UNCORRECTED PROOF

Middle Cambrian high magnetic reversal frequency(Kulumbe river section, northwestern Siberia) and reversal

behaviour during the Early Palaeozoic

Vladimir Pavlov a;b;*, Yves Gallet b

a Department of Geomagnetism and Paleomagnetism, Institute of Physics of the Earth, Moscow, Russiab Laboratoire de Paleomagnetisme, UMR7577, Institut de Physique du Globe de Paris, Paris, France

Received 30 May 2000; accepted 24 November 2000

Abstract

We present a magnetostratigraphic study of Middle Cambrian sediments along the Kulumbe river (northwesternSiberia). The deposits are V850 m thick and lie within the Mayan stage as constrained by trilobite fossils.Palaeomagnetic analyses reveal two magnetic components. The first component, parallel to the direction of the presentday field at the site, is isolated in the low to middle temperature range. A high temperature component (HTC) is thenisolated up to 580 or 680³C, carried either by magnetite and/or by hematite. The HTC defines a sequence of 28 magneticpolarity intervals, some of them being defined by only one sample. Considering that the studied section has a durationof V5 Myr, we propose that the magnetic reversal frequency was high (V4^6 reversals per Myr) during part of theMiddle Cambrian, among the highest values known within the Phanerozoic. The reversal frequency may have beenroughly similar during the Lower Cambrian. We further suggest a drastic decrease of the magnetic reversal frequencybetween the Lower^Middle Cambrian and the end of the Tremadoc (Ordovician) when a superchron probablyoccurred. However, this behaviour is still challenged by other scenarios which depend on the chosen Early Palaeozoictime scale and on the reliability of some magnetostratigraphic data. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Magnetostratigraphy; Magnetic reversal frequency; Early Paleozoic; Siberia

1. Introduction

The existence of magnetic polarity reversals isprobably as old as the generation of the geomag-netic ¢eld itself. Gubbins [1] recently speculatedthat geomagnetic reversals were much more fre-quent, approximately by a factor of 10, before the

growth of the solid inner core than after. Theinner core may indeed stabilise the magnetic ¢eldgenerated in the liquid outer core (e.g. [2,3]). Ageestimates for the inner core strongly depend onwhich cooling model is considered for the Earth,with one of the latest possibilities being youngerthan 1.7 billion years [4]. No magnetostrati-graphic record showing a succession of severalmagnetic reversals is presently known for periodsolder than 2 Ga. Between 2 and 1 Ga, magneto-stratigraphic results are very rare (e.g. [5,6]),partly because of the di¤culty of ¢nding non-re-

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 3 6 4 - 2

* Corresponding author. Tel. : +7-095-254-9105;Fax: +7-095-255-6040; E-mail: pavlov@uipe-ras.scgis.ru

EPSL 5715 18-12-00

Earth and Planetary Science Letters 5715 (2000) 1^11

www.elsevier.com/locate/epsl

UNCORRECTED PROOFmagnetised Precambrian sedimentary sections.

Gallet et al. [7] recently reported relatively numer-ous geomagnetic reversals in several sedimentarysections deposited V1050^1100 Myr ago fromthe northwestern and the southeastern parts ofthe Siberian platform. These results may indicatea moderate to high magnetic reversal frequencyduring a part of the late Mesoproterozoic, but afrequency value cannot be estimated because ofthe absence of precise dating constraints. The old-est time-constrained magnetic reversal frequencylies within the Lower Cambrian. There, Kirsch-vink and Rozanov [8] found many polarity rever-sals from biostratigraphically well-dated Tommo-tian, Atdabanian and Botomian sedimentarysections along the Lena river (Siberia). However,the interpretation of these results is controversialbecause the palaeomagnetic pole obtained byKirschvink and Rozanov [8] is signi¢cantly di¡er-ent to other Cambrian poles from the Siberiancraton and because a predominant reversed polar-ity is most often observed during this period (e.g.[9,10]). On the other hand, magnetostratigraphicdata from Morocco and China further supportthe occurrence of several, perhaps numerous,magnetic reversals during the Lower Cambrian[11]. A strong polarity bias is much more evidentduring the Ordovician (in the Arenig and theLlanvirn), which is interpreted by Algeo [12], Gal-let and Pavlov [13] and Pavlov and Gallet [14] asa superchron lasting approximately 15^20 Myr. Iftrue, it would suggest the persistence of a V150Myr-long time constant over the whole Phanero-zoic [13,15]. However, this hypothetical super-chron is in need of further con¢rmation and itmust be integrated in the general evolution ofthe magnetic reversal frequency during the EarlyPalaeozoic. For this reason, we returned to theKulumbe river area (northwestern Siberia) whereUpper Cambrian, Lower and Middle Ordovician(Tremadoc and Llanvirn stages, respectively)magnetostratigraphic results were already ob-tained by Pavlov and Gallet [14]. We present inthis study the magnetostratigraphy of the verythick Middle Cambrian part of the Kulumbe sec-tion.

2. Description of the Middle Cambrian Kulumberiver section

The studied section is located in the lowerstream valley of the Kulumbe river, about 100km to the east of the Yenisey river, in the north-west of the Siberian platform (Fig. 1; see also Fig.1 in Pavlov and Gallet [14]). Tectonically, thisarea lies on the margin of the Siberian craton,within the bounds of the Khantayka^Rybnayauplift of the Turukhansk^Norilsk dislocationzone (e.g. [16]). The Middle Cambrian depositscrop out along both sides of the Kulumbe riverover a distance of 6^7 km (latitude: 68.0³N, lon-gitude: 88.4³E). This sedimentary sequence isV850 m thick and consists of grey, greenishgrey and reddish limestones, argillaceous lime-stones and marlstones that dip gently (15^20³)to the east^southeast. The upper 630 m of thesection belongs to the Labaznaya Formation,while the lower part of the sequence belongs tothe upper part of the Ust^Brus Formation (e.g.

Fig. 1. Locality map of the studied Kulumbe river section.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^112

UNCORRECTED PROOF

[17^20]). Both formations have very similar lith-ologies, although reddish colours are more typicalof the Ust^Brus Formation. Magmatic intrusions(sills and dykes) are present at two restrictedstratigraphic intervals within the ¢rst half of theLabaznaya Formation.

The thoroughly studied biostratigraphic zona-tion of the Middle Cambrian section is essentiallybased on a rich fauna of trilobites (e.g.[17,18,20,21]). Two trilobite zones are distin-guished within the section. The upper Lejopygelaevigata/Aldanaspis truncata zone extends overthe top 340^350 m of the Labaznaya Formation,and is characterised by the following trilobite spe-cies: Maiaspis mirabilis N. Tchern, Bonneterrina

sachaica Ros., Nganasanella nganasanennsis Ros.,Kontrastina samodiica Ros., Buitella olenekensisLaz. and Bolaspidina insignis N. Tchern. The Mid-dle^Upper Cambrian boundary, which is markedby an abrupt change of trilobite assemblages, istraced to the top of the Labaznaya Formation,and at the base of the overlying Orakta Forma-tion [19,20]. The lower part of the LabaznayaFormation and the upper part of the Ust^BrusFormation exposed in the Kulumbe section liewithin the Anomocariodes limbaraeformis zone,characterised in particular by the trilobite speciesAgraulos acuminatus Ang., Phalacroma glandi-forme Ang., Aldanaspis venusta Laz., Oidalagnos-

Fig. 2. Orthogonal vector diagrams obtained by thermal and hybrid thermal-AF demagnetisation of samples from the Kulumberiver section. The closed (open) symbols refer to the horizontal (vertical) plane. All the diagrams are shown after bedding correc-tion.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^11 3

UNCORRECTED PROOFtus trispinifer Wgard., Anomocarina splendens

Lerm. and Anomocarioides sp.The two trilobite zones found in Kulumbe de-

¢ne the Mayan stage. We note that in Siberia, theMiddle Cambrian is composed, from bottom totop, of the Amgan and the Mayan stages. More-over, the Mayan stage is de¢ned by three trilobitezones in the Siberian biostratigraphic time scale,the oldest of which (the Anopolenus henrici/Cory-nexochus perforatus zone [22]) is not exposed inthe Kulumbe river valley. The Amgan stage isalso de¢ned by three trilobite zones [22]), anddespite its thickness, our section therefore con-tains only two of the six Middle Cambrian Sibe-rian trilobite zones.

3. Palaeomagnetic results

Approximately 450 oriented hand-samples werecollected in 1998 with a stratigraphic samplinginterval of 1^2.5 m. Their remanent magnetisa-tions were analysed in the palaeomagnetic labora-tory at Institut de Physique du Globe de Pariswith a CTF three-axis cryogenic magnetometerhoused in a shielded room.

Thermal demagnetisation, performed with alaboratory-built furnace, generally reveals twomagnetic components. The ¢rst (low temperaturecomponent, LTC) is often isolated over a widetemperature range from 130³C to about 500³C(Fig. 2). The LTC directions are parallel to thepresent day ¢eld direction at the site, indicatingthat this component has likely a recent viscous orchemical origin (Table 1; Fig. 3a). A high temper-ature component (HTC), which trends toward theorigin on the orthogonal vector diagrams, is iso-lated at higher temperatures, up to 580 or 680³C.However, in many samples, the demagnetisationpaths become scattered at high temperatures, pre-venting a clear estimate of the HTC directions(Fig. 2e). To avoid this e¡ect, we performed forsome samples hybrid thermal-alternating ¢eld(AF) demagnetisations. Our treatment indicatesthat the HTC component is carried by magnetiteor hematite (Fig. 2a,b,c,e,f,g and d,h, respec-tively). In some cases, AF demagnetisation re-moves only a part of the HTC component, which

may indicate that it is locally carried both bymagnetite and hematite. To characterise betterthe magnetic mineralogy, we have thermally de-magnetised a composite (¢elds of 1.2 and 0.3 Tapplied in two perpendicular directions [22]) iso-thermal remanent magnetisation (IRM). The ex-amples shown in Fig. 4 con¢rm that the magnet-isation is either carried by magnetite (Fig. 4a),hematite (Fig. 4b) or by a mixture of magnetiteand hematite (Fig. 4c).

In several samples collected close to magmaticintrusions, a third magnetic component is re-moved in the middle temperature range, betweenthe LTC and HTC components (Fig. 2a). In a fewcases, this component persists until the samplesare completely demagnetised. This componenthas a steep upward inclination and is directedtoward the northwest (Fig. 3b), which is a direc-tion similar to the one we previously found fromother intrusions located in the Upper Cambrianand Ordocician parts of the Kulumbe section [14].These intrusions are related to the emplacementof the Siberian traps, and we therefore considerthat this magnetic component was acquired at thePermo^Triassic boundary (Table 1; see also [13]).

The HTC component possesses dual magneticpolarities (Fig. 3c). Some directions are clearlybiased by the LTC component, and this is attestedby the fact that the mean normal and reversedpolarity directions are not antiparallel (Fig. 5,left ; Table 1 [23]). This bias likely results fromthe large overlap between the demagnetisationspectra of the LTC and HTC components, togeth-er with the scatter observed at high temperaturesin the demagnetisation paths. In Fig. 3d, we havetentatively selected only the HTC directions whichare obviously less contaminated by the LTC com-ponent. The mean normal and reversed directionsof this population are still not antipodal, but theangular di¡erence between these two directions ismuch less than before (Fig. 5, right; Table 1).This indicates that our selection is not fully sat-isfactory but we note that the two general meandirections computed before and after the selectionare almost identical (MEAN 1 and MEAN 2,Table 1). We therefore consider that the averagingof the normal and reversed polarity directions al-

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^114

UNCORRECTED PROOF

lows a good estimate of the geomagnetic directionprevailing at the time of the sediment deposition.

In the absence of a fold test, the arguments for

the early origin of the HTC component arise fromthe fact that (1) di¡erent magnetic minerals carrythe same HTC palaeomagnetic direction and (2)

Table 1Directions of the LTC and HTC obtained from the Middle Cambrian Kulumbe river section

N Geographic coordinates Stratigraphic coordinates

D I K K95 D I K K95

(³) (³) (³) (³) (³) (³)

Remagnetisation components :LTC component 426 342.7 78.5 25.9 1.4 81.0 81.0 24.1 1.4Trap remagnetisation 9 315.0 374.1 127.9 4.6 310.4 364.6 114.2 4.8HTC component :Normal polarity 53 143.6 319.7 19.1 4.6 146.9 335.6 20.3 4.5Reversed polarity 226 319.2 32.7 17.6 2.3 322.8 48.1 26.0 1.9MEAN 1 279 140.1 330.2 16.7 2.1 143.7 345.8 22.7 1.8HTC component after selection :Normal polarity 29 141.8 322.8 29.8 5.0 144.8 338.0 31.8 4.8Reversed polarity 99 319.5 31.1 21.9 3.1 321.9 45.4 30.8 2.6MEAN 2 128 140.0 329.2 22.4 2.7 142.6 343.7 29.8 2.3

Fig. 3. Equal-area projection of palaeomagnetic directions isolated within the Middle Cambrian part of the Kulumbe section.Closed (open) symbols refer to directions in the lower (upper) hemisphere. (a) Directions obtained before bedding correction forthe LTC. (b) Directions of reversed polarity observed in samples close to trap intrusions (before tilt correction). (c) Directions ofthe HTC isolated after bedding correction. (d) Selection of apparently unbiased directions of the HTC.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^11 5

UNCORRECTED PROOF

the HTC direction is similar (but statistically dis-tinct) to younger palaeomagnetic directions previ-ously obtained from the Kulumbe and theMoyero river sections [13,14]. The Middle Cam-brian mean direction is upward (downward) tothe southeast (northwest), and assuming that theSiberian craton was located in the southern hemi-sphere, it indicates that the present Siberian

northern shorelines were facing south at thistime (e.g. [13,14,24^26]). Another important char-acteristic, which strongly argues for the primaryorigin of the HTC component, is that the MiddleCambrian direction is in direct continuity with theMiddle Ordovician to Upper Cambrian ApparentPolar Wander Path of Siberia constructed fromthe Kulumbe and Moyero magnetostratigraphicresults (Fig. 6). Moreover, the new data fromthe Kulumbe section are in relatively good agree-ment with the Middle Cambrian (mainly of Am-gan age) palaeomagnetic pole obtained from theSiberian Olenek section (Fig. 6 [10,27]).

Fig. 7 shows the declination and inclination re-cords versus depth. The closed symbols indicatethe directions which, we believe, are biased by theLTC component. The LTC contamination whichhas a high inclination principally a¡ects the incli-nations of the HTC component, and the HTCdeclinations therefore de¢ne more distinctly themagnetic polarities. We obtain a sequence of 28magnetic polarity intervals, four of them beingde¢ned by a single sample (Fig. 7). Moreover,three magnetic intervals are only de¢ned bybiased directions (respectively by one, one andtwo samples). There are two breaks in the middlepart of the magnetostratigraphic sequence due toremagnetisation linked to the emplacement of thetraps. The upper part of the section exhibits pre-dominantly reversed polarity, whereas polarity re-versals are frequent in the lower half of the se-quence.

4. Discussion

The new section sampled along the Kulumberiver shows numerous geomagnetic polarity rever-sals. We note here that Khramov et al. [28] pre-viously reported the occurrence of several mag-netic polarity reversals from Middle Cambriandeposits exposed along the upper stream of theOlenek river. Together with other data from theliterature, these results can be used to constrainthe magnetic reversal behaviour during the Cam-brian. This is possible because some geochrono-logical data are presently available for this period,although the time frame for the di¡erent Cambri-

Fig. 4. Thermal demagnetisation of a two-axial di¡erentialIRM (0.3 and 1.2 T). The magnetisation is carried either bymagnetite (a), hematite (b) or both by magnetite and hema-tite (c).

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^116

UNCORRECTED PROOF

an subdivisions is still not yet established withoutuncertainty.

From studies dealing with the Cambrian timescale (e.g. [29^34] and references therein) it ap-pears that two main characteristics are quitewell-established: (1) the duration of the LowerCambrian (V25 Myr) is about one half of theduration of the entire Cambrian and (2) the du-ration of the Middle Cambrian is likely shortcompared to the Lower Cambrian.

The duration of the Middle Cambrian was con-sidered between V5 and V20 Myr. The Middle^Upper Cambrian boundary was recently proposedto lie between 498 Ma [33] and 505 Ma [31,32,35],while the age of the Lower^Middle Cambrianboundary ranges from 510 Ma ([36], consideredby Meert [34]) to 518 Ma ([31], used by Kirsch-vink et al. [37]). As discussed by Jago and Haines[35], most of these uncertainties depend on themeasured ages of the standards used for radio-metric datings. However, following Tucker andMcKerrow [31] and Jago and Haines [35], 505Ma is most probable for the Middle^Upper Cam-brian boundary. Age constraints for the base ofthe Middle Cambrian are much more uncertain.A duration as short as 5 Myr for the MiddleCambrian (with a Lower^Middle Cambrianboundary at 510 Ma) would be very brief withrespect to the number of trilobite zones which

occurred during this period (e.g. [35,38]). Forthis reason, we will consider here a duration ofV13 Myr for the entire Middle Cambrian, whichtakes into account the oldest (but still reasonable)age suggested for the Lower^Middle Cambrianboundary (518 Ma [31]). This duration is likelya maximum estimate and the derived magneticreversal frequency will thus be a minimum esti-mate.

There are no data to constrain the relative du-

Fig. 5. Mean palaeomagnetic directions obtained in Kulumbe. Same convention as in Fig. 3. Left: Mean directions of the di¡er-ent components obtained before bedding correction. Right: Mean normal and reversed polarity directions in stratigraphic coordi-nate system obtained for the HTC before (closed and open circles) and after (closed and open squares) a selection of the data.The respective general mean directions are indicated by triangles (MEAN 1 and MEAN 2; Table 1).

Fig. 6. Equal-area projection of the Lower Silurian to Mid-dle Cambrian. Palaeomagnetic poles obtained from theMoyero (squares; [13]), Kulumbe (circles; [14]) and Olenek(triangle; [10]) river sections.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^11 7

UNCORRECTED PROOF

rations of the Amgan and the Mayan stages de-¢ning the Middle Cambrian in Siberia. The onlypossibility is to do a very rough approximationwhich consists of assuming an equal durationfor the Middle Cambrian trilobite zones. This ap-proximation is far from satisfactory because it isuntested. However, it yields a ¢rst order estimateof V5 Myr for the duration of our section whichencompasses two of the six trilobite zones. In Ku-lumbe, 28 magnetic polarity reversals are ob-served, but 20 if only intervals de¢ned by severalsamples are retained. The latter values constrainthe magnetic reversal frequency to be betweenfour and six reversals per Myr during approxi-mately the upper one third of the Middle Cam-brian. This estimate can be tentatively tested byconsidering the correlation already proposed be-

tween the Australian and Siberian Middle Cam-brian biozonations. According to Shergold et al.[37] and Jago and Haines [35], the L. laevigataand A. limbataeformis zones from the Kulumbesection should correspond to a succession of seventrilobite zones from Australia (from the C. quasi-vespa to the G. narthorstic zones). Further assum-ing the average time span of V0.55 Myr for theAustralian Middle Cambrian (and Upper Cambri-an) trilobite zones proposed by Jago and Haines[35], the sampled Kulumbe section would have aduration of V4 Myr, roughly similar to the du-ration we previously estimated. In both cases, themagnetic reversal frequency that we derived fromthe Kulumbe data is among the highest observedduring the Phanerozoic, comparable to those pre-

Fig. 7. Magnetostratigraphy of the Middle Cambrian part of the Kulumbe river section. Magnetic intervals de¢ned by only onesample are indicated by half bars. Directions clearly biased by remagnetisation are shown by closed circles.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^118

UNCORRECTED PROOFvailing during the Miocene and the Middle^

Upper Jurassic.As mentioned above, the magnetostratigraphy

of the Lower Cambrian is presently obscured bythe problem of the reliability of the data obtainedby Kirschvink and Rozanov [8] from the Lenariver. If their results give a good picture of thereversal polarity sequence at the base of the Pa-laeozoic, then it is of interest to compare the re-versal behaviour during the Lower Cambrian withthat of the Middle Cambrian. The computation ofmagnetic reversal frequencies during the LowerCambrian strongly relies on the relative durationsof the Tommotian and Atdabanian stages. This isalso of major importance for understanding thebiological diversi¢cation within the Lower Cam-brian (e.g. [37,39]). There are several options ac-cording to published Cambrian time scales. Tuck-er and McKerrow [31] proposed that both stageswould have a similar duration of V5 Myr. Dataobtained by Kirschvink and Rozanov [8] wouldthen indicate high reversal frequencies of V4 re-versals per Myr for the Tommotian and V6^7reversals per Myr for the Atdabanian, very com-parable to the value that we propose during theMiddle Cambrian. However, zircon U^Pb agesobtained close to the base of the Tommotian[39] and to the Tommotian^Atdabanian boundary[40] suggest that the duration of the Tommotianwas perhaps longer by a factor of two. In thiscase, reversal frequencies would be notably di¡er-ent between the Tommotian, characterised by amoderate frequency, and the Atdabanian. Amore extreme possibility supported by Meert[34] is that the duration of the Tommotian isabout 15 Myr, while the Atdabanian would bemuch shorter (V2^3 Myr). These durations implyunusually rapid changes in magnetic reversal fre-quency, with a very high value (s 10 reversals perMyr) during the Atdabanian and a low value dur-ing the Tommotian (V1.5 reversal per Myr).Such a rapid behaviour has no equivalent duringthe last 160 Myr, so a conservative attitude is tofavour one of the two ¢rst options.

Computing Upper Cambrian and lowermostOrdovician (Tremadoc) magnetic reversal fre-quencies is complicated by uncertainties in corre-lating Siberian biostratigraphy with the standard

geologic time scale based on graptolite biozona-tion from British sections (e.g. [21,41]). This maybe partly the reason for the relatively poor corre-lation between magnetostratigraphic results fromKulumbe [14] and from the Dayancha and BlackMountain sections [42,43]. However, consideringa duration of V20 Myr between the Middle^Upper Cambrian boundary (505 Ma [31,35]) andthe Tremadoc^Arenig boundary (485 Ma [31]),the available magnetostratigraphic data takenglobally seem to indicate a moderate magneticreversal frequency, most probably lower than dur-ing the Middle Cambrian.

At this stage of our knowledge, we propose adrastic decrease in magnetic reversal frequencybetween the Lower (?)^Middle Cambrian andthe end of the Tremadoc stage when a superchronprobably occurred [12,13]. This behaviour wouldbe similar to the one prevailing during the LowerCretaceous, before the Cretaceous Long Normalsuperchron. However, we recognise that it is pres-ently challenged by other scenarios which dependon the chosen Cambrian time scale and on thereliability of some magnetostratigraphic data. Fi-nally, the new data from Kulumbe clearly do notsupport the idea that geomagnetic reversals wereless frequent during the Lower Palaeozoic thanduring the last 300 Myr, as is suggested by Figure1 of Eide and Torsvik [44]. The Palaeozoic mag-netostratigraphic database is probably too limitedto support such an inference.

Acknowledgements

We thank Stuart Gilder for stimulating discus-sions. We are also pleased to thank L. Tauxe andW. Lowrie for useful comments on the manu-script. Field work was ¢nanced by the RussianBasic Research Foundation grant No. 98-05-65082. IPGP supported V.P. during his stay inParis. This is IPGP contribution No. 1721 andINSU contribution No. 252.[AC]

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^11 9

UNCORRECTED PROOFReferences

[1] D. Gubbins, The distinction between geomagnetic excur-sions and reversals, Geophys. J. Int. 137 (1999) F1^F3.

[2] R. Hollerbach, C. Jones, In£uence of the Earth's innercore on geomagnetic £uctuations and reversals, Nature365 (1993) 541^543.

[3] S. Gilder, J. Glen, Magnetic properties of hexagonalclosed-packed iron deduced from direct observations ina diamond anvil cell, Science 279 (1998) 72^74.

[4] S. Labrosse, J.-P. Poirier, J.-L. Le Moue«l, On cooling ofthe Earth's core, Phys. Earth Planet. Int. 99 (1997) 1^17.

[5] D. Bingham, M. Evans, Precambrian geomagnetic ¢eldreversal, Nature 253 (1975) 332^333.

[6] D. Elston, S. Bressler, Paleomagnetic poles and polarityzonation from the Middle Proterozoic Belt supergroup,Montana and Idaho, J. Geophys. Res. 85 (1980) 339^355.

[7] Y. Gallet, V. Pavlov, M. Semikhatov, P. Petrov, LateMesoproterozic magnetostratigraphic results from Sibe-ria: Paleogeographic implications and magnetic ¢eld be-haviour, J. Geophys. Res. 105 (2000) 16481^16500.

[8] J. Kirschvink, A. Rozanov, Magnetostratigraphy of Low-er Cambrian strata from the Siberian platform: a paleo-magnetic pole and a preliminary polarity time-scale, Geol.Mag. 121 (1984) 189^203.

[9] A. Khramov, V. Rodionov, The geomagnetic ¢eld duringthe Palaeozoic time, J. Geomagn. Geoelectr. 32 (sup. III)(1980) 99^115.

[10] S. Pisarevsky, E. Gurevich, A. Khramov, Paleomagnetismof the Lower Cambrian sediments from the Olenek riversection (northern Siberia): paleopoles and the problem ofthe magnetic polarity in the Early Cambrian, Geophys. J.Int. 130 (1997) 746^756.

[11] J. Kirschvink, M. Margaritz, R. Ripperdam, Y. Zhurav-lev, A. Rozanov, The Precambrian^Cambrian boundary:magnetostratigraphy and Carbon isotopes resolve correla-tion problems between Siberia, Morocco, and South Chi-na, GSA Today 1^4 (1991) 69^91.

[12] T. Algeo, Geomagnetic polarity bias patterns through thePhanerozoic, J. Geophys. Res. 101 (1996) 2785^2814.

[13] Y. Gallet, V. Pavlov, Magnetostratigraphy of the Moyeroriver section (north-western Siberia): constraints on geo-magnetic reversal frequency during the early Palaeozoic,Geophys. J. Int. 125 (1996) 95^105.

[14] V. Pavlov, Y. Gallet, Upper Cambrian to Middle Ordo-vician magnetostratigraphy from the Kulumbe river sec-tion (northwestern Siberia), Phys. Earth Planet. Int. 108(1998) 49^59.

[15] Y. Gallet, J. Besse, L. Krystyn, J. Marcoux, H. Theve-niaut, Magnetostratigraphy of the Bolu«cektasi Tepe sec-tion (southwestern Turkey): implications for changes inmagnetic reversal frequency, Phys. Earth Planet. Int. 73(1992) 85^108.

[16] N. Malich, Tectonic Development of the SedimentaryCover of the Siberian Platform, Nedra, Leningrad, 1975,p. 216 (in Russian).

[17] A. Rozova, Biostratigraphy and Description of the Trilo-

bites of Middle and Upper Cambrian of the Siberian Plat-form, Nauka, Moscow, 1964, p. 146 (in Russian).

[18] V. Datsenko, I. Zhuravleva, N. Lazarenko, Y. Popov, N.Chernychova, Biostratigraphy and Fauna of CambrianDeposits of the Northwest of the Siberian Platform, Ne-dra, Leningrad, 1968, p. 242 (in Russian).

[19] V. Astahkin, T. Pegel, Y. Shabanov, S. Sukhov, V. Sun-dukov, L. Repina, A. Rozanov, A. Zhuravlev, The Cam-brian system on the Siberian Platform, in: J. Shergold, A.Rozanov, A. Palmer (Eds.), International Union of Geo-logical Sciences, vol. 27, 1991, p. 134 (in English).

[20] A. Rozanov, L. Repina, M. Apollonov, Cambrian of Si-beria, Trans. of Geological and Geophysical In-te 788,Nauka, Novosibirsk, 1992, p. 136 (in Russian).

[21] A. Rozova, Correlation of upper Cambrian sections ofSouth Kazakstan and Siberian Platform, In: Biostratigra-phy and Paleontology of Cambrian of Northern Asia,Moscow, 1986, pp. 25^39 (in Russian).

[22] W. Lowrie, Identi¢cation of ferromagnetic minerals in arock by coercivity and unblocking temperature properties,Geophys. Res. Lett. 17 (1990) 159^162.

[23] P. McFadden, M. McElhinny, Classi¢cation of the rever-sal test in palaeomagnetism, Geophys. J. Int. 103 (1990)725^729.

[24] A. Khramov, Standard nets of paleomagnetic poles forplates of North Eurasia: the connection with the problemof geodynamics of the territory of USSR (in russian), In:Paleomagnetism and Paleogeodynamics of the territory ofUSSR, VNIGRI, Leningrad, 1991, pp. 154^176.

[25] R. Van der Voo, Paleomagnetism of the Atlantic, Tethysand Iapetus Oceans, Cambridge Univerity Press, Cam-bridge, 1993, p. 411.

[26] M. Smethurst, A. Khramov, T. Torsvik, The Neoproter-ozoic and Palaeomagnetic data for the Siberian platform:from Rodinia to Pangea, Earth Sci. Rev. 43 (1998) 1^24.

[27] E.P. Osipova, Paleomagnetic directions and palaeomag-netic poles, in: A. Khramov (Ed.), Reference Book forthe USSR, Moscow, 1973, p. 88 (in Russian).

[28] A. Khramov, R. Komissarova, G. Goncharov, in: A.Khramov (Ed.), Paleomagnetism of Paleozoic, Nedra,Leningrad, 1974, p. 236 (in Russian).

[29] S. Odin, Geological time scale, C.R. Acad. Sci. Paris 318(1994) 59^71.

[30] J. Shergold, Timescales calibration and development 1:Cambrian, Aust. Geol. Surv. Org. Rec. 30 (1995) 1^32.

[31] R. Tucker, W. McKerrow, Early Paleozoic chronology: areview in light of new U^Pb zircon ages from Newfound-land and Britain, Can. J. Earth Sci. 32 (1995) 368^379.

[32] F. Gradstein, J. Ogg, Phanerozoic time scale, Episodes 19(1996) 3^5.

[33] G. Young, J. Laurie, An Australian Phanerozoic Time-scale, Oxford University Press, Melbourne, 1996, p. 279.

[34] J. Meert, A paleomagnetic analysis of Cambrian true po-lar wander, Earth Planet. Sci. Lett. 168 (1999) 131^144.

[35] J. Jago, P. Haines, Recent radiometric dating of someCambrian in southern Australia: relevance to the Cam-brian time scale, Rev. Espan. Paleont. (1998) 115^122.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^1110

UNCORRECTED PROOF[36] E. Landing, S. Bowring, K. Davidek, S. Westrop, G.

Geyer, W. Heldmaier, Duration of the Early Cambrian:U^Pb date from Avalonian Wales and the age of theCambrian^Ordovician boundary, Can. J. Earth Sci. 35(1998) 303^309.

[37] J. Kirschvink, R. Ripperdam, D. Evans, Evidence for alarge-scale reorganization of early Cambrian continentallandmasses by inertial interchange true polar wander, Sci-ence 277 (1997) 541^545.

[38] J. Shergold, J. Laurie, S. Xiao Wen, Classi¢cation andreview of the trilobite order Agnostida Salter, 1864: anAustralian perspective, Geol. Geophys. Rep. No. 296, Bu-reau of Mineral Resources, Australia, 1990, pp. 1^93.

[39] S. Bowring, J. Grotzinger, C. Isachsen, A. Knoll, S. Pe-lechaty, P. Kolosov, Calibrating rates of Early Cambrianevolution, Science 261 (1993) 1293^1298.

[40] W. Compston, I. Willimas, J. Kirschvink, Zichao Zhang,Guogan Ma, Zircon U^Pb ages for the Early Cambriantime-scale, J. Geol. Soc. London 149 (1993) 171^184.

[41] A. Kanygin, T. Moscalenko, A. Yadrenkina, OrdovicianSystem of the Siberian Platform, Nauka, Novosibirsk,1987 (in English).

[42] R. Ripperdam, J. Kirschvink, Paleomagnetic results fromthe Cambrian-Ordovician boundary section at BlackMountain, Georgina basin, western Queensland, Austra-lia, in: Webey Laurie (Ed.), Global Perspectives on Ordo-vician Geology, Balkema, Rotterdam, 1992, pp. 93^103.

[43] R. Ripperdam, M. Margaritz, J. Kirschvink, Carbon iso-tope and magnetic polarity evidence for non-depositionalevents within the Cambrian^Ordovician boundary sectionnear Dayangcha Jilin Province, China, Geol. Mag. 130(1993) 443^452.

[44] E. Eide, T. Torsvik, Paleozoic supercontinental assembly,mantle £ushing, and genesis of the Kiaman superchron,Earth Planet. Sci. Lett. 144 (1996) 389^402.

EPSL 5715 18-12-00

V. Pavlov, Y. Gallet / Earth and Planetary Science Letters 5715 (2000) 1^11 11