb age of 599.3 9 4.2 Ma obtained for the Doushantuo phosphorites in the Weng’an area represent true isochrons (Figs. 2a and 3a). The smaller error for the Pb^Pb age relative to the Lu^Hf isochron age is related to the signi¢cantly higher radiogenic ingrowth in Pb over time. The minor scatter observed for the two Lu^Hf errorchrons (Fig. 2b,c) most likely derives from either initial isotopic heterogeneity, an inherited component, or post-depositional diagenetic disturbance. The initial isotopic ratios of phosphorites could

Fig. 2. Lu^Hf ‘isochrons’ for the Lanmaao (circles), Mofen (squares) and Yongshaba (diamonds) pro¢les of the Doushantuo Formation. Analytical errors (2c) are smaller than symbols. Data regression was done using ISOPLOT after the method of Ludwig [38].

4. Discussion It is striking that despite the physical distance and the di¡erences in occurrence and geochemistry of the phosphorites at Weng’an and Kaiyang, all obtained ages overlap within uncertainty of the most precise Pb^Pb age of 599.3 9 4.2 Ma. An

Fig. 3. Pb^Pb ‘isochrons’ for the Mofen (squares) and Yongshaba (diamonds) phosphorites. Analytical errors are smaller than symbols. Also presented as error ellipses are Pb^Pb analyses of clay fraction removed from the phosphorites samples. Arrows from clay analyses point towards the phosphorite samples, from which the individual clay fractions were separated. Data regression and estimation of 207 Pb/ 206 Pb age uncertainties (2c) were calculated using the ISOPLOT method of Ludwig [38].

EPSL 6251 17-6-02

G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

lation are thus indicated to have changed markedly during the time of Doushantuo phosphate deposition [19] (Table 1). This is supported by the deposition of these phosphorites in a shelf environment, which was likely characterized by mixing of upwelling deep water with water masses a¡ected by severe post-glacial weathering of the adjacent continent [13]. In such an environment, it is not surprising that the phosphorites show geochemical and isotopic variations on a regional (between pro¢les) as well as on a local scale (within one pro¢le) (e.g., [20]). Local initial Hf isotopic heterogeneity may thus be the reason for the minor scatter on the Lanmaao and Yongshaba correlation lines. By contrast, the relatively large scatter on the Mofen errorchron, the high OHf (600 Ma) value of +17 for sample MF-1 and the di¡erence of more than 10 OHf (600 Ma) units for the two phosphorite fractions from the stratigraphical level MF-2 cannot be solely accounted for by initial isotopic heterogeneity (Table 1). If these features are related to post-depositional disturbance, this disturbance must have occurred shortly after phosphorite deposition in order to retain a Lu^Hf age close to 600 Ma. However, the lack of similar scatter for the Pb systematics in these samples (Fig. 3a) as well as the preservation here of delicate biological features in fossils suggest the absence of any diagenetic perturbation. We therefore suspect that minor clay leaching during sample digestion contributed to the observed scatter in Lu^Hf systematics for the Mofen samples. In contrast to the absence of diagenetic perturbation for Doushantuo phosphorites at Weng’an, the Pb systematics for the Yongshaba samples at Kaiyang strongly suggests that Pb mobilization occurred here (Fig. 3b). By analogy with the maximum variations in 207 Pb/204 Pb ratios observed in

Fig. 4. Comparison between the Lu^Hf data for centrifuged (open diamonds; Table 1) and non-centrifuged (¢lled diamonds) subset of samples from the Yongshaba pro¢le.

re£ect the local isotopic variation of the seawater masses in which the phosphates formed. At present, the Hf isotopic heterogeneity observed in ocean waters is largely controlled by mixing between a mantle component and a continental ‘zircon-free’ component [17]. The Hf isotopic characteristics of these two components at 600 Ma are best represented by the composition of the depleted mantle, sampled at oceanic spreading ridges, and by Archean detrital sediments. Vervoort and Blichert-Toft [18] predict an OHf (600 Ma) value for depleted mantle of maximum +16, whereas Archean shales reported by Vervoort et al. [10] had OHf values between 330 and 340 at 600 Ma. Initial OHf values were calculated for the Doushantuo phosphorites assuming a depositional age of 600 Ma (Table 1). These show that the positive OHf (600 Ma) values of the Mofen section indicate a much stronger signal from the mantle-derived Hf component than the intermediate and the mostly negative OHf (600 Ma) values observed for the Yongshaba and Lanmaao samples, respectively. The patterns of water circu-

Table 2 Pb isotopic data for clay in the Mofen and Yongshaba phosphorites 206

MF-1.clay MF-2.clay MF-5.clay YSB-12.clay YSB-14.5.clay

Pb/204 Pb

35.6 9 12 88.5 9 14 64.38 9 54 39.67 9 15 34.56 9 29



Pb/204 Pb

16.18 9 27 19.325 9 73 18.278 9 50 16.686 9 28 16.807 9 70

EPSL 6251 17-6-02


Pb/204 Pb

35.2 9 11 36.72 9 27 37.16 9 23 38.68 9 14 37.78 9 35


G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

modern seawater [21], the scatter in Pb for the Yongshaba phosphorites is too large to be accounted for solely by initial isotopic heterogeneity of the ambient seawater. Instead, the lack of wellpreserved fossils and the observation of sul¢des, mostly pyrite, indicate that these rocks experienced Pb mobility associated with post-depositional sul¢de formation. This would explain the lack of similar scatter in the Lu^Hf data (Fig. 2c), since these elements, having only one state of oxidation, are not expected to form sul¢de complexes in solution. The above discussion shows that there is a striking absence of correlation between Lu^Hf age precision and the degree of diagenetic disturbance suggested by Pb systematics and fossil preservation within the same phosphorite units. It is therefore the conclusion of this study that the Lu^ Hf isotopic system in phosphorites is robust against diagenetic alteration. Instead, the observed scatter for this system is mainly related to initial isotopic heterogeneity of Hf, preserved due to limited (relative to Pb in the Mofen samples) radiogenic ingrowth. Additionally, leaching of clay by the stronger acid needed to dissolve the larger phosphorite fractions for Hf determination must have contributed to the observed Lu^Hf scatter for the Mofen samples. We can only speculate whether this is due to the type of clay present here. Prior to this study, the best direct radiometric constraints on Doushantuo deposition were U^Pb zircon dates of 748 9 12 Ma from tu¡s in red beds that underlie the Nantuo glacial rocks [22] and 539 9 34 Ma from a bentonite bed within the overlying Lower Cambrian succession [23]. This provides a ca. 200 Myr window for the deposition of the Doushantuo phosphorites. Previous attempts to directly date the Doushantuo rocks using the Rb^Sr and K^Ar isotopic systems have yielded ages ranging from 600 to 700 Ma (see [24] for review). A serious concern regarding these Rb^Sr and K^Ar determinations is the in£uence of inherited clay minerals in the analyses [24]. In accordance to this, our study con¢rms that the clay in the Doushantuo Formation is an older detrital component, which signi¢cantly a¡ects isotopic dates if it is not carefully removed.

Based on the above discussion, we conclude that the Pb^Pb age of 599.3 9 4.2 Ma represents the most precise depositional age existing for the Doushantuo phosphorites. A depositional age close to 600 Ma indicates that the fossil assemblage in the Doushantuo Formation predates all known assemblages of diverse Ediacaran animals [25], and is the earliest paleontological record of animal evolution. Equally important are the implications of this age for the understanding of the dramatic climate changes that occurred during the Neoproterozoic. At present, there is no consensus on the scale, number or duration of Neoproterozoic ice ages. It is widely agreed that at least two glaciations, an older, Sturtian, glaciation at about 700 Ma and a younger, Marinoan, glaciation at about 600 Ma, reached truly global proportions. Whether these were the sole Neoproterozoic ice ages [26] or the most prominent of at least four glaciations [27] can only be resolved by radiometric constraints from glaciogenic rocks. At Yangtze Gorges, the ‘typical’ post-Marinoan N13 C pro¢le is observed (e.g., [2,28]). Negative N13 C values just above the Nantuo Formation are followed by two (perhaps three) positive excursions during the Doushantuo and overlying Dengying successions and ¢nally succeeded by relatively constant values around +2x (e.g., [29,30]). Thus, the carbon isotopic composition of the Doushantuo sediments suggest that these are post-Marinoan, which is also in agreement with available Sr isotopic data [30] and biostratigraphy for Doushantuo rocks [31]. This has led to the conventional correlation of the Nantuo tillite with the Marinoan ice age (e.g., [24]). As the Nantuo tillite is bracketed by unconformities, it could, in principle, have been deposited during the earlier Sturtian glaciation. However, by providing minimum ages for tillite deposition, the radiometric ages for the Doushantuo Formation still allow correlations of the Nantuo rocks with glaciogenic successions from other regions. Only two glaciogenic deposits believed to be Marinoan have been constrained by U^Pb radiometric ages. On the Avalon Peninsula, Newfoundland, glacial deposits are bracketed by ash beds yielding U^Pb zircon ages of 606 9 3 Ma and

EPSL 6251 17-6-02

G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

565 9 3 Ma [32,33]. To the south, in Massachusetts, U^Pb zircon ages of detrital zircons from sandstone underlying the possibly correlative Squantum tillite provide a maximum age of 595 Ma for the glaciation [34]. The Pb^Pb date of 599.3 9 4.2 Ma for the Doushantuo phosphorites in the Mofen pro¢le provides a minimum age for the Marinoan glaciation in China. Comparison of this age, which was obtained at a relatively high stratigraphic layer in the Doushantuo formation, with the 595 Ma age constraint from below the Squantum tillite might suggest that the Squantum tillite post-dates the Nantuo glaciation. In Norway, glaciogenic rocks of the Moelv tillite lie above conglomerates that contain phosphatic clasts with complex acritarchs like those found in the Doushantuo rocks. This observation therefore provides biostratigraphic support for the hypothesis of a regionally extensive post-Marinoan glaciation, represented by the Squantum and the Moelv glaciogenic deposits [2,35]. If, however, the Squantum and Nantuo tillites are correlative, the ice age is indicated to have been of short duration. Further chronological investigation of these successions in China, North America and Norway is required to fully substantiate this hypothesis.

5. Conclusion The mutual consistency of indirect age constraints, Lu^Hf dates, and a Pb^Pb age for the Doushantuo phosphorites demonstrates that the Lu^Hf and Pb isotopic systems can preserve near-depositional ages in old sedimentary phosphates. The samples from the Mofen pro¢le show that Pb^Pb dating of phosphorites may yield very precise ages. However, the disturbance of Pb systematics in the Yongshaba samples, where a near-depositional Lu^Hf age is preserved, suggests that the U^Pb isotopic system is prone to diagenetic perturbation. The Lu^Hf isotopic system therefore represents a potentially robust system for radiometric dating of sedimentary rocks, albeit with larger uncertainties relative to Pb^Pb dating. We believe that a combination of Pb^Pb and Lu^Hf dating of sul¢de-free samples containing well-preserved fossils o¡ers the best opportu-


nity for direct dating of Proterozoic phosphatic deposits. Even at their current level of resolution, Lu^Hf and Pb^Pb dates could, however, provide useful age estimates for many Proterozoic strata that are poorly constrained from indirect dating. Further work is required to determine approaches for sample selection and preparation that will yield the most robust age determinations from our method. Success in this e¡ort may then help resolve age relationships among Neoproterozoic sedimentary successions, contributing to a better understanding of the relationship between tectonics, climate, and biological evolution during the Neoproterozoic Era.

Acknowledgements This research was ¢nanced by the Danish Lithosphere Centre and Ecole Normale Supe¤rieure de Lyon. We are grateful to J. BlichertToft for her preliminary work with the phosphate chemistry and thank D. Sumner, I. Montanez, C.E. Lesher and S. Bowring for discussions and comments. A.H.K. acknowledges the NASA Astrobiology Institute and S.X. the Louisiana Board of Regents Support Fund (541344) and the Chinese Ministry of Science and Technology (G2000077700). This paper bene¢ted from careful reviews by Ariel D. Anbar, Louis A. Derry and an anonymous reviewer.[BARD] References [1] S.A. Bowring, D.H. Erwin, A new look at evolutionary rates in deep time; uniting paleontology and high-precision geochronology, GSA Today 8 (1998) 1^8. [2] A.H. Knoll, Learning to tell Neoproterozoic time, Precambrian Res. 100 (2000) 3^20. [3] S. Xiao, Y. Zhang, A.H. Knoll, Three-dimensional preservation of algae and animal embryos in a Neoproterozoic phosphorite, Nature 391 (1998) 553^558. [4] S. Xiao, A.H. Knoll, Phosphatized animal embryos from the Neoproterozoic Doushantuo Formation at Weng’an, Guizhou, South China, J. Paleontol. 74 (2000) 767^788. [5] C.W. Li, J.Y. Chen, T.E. Hua, Precambrian sponges with cellular structures, Science 279 (1998) 879^882. [6] S. Xiao, X. Yuan, A.H. Knoll, Eumetazoan fossils in terminal Proterozoic phosphorites?, Proc. Natl. Acad. Sci. USA 97 (2000) 13684^13689.

EPSL 6251 17-6-02


G.H. Barfod et al. / Earth and Planetary Science Letters 201 (2002) 203^212

[7] P.F. Ho¡man, A.J. Kaufman, G.P. Halverson, D.P. Schrag, A Neoproterozoic Snowball Earth, Science 281 (1998) 1342^1346. [8] W.T. Hyde, T.J. Crowley, S.K. Baum, W.R. Peltier, Neoproterozoic ‘snowball Earth’ simulations with a coupled climate/ice-sheet model, Nature 405 (2000) 425^429. [9] B. Jahn, H. Cuvellier, Pb-Pb and U-Pb geochronology of carbonate rocks: an assessment, Chem. Geol. 115 (1994) 125^151. [10] J.D. Vervoort, P.J. Patchett, J. Blichert-Toft, F. Albare'de, Relationships between Lu-Hf and Sm-Nd isotopic systems in the global sedimentary system, Earth Planet. Sci. Lett. 168 (1999) 79^99. [11] Y. Li, Regional review, in: P.J. Cook, J.H. Shergold (Eds.), Proterozoic and Cambrian Phosphorites, Cambridge University Press, Cambridge, 1986, pp. 42^62. [12] D.E.G. Briggs, A.J. Kear, Fossilization of soft tissue in the laboratory, Science 259 (1993) 1439^1442. [13] S. Xiao, A.H. Knoll, Fossil preservation in the Neoproterozoic Doushantuo phosphorite Lagersta«tte, South China, Lethaia 32 (1999) 219^240. [14] J. Blichert-Toft, C. Chauvel, F. Albare'de, Separation of Hf and Lu for high-precision isotope analysis of rock samples by magnetic sector-multiple collector ICP-MS, Contrib. Miner. Petrol. 127 (1997) 248^260. [15] W. Todt, R.A. Cli¡, A. Hanser, A.W. Hofmann, Recalibration of NBS lead standards using a 202 Pb+205 Pb double spike, Terra Abstr. 5 (1993) 396. [16] C. Brooks, S.R. Hart, T. Wendt, Realistic use of twoerror regression treatments as applied to rubidium-strontium data, Rev. Geophys. Space Phys. 10 (1972) 551^ 577. [17] F. Albare'de, A. Simonetti, J.D. Vervoort, J. Blichert-Toft, W. Abouchami, A Hf-Nd isotopic correlation in ferromanganese nodules, Geophys. Res. Lett. 25 (1998) 3895^3898. [18] J.D. Vervoort, J. Blichert-Toft, Evolution of the depleted mantle: Hf isotope evidence from juvenile rocks through time, Geochim. Cosmochim. Acta 63 (1999) 533^556. [19] K. David, M. Frank, R.K. O’Nions, N.S. Belshaw, J.W. Arden, The Hf isotope composition of global seawater and the evolution of Hf isotopes in the deep Paci¢c Ocean from Fe-Mn crusts, Chem. Geol. 178 (2001) 23^42. [20] C. Holmden, R.A. Creaser, K. Muehlenbachs, S.A. Leslie, S.M. Bergstrom, Isotopic evidence for geochemical decoupling between ancient epeiric seas and bordering oceans; implications for secular curves, Geology 26 (1998) 567^ 570. [21] W. Abouchami, S.L. Goldstein, A lead isotopic study of Circum-Antarctic manganese nodules, Geochim. Cosmochim. Acta 59 (1995) 1809^1820. [22] Z. Zhao, Y. Xing, G. Ma, Y. Chen, Biostratigraphy of the Yangtze Gorges Area (1) Sinian, Geological Publishing House, Beijing, 1985, 143 pp. [23] W. Compston, I.S. Williams, J.L. Kirchvink, Z. Zhang,

[24] [25]











[36] [37] [38]

G. Ma, U-Pb ages for the Early Cambrian time-scale, J. Geol. Soc. London 149 (1992) 171^184. A.H. Knoll, S. Xiao, On the age of the Doushantuo Formation, Acta Micropalaeontol. Sin. 16 (1999) 225^236. M.W. Martin, D.V. Grazhdankin, S.A. Bowring, D.A.D. Evans, M.A. Fedonkin, J.L. Kirschvink, Age of Neoproterozoic bilaterian body and trace fossils, White Sea, Russia: implications for metazoan evolution, Science 288 (2000) 841^845. M.J. Kennedy, B. Runnegar, A.R. Prave, K.H. Ho¡mann, M.A. Arthur, Two or four Neoproterozoic glaciations?, Geology 26 (1998) 1059^1063. A.J. Kaufman, J.M. Hayes, A.H. Knoll, G.J.B. Germs, Isotopic compositions of carbonates and organic carbon from upper Proterozoic successions in Namibia: stratigraphic variation and the e¡ects of diagenesis and metamorphism, Precambrian Res. 49 (1991) 301^327. B.Z. Saylor, A.J. Kaufman, J.P. Grotzinger, F.A. Urban, A composite reference section for terminal Proterozoic strata of Southern Namibia, J. Sediment. Res. 68 (1998) 1223^1235. I.B. Lambert, M.R. Walter, W. Zang, S. Lu, G. Ma, Paleoenvironment and carbon isotope stratigraphy of Upper Proterozoic carbonates of the Yangtze Platform, Nature 325 (1987) 140^142. J. Yang, W. Sun, Z. Wang, Y. Xue, X. Tao, Variations in Sr and C isotopes and Ce anomalies in successions from China; evidence for the oxygenation of Neoproterozoic seawater?, Precambrian Res. 93 (1999) 215^233. Y. Zhang, Multicellular thallophytes with di¡erentiated tissues from late Proterozoic phosphate rocks of South China, Lethaia 22 (1989) 113^132. T.E. Krogh, D.F. Strong, S.J. O’Brien, V. Papezik, Precise U-Pb zircon dates from the Avalon Terrane in Newfoundland, Can. J. Earth Sci. 25 (1988) 442^453. A.P. Benus, Sedimentological context of a deep-water Ediacaran fauna (Mistaken Point Formation, Avalon Zone, eastern Newfoundland), In: E. Landing, G.M. Narbonne, P. Myrow (Eds.), Trace Fossils, Small Shelly Fossils and the Precambrian-Cambrian Boundary, Bulletin of the New York State Museum, 1988, pp. 8^9. M.D. Thompson, S.A. Bowring, Age of the Squantum ‘tillite’, Boston Basin, Massachusetts; U-Pb zircon constraints on terminal Neoproterozoic glaciation, Am. J. Sci. 300 (2000) 630^655. G. Vidal, Giant acanthomorph acritarchs from the upper Proterozoic in Southern Norway, Palaeontology 33 (1990) 287^298. Y. Nir-El, N. Lavi, Measurement of half-life of 176 Lu, Appl. Radiat. Isot. 49 (1998) 1653^1655. E. Scherer, C. Mu«nker, K. Mezger, Calibration of the Lutetium-Hafnium Clock, Science 293 (2001) 683^687. K.R. Ludwig, ISOPLOT; a plotting and regression program for radiogenic-isotope data; version 2.53, U.S.G.S. Open ¢le Rep. 91-0445 (1991).

EPSL 6251 17-6-02