Meteoritics & Planetary Science 40, Nr 4, 591–607 (2005) Abstract available online at http://meteoritics.org

Laser argon dating of melt from the , : Implications for a possible relationship to Late extinction events

Wolf U. REIMOLD1*, Simon P. KELLEY2, Sarah C. SHERLOCK2, Herbert HENKEL3, and Christian KOEBERL4

1Impact Cratering Research Group, School of Geosciences, University of the Witwatersrand, Private Bag 3, P. O. Wits 2050, Johannesburg, South Africa 2Department of Earth Sciences, Open University, Walton Hall, Milton Keynes MK7 6AA, UK 3Department of Land and Water Resources Engineering, Division of Engineering Geology and Geophysics, Royal Institute of Technology, Teknikringen 72, SE 100-44 Stockholm, Sweden 4Department of Geological Sciences, University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria *Corresponding author. E-mail: [email protected] (Received 12 July 2004; revision accepted 08 February 2005)

Abstract–In earlier studies, the 65–75 km diameter Siljan impact structure in Sweden has been linked to the Late Devonian mass extinction event. The Siljan has previously been dated by K- Ar and Ar-Ar chronology at 342–368 Ma, with the commonly quoted age being 362.7 ± 2.2 Ma (2 σ, recalculated using currently accepted decay constants). Until recently, the accepted age for the /Famennian boundary and associated extinction event was 364 Ma, which is within error limits of this earlier Siljan age. Here we report new Ar-Ar ages extracted by laser spot and laser step heating techniques for several melt samples from Siljan (interpreted to be impact melt breccia). The analytical results show some scatter, which is greater in samples with more extensive alteration; these samples generally yield younger ages. The two samples with the least alteration yield the most reproducible weighted mean ages: one yielded a laser spot age of 377.2 ± 2.5 Ma (95% confidence limits) and the other yielded both a laser spot age of 376.1 ± 2.8 Ma (95% confidence limits) and a laser stepped heating plateau age over 70.6% 39Ar release of 377.5 ± 2.4 Ma (2 σ). Our conservative estimate for the age of Siljan is 377 ± 2 Ma (95% confidence limits), which is significantly different from both the previously accepted age for the Frasnian/Famennian (F/F) boundary and the previously quoted age of Siljan. However, the age of the F/F boundary has recently been revised to 374.5 ± 2.6 Ma by the International Commission for Stratigraphy, which is, within error, the same as our new age. However, the currently available age data are not proof that there was a connection between the Siljan impact event and the F/F boundary extinction. This new result highlights the dual problems of dating impacts where fine-grained melt rocks are often all that can be isotopically dated, and constraining the absolute age of biostratigraphic boundaries, which can only be constrained by age extrapolation. Further work is required to develop and improve the terrestrial impact age record and test whether or not the terrestrial impact flux increased significantly at certain times, perhaps resulting in major extinction events in Earth’s biostratigraphic record.

INTRODUCTION known impact structure in Europe. Its diameter was originally estimated at 52 km (Grieve 1988), but Von Dalwigk and The Siljan impact structure (Fredriksson and Wickman Kenkmann (1999) and Kenkmann and Von Dalwigk (2000) 1963; Wickman et al. 1963; Svensson 1971, 1973; Rondot made a case for a larger diameter of at least 65 km on the basis 1975; BodÈn and Eriksson 1988; Juhlin and Pedersen 1987; of structural geological considerations and by applying the Kenkmann and von Dalwigk 2000; Henkel and Aaro 2005) is empirical morphometric scaling laws provided by Therriault located in the region of south-central Sweden, et al. (1997). In contrast, Henkel and Aaro (2005) observe a centered at 61°02′N/14°52′E (Fig. 1). Siljan is the largest 75 km wide current topographic expression and estimate that

591 © The Meteoritical Society, 2005. Printed in USA. 592 W. U. Reimold et al.

Fig. 1. The geology of the Siljan impact structure. The inset shows the location of Siljan in Scandinavia. the pre-erosional crater diameter could have been as much as Consequently, very little material that could be used to obtain 85 km. a reliable age for this impact structure has been produced. The evidence for impact at Siljan includes the presence Only “allochthonous breccia, small and narrow breccia dikes, of shatter cones and planar deformation features (PDFs) in and float of melt breccia” were reported by Rondot (1975). quartz. Exposure is poor in much of the Siljan structure, and Svensson (1971, 1973) suggested a post- age for the bona fide impact melt rock has, to date, remained elusive. impact event. Laser Ar dating of melt breccias from Siljan 593

In the 1980s, Siljan experienced an “impact exploration intriguing part of the Late Devonian, during which several boom” in the wake of Gold’s proposal (reviewed in Gold and important events occurred (geological time scale of Harland Soter 1980; Gold 1987, 1988) that the structure could provide et al. 1989). In this scheme, the Devonian was placed between access to significant mantle-derived resources 417 and 354 Ma, with the Givetian stage between 380 and that might have infiltrated into the impact-deformed basement 370 Ma, the Frasnian from 370–364 Ma, and the Famennian of the structure. Accordingly, Siljan was extensively and from 364 to 354 Ma. It should also be noted that Ellwood deeply drilled (BodÈn and Eriksson 1988), but no economic et al. (2003) suggested the presence of evidence for impact at potential could be established. Some impact-related the Eifelian/Givetian stage boundary of the mid-Devonian, a hydrothermal Pb-Zn mineralization does, however, occur, and suggestion that has remained controversial (Racki and has been mined locally (e.g., Reimold et al. 2005 and Koeberl 2004). As reviewed by Sandberg et al. (2002), the references therein; Hode et al. 2002). Late Frasnian mass extinction occured just prior to 364 Ma Bottomley et al. (1978) referred to an outcrop with a (within 20,000 years), representing a major extinction event “small dikelet [of melt breccia] at locality 3 of that decimated most groups of marine organisms. This event Svensson (1973).” Two samples from this site were described has been associated with alleged impact evidence, including a as containing 30–25% inclusions, predominantly quartz that weak enrichment found at the Frasnian/Famennian occasionally shows shock deformation, but also with clasts of boundary in southern China (Wang et al. 1991) and in a cross- feldspar and brecciated granite with incipient recrystallization boundary section in the state of New York (Over et al. 1997). and rare inclusions of sandstone. Melt matrix was In addition, the presence of so-called “microtektite-like glass” microcrystalline and granular. Interstitial devitrified glass did at a locality in Belgium was reported by Claeys and Casier also occur. One sample yielded a humped spectrum with a (1994) and discussed by Sandberg et al. (1988). The Siljan maximum age of 380 Ma but no plateau. The second yielded impact structure was proposed by Claeys and Casier as the a similar pattern but with a three-step plateau comprising 92% possible source for these Belgian microtektites. Furthermore, of release, but only because the analytical error on the middle the Amˆnau catastrophic event of central Germany has also step was 4.5% (around 10 times the two adjacent plateau been tentatively linked with impact (Sandberg et al. 2000, steps). Re-analysis of the Bottomley et al. (1978) data using 2002), although no bona fide impact evidence has been ISOPLOT (Ludwig 1999) yields an age of 358.3 ± 4.8 Ma reported for this event. The Amˆnau event was placed at the (2 σ). In addition, the final age quoted by Bottomley et al. Givetian/Frasnian boundary at 370 Ma. The (1978) was quoted at the 1 σ level and was not the plateau age biochronologically dated Alamo impact breccia of southern but an integrated total fusion age, in effect a K-Ar age. It is Nevada (e.g., Warme et al. 2002) occurred in the early well documented that fine-grained whole rock samples Frasnian punctata zone at about 367 Ma (Sandberg and showing younger ages in the highest temperature release Warme 1993; Sandberg and Morrow 1998; Sandberg et al. result from 39Ar recoil (McDougall and Harrison 1999) and 2002). The latter authors proposed links between this Alamo are more likely obtained from altered samples. Ar-Ar impact breccia and the Siljan and Flynn Creek impact analyses presented below demonstrate that the hydrothermal structures and suggested that these events could have resulted alteration around Siljan has led to alteration of many melt from a comet shower. Such a link is, however, highly unlikely, samples and it is likely that the age of 362.7 ± 2.2 Ma as the geometry of the distribution of the Alamo breccia rather significantly underestimates the true age of the impact at indicates a nearby source crater in Nevada. A further mass Siljan. In fact, a later publication by Bottomley et al. (1990) extinction close to the Devonian/Carboniferous boundary was quotes an age of 368 ± 1.1 Ma, apparently from the same placed at 357 Ma. dataset, although no explanation was given for the difference. Recently, the International Commission on Stratigraphy A K-Ar age of 349 Ma for “shock melt” from another (ICS) proposed a revised geological time scale (Gradstein locality was cited by Åberg and Bollmark (1985); Juhlin et al. et al. 2004; Ogg 2004; Gradstein and Ogg 2004), resulting in a (1991) cite a 40Ar–39Ar date for a “granitic pseudotachylite” shift of the ages for the Givetian, Frasnian, and Famennian of 359 ± 4 Ma, as well as two K-Ar ages for “doleritic stages of the Devonian period from 370–380 to 385.3–392, pseudotachylite” of 342 ± 3 Ma and 349 ± 2 Ma (whereby it 364–370 to 374.5–385.3, and 359.2–374.5 Ma (errors at 2.5– is assumed that these breccias would have been formed by the 2.7 Ma), respectively. This means that the Alamo impact impact event). The published constraints on the age for the event historically dated at about 367 Ma and the Siljan impact, large Siljan impact event thus clearly define a post-Silurian as dated previously, move both into the Famennian stage. age but are themselves not tightly constrained. The limited The Frasnian/Famennian event(s?) is (are) of global data available demand further dedicated chronological work, importance and represent(s) one of the five most significant especially in the light of the widely quoted possibility of a mass extinction events in the Phanerozoic (McGhee 1996). causal link between the Siljan impact and environmental Sandberg et al. (2002) suggested that several subcritical catastrophe in the Late Devonian. oceanic impacts could best explain the evidence from a The early ages for the impact event seemed to fall into an number of widely separated regions in the world. 594 W. U. Reimold et al.

While it is still debated what magnitude impact is (Juhlin et al. 1991). Planar deformation features (PDFs) in required to cause a significant global extinction event, to date quartz from the central region of the basement complex only one mass extinction—the K/T boundary event—has indicate shock pressures between 12 and 17 GPa (Svensson been unambiguously linked with a large impact event, i.e., the 1973; Tamminen and Wickman 1980; Grieve 1984; Åberg about 180 km wide, 65 Ma old Chicxulub impact structure in and Bollmark 1985). Apparent melt veinlets from small Mexico. exposures and from drill core have been compared to Clearly, it is important to investigate whether the Siljan pseudotachylite. Pseudotachylitic breccias (Reimold et al. impact event is indeed coeval with any of the Late Devonian 2005) of both doleritic and granitic composition have been catastrophic events, especially in light of the recent changes reported (Collini 1988). According to BodÈn and Eriksson, a of the geological time scale and with regard to the relatively sample of the former was used by Bottomley et al. (1978) for high uncertainty on the Siljan impact age (361–368 Ma). dating. Two of us (WUR, HH) observed millimeter-thick Improvement of the terrestrial impact cratering record, which melt-like covers on slickensides as well as veinlets of is still far from complete, is also required, with special regard pseudotachylitic breccia in a limestone quarry near Kallholen to the possibility of periodic increase in cratering activity and in the northwestern part of the ring of strata. possible relationship to terrestrial mass extinction events, as Microscopic analysis of these breccia veinlets revealed that discussed extensively in recent years (e.g., Rampino 2002 and all investigated occurrences from this quarry represent pure references therein). To this effect, we have carried out laser cataclastic breccia devoid of any evidence of melting spot argon as well as laser stepheating (40Ar–39Ar) dating on (frictional or other). Juhlin et al. (1991) stated that “true several samples of melt breccia that have recently been impact melts have not, so far, been found.” They proceeded, retrieved from the Siljan structure. however, to refer to two localities where, according to Åberg et al. (1988), impact melt had been suspected, but resolved GEOLOGY OF THE SILJAN IMPACT STRUCTURE that it was unlikely that this material represented impact melt rock. In the course of recent fieldwork, a number of melt rock The Siljan structure (Fig. 1) was formed in Svecokarelian occurrences were identified at the locations shown in Fig. 2 crystalline basement overlain by supracrustals of (coordinates are listed in Table 1). These samples provide the and Silurian age. Juhlin et al. (1991) provided a variety of basis for this chronological investigation. 40Ar–39Ar mineral ages, and U-Pb zircon as well as titanite ages for granitic lithologies ranging from 1436 Ma to SAMPLES 1702 Ma. Argon chronology on a number of dolerite samples yielded ages between 789 and 1098 Ma. An undefined “melt” Sample Si-1 originates from a dyke- or pod-like exposure indicated an age of 1163–1193 Ma. It is obvious that the post- approximately 1 m wide and more than 10 m long at Silurian impact event is chronologically well separated from Trollberget near the center of the impact structure (Fig. 2). these various target rock/basement ages. This melt rock cuts across granite as well as a mafic dike. On The 28–30 km diameter central part of the Siljan the basis of published geographic information, it would not be complex impact structure (Fig. 1) comprises shocked and impossible that this melt rock could represent the same brecciated granites (the so-called Dala granites) that have material dated by Bottomley et al. (1978, 1990). It comprises mostly been related to the Svecokarelian, but that also an extremely fine-grained matrix that, in reflected light, include several younger intrusives (compare above). This appears fully crystalline (Figs. 3a and 3b). The mode includes area represents a topographic high that is surrounded by a quartz, feldspar, and pyroxene or amphibole, as well as an relatively depressed, ring-shaped zone (up to a diameter of opaque phase (either magnetite or ilmenite). Clasts, in excess approximately 44 km), which is partly covered by lakes and of the matrix grain size, amount to about 15 vol% and include in which predominantly sedimentary strata of Ordovician and alkali feldspar and granite-derived (quartz plus feldspars) Silurian age occur. These strata include downfaulted lithic clasts. Most clasts of 0.3 mm grain size or larger are at Ordovician conglomerate and limestone and Silurian shale least partially annealed. Clast shapes are generally angular to and sandstone. Several authors have commented that the subrounded, but a small number of plastically deformed and central area could represent either a central uplift structure or well-rounded to folded clasts are also present. The matrix the remnant of a peak-ring structure (BodÈn and Eriksson appears altered in places where reddish patches of tiny 1988; Grieve 1988; Kenkmann and Von Dalwigk 2000). The crystallites of hematite occur. Some of these patches can be latter authors also presented a detailed structural analysis of recognized as loci of felsic ghost clasts. Other clasts display the Siljan structure. BodÈn and Eriksson (1988) reported that reaction rims that are also strongly hematite-stained. Besides the sedimentary strata occur, in part, as chaotically arranged the obvious thermal overprint on clasts, no shock deformation mega-blocks. could be discerned. Shatter cones have been observed throughout the uplifted In contrast, however, Bottomley et al. (1978) reported the central part of the structure and at some peripheral locations presence of planar shock deformation features in quartz in Laser Ar dating of melt breccias from Siljan 595

Fig. 2. The locations in the Siljan structure where samples for this study were taken. their samples, which could indicate that our sample does not have occurred. The brecciated parts of our thin section show necessarily represent the same material analyzed by these extensive aggregates of euhedral, medium-grained galena and authors. A single, 2.5 mm wide, strongly altered clast with a especially sphalerite, plus trace amounts of chalcopyrite. No subophitic texture of laths that likely originally represented indication of shock deformation was noted in quartz or feldspar could represent an inclusion derived from an igneous feldspar. precursor rock or of crystalline impact melt. Sample Si-3 (Figs. 3c and 3d) was obtained from the Sample Si-2 is a granitoid that is locally transected by a Museum of Natural History in Stockholm, where only the dense network of millimeter-wide breccia veinlets. These approximate locality of origin information was available veinlets are generally thinner than 3 mm and enclose or (compare Table 1). The sample is derived from a melt breccia infiltrate into cm- to dm-sized host rock clasts. The brecciated with granitic clasts from a locality close to that shown in parts of the sample are strongly impregnated with secondary Fig. 2 on the central uplift. The sample is a fluidal-textured calcite; feldspar in such areas is strongly altered to carbonate. melt rock with a matrix that optically appears glassy (locally) The sample originates from Stumsn‰s near the edge of the to crypto-crystalline. Matrix seems to flow around strongly central uplift (Fig. 2). It is impossible to ascertain whether the deformed (brecciated, partially annealed, and locally melted) breccia represents a pure cataclasite or if locally melting may clasts, most of which are granite-derived. Locally, the glass is 596 W. U. Reimold et al.

Fig. 3. Photomicrographs of the analyzed breccia samples from Siljan in plane-polarized light; all widths of view are 3.5 mm: a) and b) show sample Si-1, which is a clast-poor, aphanitic to microcrystalline melt breccia with most clasts clearly granitoid derived. c) and d) show sample Si-3, which is a fluidal-textured and variegated melt rock with a significant clast component. Again, most clasts are derived from granitoid precursors. Many clasts display evidence for plasticity and have been at least partially melted.

Table 1. Geographic coordinates of sample locations. Note clasts display only cataclasis. Matrix also contains some tiny, that the location for sample Si-3, which was provided by euhedral crystals of rutile. Planar deformation features the Stockholm Museum of Natural History, is not as (PDFs) have been observed in several quartz and feldspar precise as the others. grains within granitic clasts. Sample 1 Longitude/latitude Sample Si-4 is from a boulder in the northwest part of the Si-1 14°50.2′/61°03.0 structure (Fig. 2). It comprises a relatively clast-rich breccia Si-2 14°49.7′/60°53.2 (Figs. 4a and 4b) in which internally brecciated clasts are Si-3 ∼15°/61° prominent. The clast distribution is quite heterogeneous, and Si-4 15°50′/61°05’ it was obviously attempted to separate relatively clast-poorer Si-5 14°50.3′/61°03 material for the dating experiments. The matrix is essentially Si-6 14°50.1′/61°04’ clastic, but contains some hematite-bearing patches that optically appear as glass. They are characterized by the oxidized, mostly where it carries remnants of a mafic (gabbro presence of numerous tiny quartz clasts. There are also or amphibolitic) precursor rock. Aggregates of tiny crystals of fragments of melt, some of which are strongly extended and hematite lend these patches a reddish color. At least 30% of form stringers or schlieren. Larger granitic clasts are all clasts are completely annealed and many display plastic brecciated and partially annealed and locally even melted. No deformation in the form of folded shapes. It thus appears PDFs were observed in quartz or feldspar. likely that such clasts were melted and recrystallized. Other Sample Si-5 is also from a boulder near the center of the Laser Ar dating of melt breccias from Siljan 597

Fig. 4. Photomicrographs of the analyzed breccia samples from Siljan in plane-polarized light; all widths of field of view are 3.5 mm. (a) and (b) show sample Si-4, which is an aphanitic melt rock with locally very variable clast content (Fig. 4b shows a very clast poor area) and local hematite staining. The rare large granitoid clasts have been partially melted or locally annealed. c) Sample Si-5, which is a melt rock that is very similar to Si-3, but does not exhibit fluidal texture to the same degree. Plastic deformation and evidence of melting in clasts is, however, very evident. d) Sample Si-6 with several partially assimilated clasts in an aphanitic, locally microlithic melt matrix. structure, close to the Si-1 locality. This sample resembles Si- represent strongly extended (schlieren) granitic clasts (Fig. 4, but contains significantly less clastic component. 4d). The overall appearance could suggest that this matrix Nevertheless, it still is a clast-rich melt breccia (Fig. 4c). The was melt. The matrix is locally altered. Clasts are strongly rock is strongly hematite-stained. The clast content is brecciated and annealed. Several larger clasts have strongly generally granite-derived. PDFs occur in quartz of lithic clasts sericitized feldspar. Several large, brecciated granitic clasts as well as in several feldspar clasts. In some patches, the are impregnated with secondary carbonate. Locally, patches matrix is glassy or cryptocrystalline; in others, incipient of strongly altered melt matrix have remnants of small devitrification in the form of tiny microlites of feldspar is seen. feldspar laths. Shocked feldspar clasts with alternate twin Locally, microlites form dense aggregates indicating flow. lamellae converted to maskelynite are noted and a number of Flow directions are not uniform, which is interpreted as this diaplectic quartz or feldspar glass clasts occur; they only breccia representing an agglomeration of different melt display limited alteration. fragments or as a result of turbulent flow. Shocked plagioclase With the general lack of field control on the occurrences (diaplectic glass in alternate lamellae of polysynthetically of these breccias, it is basically impossible to evaluate twinned crystals) and fused feldspar and quartz (as identified whether they represent impact melt injections or on the basis of rosettes and spherulitic aggregates of pseudotachylitic breccia formed locally within the basement microcrystals in granite-derived clasts) are distinct. of the central uplift. The generally moderate degree of shock Sample Si-6 was taken from a local boulder in the deformation (12–17 GPa) reported for basement at the current northwest part of the central uplift, near H‰ttberg, close to a level of exposure favors the origin of the breccias in situ as 50 × 50 m large, partly excavated outcrop of granite, in which pseudotachylite or other pseudotachylitic breccia (for detail shatter cones are prominent. This sample represents a narrow on such breccias, refer to, e.g., Gibson and Reimold [2005] or (<10 cm wide) melt dikelet. The sample has a variegated Dressler and Reimold [2004]). fluidal-textured matrix with several narrow bands that Based on the above descriptions, samples Si-3 and Si-5 598 W. U. Reimold et al.

Table 2. Chemical compositions of Siljan samples. All Fe as Fe2O3. Wt% Si-1 Si-2 Si-3 Si-4 Si-5 Si-6

SiO2 54.06 54.38 59.46 63.83 61.46 55.85 TiO2 1.76 0.11 0.68 0.90 0.56 1.40 Al2O3 14.90 4.64 16.81 13.68 16.87 16.26 Fe2O3 9.13 2.67 5.17 5.69 3.86 8.48 MnO 0.18 0.14 0.05 0.09 0.03 0.16 MgO 3.53 0.33 1.15 1.41 0.47 3.97 CaO 4.04 12.18 1.30 1.94 0.95 2.24 Na2O 2.87 0.99 3.12 2.61 1.99 4.88 K2O 6.22 3.55 9.26 6.99 11.48 2.31 P2O5 0.34 1.44 0.18 0.19 0.10 0.27 LOI 2.66 9.67 2.08 1.58 1.32 3.80 Total 99.69 90.10 99.26 98.91 99.09 99.62 ppm Sc 14.3 0.90 7.09 12.7 11.4 12.4 V 121 11 43 72 15 116 Cr 19.8 25.1 12.9 24.5 6.1 16.5 Co 23.3 10.7 4.04 8.51 1.56 15.0 Ni 20 50 12 20 10 28 Cu 13 186 24 26 8 14 Zn 350 58100 87 75 178 150 As 0.15 20.0 0.20 0.25 0.50 1.05 Se 0.15 0.27 0.65 0.3 0.6 0.5 Br 0.3 3.2 0.3 1.5 0.5 0.4 Rb 167 156 196 296 348 105 Sr 410 233 357 295 206 386 Y376346454339 Zr 290 90 570 280 555 300 Nb 18 <323232521 Sb 0.05 6.49 0.04 0.13 0.06 0.17 Cs 1.12 1.78 1.42 7.70 2.46 1.09 Ba 830 720 1420 1650 690 520 La 77.5 36.9 90.2 160 365 46.2 Ce 98.8 86.1 118 253 466 91.8 Nd 49.1 45.8 60.3 95.5 247 45.6 Sm 8.48 10.2 8.02 11.7 25.8 7.81 Eu 2.03 1.38 1.58 1.57 2.06 1.26 Gd 6.85 8.11 6.45 8.6 14.7 7.27 Tb 1.01 1.26 0.85 1.10 1.81 0.94 Tm 0.49 0.49 0.44 0.61 0.92 0.54 Yb 3.02 2.63 3.21 4.02 5.74 3.97 Lu 0.47 0.35 0.50 0.67 0.88 0.62 Hf 6.61 0.71 11.0 7.67 13.8 8.32 Ta 0.62 0.30 0.34 1.01 0.65 0.94 W 0.2 0.2 0.1 0.7 1.4 0.5 Ir (ppb) <0.6 <0.8 <0.3 <0.2 <0.4 <0.6 Au (ppb) 0.3 <0.5 0.2 12.5 2.1 2.5 Th 4.59 3.15 6.02 40.2 15.5 13.7 U 0.71 3.03 1.34 3.23 1.72 2.49 have the largest amounts of relatively fresh melt material, and Analytical Methods it was anticipated that they would present the best chances for obtaining argon chronological results. The presence of bona The six rock samples were powdered and analyzed for fide shock deformation (PDFs and diaplectic glass) in both major element abundances in the X-ray fluorescence these samples forms a direct link between melt breccia laboratory of the School of Geosciences, University of the formation and the impact event. Witwatersrand, Johannesburg. A range of international and Laser Ar dating of melt breccias from Siljan 599

SARM reference materials were analyzed for calibration Trace element data are also quite variable and generally purposes. Accuracies from duplicate analyses are similar to in keeping with concentrations that one would expect for those reported by Reimold et al. (1994). The samples were granitoid dominated materials. The somewhat elevated Cu, also analyzed for 35 trace elements by instrumental neutron Co, Ni, and As values, as well as the very high Zn content for activation analysis at the Department of Geological Sciences, sample Si-2 are in line with the presence of secondary University of Vienna (for details on the methodology, sulfides, in particular sphalerite, in this sample. Iridium including information on instrumentation, standards, data concentrations in all six samples are below the detection limit reduction, accuracy, and precision, see Koeberl [1993]). The (0.5–1 ppb), indicating a maximum chondritic contribution to results are listed in Table 2. the melt rocks of less than 0.5%. The rare earth element Samples for argon chronology were prepared initially as (REE) patterns for this suite of samples are all very similar. square, 5 mm thick slabs, from which 100–300 µm thick They are relatively enriched in the light REE (LREE), with polished sections were prepared. Sections selected to contain relatively high concentrations, as expected for felsic crustal few clasts were released from the glass slide and rocks (chondrite-normalized La abundances between ultrasonically cleaned using methanol and deionized water. approximately 100 and 1000). The LREE patterns are flat; Sample Si-6 is clast-rich, and thus the area exhibiting most negative Eu anomalies are prominent but somewhat variable. melt was selected. Specimens were wrapped in aluminum Overall, the trace element characteristics of these samples are foils and irradiated at the McMaster Nuclear Reactor, Canada, consistent with their derivation from mainly granitic material, together with biotite standard GA1550 (98.79 ± 0.96 Ma) with a limited but significant contribution from mafic (Renne et al. 1998) to monitor neutron flux. The samples were material (see petrographic descriptions). packed adjacent to each other and represented a package only 3 mm long, sandwiched by standards. The J values calculated Argon Chronology from the two GA1550 standards were within 0.2% and thus a single J value is assigned to all samples with a 0.5% error. Sample Si-1 yielded a range of Ar-Ar laser spot ages Samples were analyzed using techniques outlined in Kelley from 350.6 ± 8.3 Ma to 375.9 ± 4.8 Ma (Fig. 5a). The data and Gurov (2002). The individual laser spot data are given in have an average of 2.9% atmospheric contamination but Table 3 and stepped heating data in Table 4. Twelve to fifteen exhibit little correlation between age and 36Ar/39Ar. The points were analyzed on each sample except Si-2, where just majority of the data points form a vertical array on the 36Ar/ five points were analyzed. Final weighted mean ages were 40Ar versus 39Ar/40Ar diagram, similar to those seen in glassy calculated using ISOPLOT-Ex after Ludwig (1999), which volcanic rocks containing devitrified glass (e.g., Turner et al. enhances the errors using the sum of students ‘t’ and square 1994). The scatter of data points is insufficient to form an root of the MSWD. isochron. Sample Si-2 (not shown in Fig. 5) yielded rather scattered RESULTS ages ranging from 588 ± 6 Ma to 788 ± 32 Ma. Although an attempt was made to target breccia veinlets, the resulting ages Chemical Composition were strongly variable and reflected mainly Ar extracted from partially reset host rock grains. The major element data indicate significant chemical Sample Si-3 yielded ages ranging from 371.7 ± 1.8 Ma to variability within this sample suite. Samples have 384.4 ± 2.7 Ma, neglecting two points (not shown in Fig. 5) intermediate SiO2 concentrations (54 to 64 wt%), with which fell more than 4 sigma below the mean value. The 39 relatively high Al2O3, Fe2O3, and alkali element contents. individual spot ages form a very tight cluster close to the Ar/ These compositions are strongly suggestive of mixing 40Ar axis (Fig. 5b) and an average atmospheric contamination between relatively more felsic (granite) and more mafic of only 0.3% (considerably less than, for example, Si-1), and precursor materials. Sample Si-2 is characterized by low total thus do not form an isochron. The data yield a weighted mean and elevated loss on ignition, concomitant with relative (Ludwig 1999) of 377.2 ± 2.5 Ma. enrichment in CaO, in accordance with petrographic Sample Si-4 yielded ages in the range 363.2 ± 4.9 Ma to observations of secondary carbonate and presence of 378.5 ± 3.8 Ma, neglecting one point. Analyses of this sample significant amounts of sulfide. The chemical compositions yielded an average atmospheric contamination of 5.8% alone do not allow identification of the true nature of these (Fig. 5c) and data scatter along a regression line, which samples as either impact melt injections into basement or corresponds to an age of 366.3 ± 9.0 Ma with a 40Ar/36Ar local formations of pseudotachylitic breccia in the central intercept of 343 ± 150 and an MSWD of 6.4. uplift. Notably, samples Si-3 and Si-5 have high K2O Sample Si-5 yields ages in the range 367.2 ± 2.9 Ma to contents, which favor these samples for argon dating 384.4 ± 1.8 Ma. Like Si-3, the analyses contained very low attempts, but could be an indication of secondary alteration atmospheric contamination with an average of just 0.5%. The (compare petrographic descriptions). data form a cluster close to the 39Ar/40Ar axis (Fig. 5d), but do 600 W. U. Reimold et al.

Table 3. Argon chronological data. Summary of laser spot data (amounts of 39Ar in cc STP × 10−12). J value = 0.001189 ± 0.000055 Siljan 1 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 19.691 0.0113 0.445 −0.00012 1.60 19.726 380.1 27.1 Spot 2 19.702 0.0124 0.418 0.00075 9.76 19.481 375.9 4.8 Spot 3 19.460 0.0120 0.352 0.00186 15.06 18.911 365.9 3.4 Spot 4 19.307 0.0125 0.380 0.00174 11.44 18.794 363.9 4.6 Spot 5 19.248 0.0126 0.389 0.00278 11.66 18.427 357.4 4.7 Spot 6 19.159 0.0134 0.375 −0.00010 10.00 19.188 370.8 5.2 Spot 7 19.212 0.0114 0.389 0.00134 8.76 18.817 364.3 5.4 Spot 8 19.172 0.0127 0.407 0.00365 7.82 18.092 351.5 6.6 Spot 9 19.486 0.0123 0.358 0.00256 7.91 18.729 362.7 5.9 Spot 10 19.429 0.0111 0.305 0.00164 14.64 18.944 366.5 3.6 Spot 11 19.180 0.0092 0.352 0.00273 5.96 18.375 356.5 7.8 Spot 12 19.268 0.0111 0.353 0.00333 4.90 18.283 354.9 9.4 Spot 13 18.497 0.0126 0.412 0.00093 8.05 18.221 353.8 5.8 Spot 14 18.907 0.0146 0.444 0.00293 5.51 18.041 350.6 8.3 J value = 0.001190 ± 0.000055 Siljan 2 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 51.766 0.0240 14.105 0.01956 2.16 45.987 787.5 16.6 Spot 2 38.130 0.0139 1.703 0.00488 44.11 36.687 653.6 3.0 Spot 3 32.730 0.0116 0.111 0.00107 26.66 32.412 588.5 2.9 Spot 4 33.928 0.0121 0.054 −0.00073 11.75 34.144 615.1 4.3 Spot 5 43.621 0.0121 0.145 0.00431 59.92 42.348 736.2 3.9 J value = 0.001192 ± 0.000055 Siljan 3 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 19.486 0.0117 0.055 −0.00001 29.66 19.491 376.9 2.3 Spot 2 19.465 0.0115 0.049 0.00013 28.69 19.425 375.7 2.3 Spot 3 19.236 0.0116 0.057 0.00014 80.65 19.194 371.7 1.8 Spot 4 19.372 0.0113 0.065 0.00029 41.14 19.286 373.3 2.0 Spot 5 18.399 0.0111 0.037 0.00032 50.91 18.305 356.1 1.9 Spot 6 19.443 0.0104 0.043 0.00011 34.17 19.410 374.9 4.4 Spot 7 19.877 0.0099 0.035 −0.00001 50.12 19.879 383.1 2.4 Spot 8 19.412 0.0080 0.041 0.00031 39.52 19.321 373.4 3.6 Spot 9 18.438 0.0100 0.000 0.00014 58.40 18.395 357.1 1.9 Spot 10 19.566 0.0101 0.042 0.00008 47.03 19.543 377.2 2.2 Spot 11 19.941 0.0096 0.046 0.00037 33.28 19.833 382.3 2.3 Spot 12 19.907 0.0098 0.039 −0.00016 28.33 19.954 384.4 2.7 Spot 13 20.007 0.0088 0.046 0.00079 36.79 19.775 381.3 2.1 Spot 14 19.577 0.0107 0.048 0.00055 52.61 19.416 375.0 2.3 Spot 15 19.597 0.0102 0.049 0.00019 104.10 19.541 377.2 1.8 Weighted mean of 13 points (95% confidence limit) 377.2 2.5 MSWD 13.0 J value = 0.001192 ± 0.000055 Siljan 4 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 19.558 0.0110 0.175 0.00146 69.78 19.127 370.5 2.3 Spot 2 19.928 0.0108 0.160 0.00241 46.27 19.216 372.1 2.0 Spot 3 19.743 0.0102 0.143 0.00242 27.39 19.029 368.8 2.4 Spot 4 19.967 0.0100 0.135 0.00422 30.59 18.720 363.4 2.4 Spot 5 19.815 0.0100 0.177 0.00292 33.95 18.951 367.4 2.3 Spot 6 19.967 0.0100 0.148 0.00425 38.21 18.711 363.2 2.1 Spot 7 20.512 0.0116 0.265 0.00395 21.96 19.346 374.4 3.1 Spot 8 20.289 0.0096 0.124 0.00320 14.29 19.344 374.3 3.5 Spot 9 20.397 0.0101 0.131 0.00509 9.84 18.893 366.4 5.2 Spot 10 19.357 0.0104 0.162 0.00418 9.95 18.123 352.9 4.9 Spot 11 21.228 0.0113 0.130 0.00677 14.82 19.229 372.3 3.4 Spot 12 21.836 0.0112 0.150 0.00762 16.46 19.585 378.5 3.8 Laser Ar dating of melt breccias from Siljan 601

Table 3. Continued. Argon chronological data. Summary of laser spot data (amounts of 39Ar in cc STP × 10−12). J value = 0.001192 ± 0.000055 Siljan 5 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 19.431 0.0091 0.025 0.00054 38.13 19.270 373.0 3.1 Spot 2 19.510 0.0097 0.026 0.00029 128.56 19.426 375.8 2.0 Spot 3 19.629 0.0098 0.026 0.00018 90.68 19.576 378.4 1.9 Spot 4 19.456 0.0098 0.032 0.00025 32.14 19.381 375.0 2.2 Spot 5 19.609 0.0100 0.021 0.00054 30.56 19.448 376.1 2.3 Spot 6 19.235 0.0103 0.025 0.00022 36.81 19.170 371.3 2.1 Spot 7 19.674 0.0091 0.032 0.00023 124.52 19.606 378.9 1.8 Spot 8 20.019 0.0082 0.037 0.00033 85.44 19.920 384.4 1.8 Spot 9 19.064 0.0100 0.033 0.00042 87.60 18.939 367.2 2.9 Spot 10 19.408 0.0107 0.029 0.00031 78.49 19.316 373.8 2.0 Spot 11 19.647 0.0106 0.020 −0.00015 28.94 19.692 380.4 4.8 Spot 12 19.431 0.0099 0.043 0.00052 55.08 19.276 373.1 1.9 Weighted mean of 11 points (95% confidence limit) 376.1 2.8 MSWD 16.0 J value = 0.001192 ± 0.000055 Siljan 6 40Ar/39Ar 38Ar/39Ar 37Ar/39Ar 36Ar/39Ar 39Ar 40Ar*/39Ar Age (Ma) +/− Spot 1 20.578 0.0107 0.425 0.00391 7.31 19.422 375.7 6.7 Spot 2 21.077 0.0083 0.491 0.00340 4.76 20.073 387.0 10.1 Spot 3 20.087 0.0099 0.315 0.00185 41.30 19.541 377.8 2.7 Spot 4 19.076 0.0105 0.302 0.00279 20.51 18.252 355.1 3.7 Spot 5 20.406 0.0105 0.393 0.00303 20.11 19.510 377.2 3.6 Spot 6 20.595 0.0119 1.018 0.00454 14.85 19.254 372.8 5.5 Spot 7 20.559 0.0106 0.938 0.00436 8.20 19.270 373.0 7.0 Spot 8 22.067 0.0109 0.237 0.00179 22.64 21.537 412.3 4.4 Spot 9 20.863 0.0099 0.195 0.00156 23.45 20.402 392.7 2.6 Spot 10 20.612 0.0096 0.244 0.00157 35.97 20.148 388.3 2.5 Spot 11 20.021 0.0088 0.491 0.00406 5.02 18.823 365.2 9.4 Spot 12 20.500 0.0106 0.577 0.00353 4.56 19.456 376.3 10.3 not allow an isochron to be constructed. The data yield a sample with lower atmospheric contamination (Si-5) and one weighted mean (Ludwig 1999) age of 376.1 ± 2.8 Ma. with higher atmospheric contamination (Si-4). Sample Si-6 yielded ages in the range 355.1 ± 3.7 Ma to Sample Si-4 yielded a relatively flat release spectrum but 412 ± 4.4 Ma, for an average atmospheric contamination of no plateau. The total gas age was 410 ± 5.8 Ma (Fig. 6a), 4.4% (Fig. 5e). Again, there was too much scatter to construct which is older than any of the individual laser spot ages for an isochron. the same sample. The high total gas age is caused almost Figure 5f shows the variability of the laser spot data, in entirely by step three (438.8 ± 3.3 Ma), without which the age particular highlighting the difference in atmospheric would have been within error of the weighted mean spot age. contamination between the low contamination samples, Si-3 Sample Si-5 yielded a plateau over 70.6% of the 39Ar release and Si-5, and the others. Si-3 and Si-5 are also the least altered with an age of 377.5 ± 4.1 Ma (Fig. 6b), which is within errors samples, and thus alteration is the most likely cause of the identical to the total gas age of 380.7 ± 4.0 Ma and the higher atmospheric contamination and less reproducible Ar weighted mean laser spot age of 376.1 ± 2.8 Ma for the same isotope data for the other samples. sample. The low Ca/K ratios of both samples are reflected in The spot data indicate that old ages resulting from low 37Ar/39Ar ratios throughout gas release (Figs. 6c and 6d). inherited argon released from clasts, which were not degassed Si-4 exhibits slowly falling 37Ar/39Ar ratios indicating some or only partly degassed in the impact event, are not a major Ca contamination in low temperature phases, whereas Si-5 problem in these melt samples. This observation is exhibits near zero 37Ar/39Ar concentrations. The difference corroborated by the annealed nature of many of the clasts. The between atmospheric contents of Si-4 and Si-5 in the spot data only “high” clast ages were those determined for sample Si-2, is again mirrored in the stepped heating data. which appears to have undergone only cataclastic In summary, the samples which are least altered yielded deformation and no melting, and several higher ages obtained the most reliable age data. Weighted mean ages for the for clast-rich sample Si-6. In view of the low clast samples with the lowest contamination are 377.2 ± 2.5 Ma contribution, we decided to test the difference between altered (Si-3) and 376.1 ± 2.8 Ma (Si-5), and the stepped heating and less altered samples by step-heating fragments of one plateau age for Si-5 falls within errors at 377.5 ± 4.1 Ma. It 602 W. U. Reimold et al. σ σ though no temperature though 377.5 4.1 Ar Age (Ma) ±2 Ar Age (Ma) ±2 39 39 / / * * Ar Ar 40 40 Ar (%) Ar% 39 39 . The steps indicate increasing laser power Ar Ar 39 39 Ar/ Ar/ 36 36 Ar Ar 39 39 Ar/ Ar/ 0.0212 0.00001 100.0 21.573 376.3 40.0 − 37 37 and include J error of 0.5%; J values as shown in Table 3). and include J error of 0.5%; values as shown in Table ep-heated samples of Siljan melt breccias σ Ar Ar 39 39 Ar/ Ar/ 38 38 Ar release 39 Ar Ar 39 39 Ar/ Ar/ 40 40 No plateau Siljan 5 Laser step Step 1Step 2Step 3Step 4Step 5Step 23.132 6Step 23.248 7Step 26.502 8Step 22.444 9Step 20.440 0.0072 10Step 20.519 0.0065 11Step 21.707 0.0062 12Step 20.121 0.0059 13Step 20.018 0.0065 0.0083 21.896 0.0069 0.0134 21.499 0.0055 0.0078 22.271 0.0069 0.0055 21.575 0.0066 1Step 0.0107 0.0064 2Step 0.00596 0.0044 0.0070 3Step 0.00243 0.0032 0.0070 4Step 0.00300 0.0062 0.0141 5Step 0.00180 0.0037 18.899 0.0038 6Step 0.00174 19.713 0.0013 14.8 7Step 0.00139 19.415 0.0015 20.7 8Step 0.00106 19.540 30.6 10Step 0.00085 20.060 0.0107 37.7 11Step 0.00134 19.561 0.0095 0.00166 39.8 12Step 19.625 0.0098 0.00199 21.371 50.9 13Step 19.503 0.0106 0.00250 19.838 22.530 51.9 14Step 0.0118 0.0001 20.338 25.617 60.9 15Step 0.0088 0.0000 24.988 21.911 72.3Plateau age over 70.6% of 0.0097 81.6 0.0006 20.410 19.925 0.0078 84.1 373.1 0.0007 24.936 20.109 0.0035 99.7 391.3 0.0000 19.857 21.393 0.0063 0.00087 438.8 0.0008 19.869 0.0057 0.00012 381.6 0.0007 19.621 0.0055 21.405 0.00020 350.1 0.0007 0.0100 0.0015 20.912 0.00027 353.1 2.6 0.0067 0.0022 21.531 0.00083 373.4 3.7 0.0051 10.4 0.00013 349.2 3.3 0.0034 13.9 0.00027 345.3 2.7 373.6 0.0008 25.6 0.00000 9.1 0.00043 365.8 0.0016 44.6 3.9 0.00052 375.6 48.0 12.2 0.00261 18.641 51.0 4.7 0.00010 19.679 54.2 2.2 0.00081 19.357 63.8 2.5 81.0 0.00039 19.459 6.7 88.9 12.0 19.816 89.2 362.0 19.521 91.3 380.2 19.545 94.6 374.5 19.503 100.0 19.710 376.3 20.186 382.6 24.217 377.4 2.0 20.381 377.8 3.7 24.696 377.1 1.6 19.741 380.7 1.6 389.0 3.0 457.5 5.4 392.4 2.3 465.5 381.2 1.8 1.9 9.3 21.8 7.0 9.8 7.2 Siljan 4 Laser step Table 4. Ar geochronology data for laser st Table measurements were possible (errors are 2 Laser Ar dating of melt breccias from Siljan 603

Fig. 5. Inverse isochron diagrams illustrating laser spot data for samples Si-1 to Si-6. Note that all of the samples plot close to the 39Ar/40Ar axis and are scaled accordingly. a) Sample Si-1; b) sample Si-3; c) sample Si-4. The regression line represents a fit obtained using ISOPLOT, which yields an age of 366.3 ± 9.0 Ma with a 40Ar/36Ar intercept of 342 ± 150 and an MSWD of 6.4. This is within error of the mean spot ages and plateau age obtained from other samples, though with poorer precision. d) Sample Si-5; e) sample Si-6; f) all samples plotted showing variation in atmospheric contamination: open symbols are Si-1, Si-4, Si-6 exhibiting higher atmospheric contamination, closed symbols are Si-3 and Si-5 symbols, and the dashed lines represent 1% and 10% atmospheric contamination. 604 W. U. Reimold et al.

Fig. 6. Results for stepped heating on selected whole rock chips: (a) shows sample Si-4, which did not yield a plateau age; (b) shows sample Si-5 showing a plateau over 70.6% of 39Ar release and an age of 377.5 ± 4.1 Ma; (c) shows 37Ar/39Ar release spectrum for sample Si-4; and (d) shows 37Ar/39Ar release spectrum for sample Si-5. seems likely that the larger scatter on ages of the other errors) and one that did not form a plateau. The analyses were samples is the result of devitrification, alteration, and also affected by 39Ar recoil during irradiation indicating the incomplete degassing of lithic and mineral clasts. We find that presence of a component of fine-grained potassium-bearing the best estimate of the age for the formation of the Siljan phyllosilicate that grew during post-impact hydrothermal crater is a combination of the best ages in a weighted mean alteration. The presence of a fine-grained phyllosilicate allowing for geological scatter by multiplying the error by component indicates that the resulting age could be an students ‘t’ multiplied by square root of MSWD, yielding an underestimate of the true age of the Siljan impact. New laser age of 377 ± 2 Ma (95% confidence limit). argon spot data and step-heating data presented above show some scatter, but this scatter is correlated with the alteration DISCUSSION AND CONCLUSIONS state of the samples. We have illustrated all data in order to emphasize the correlation, which appears to result in slightly The commonly quoted Ar-Ar age for Siljan (Bottomley lower ages for more altered samples. The two least altered et al. 1978) is an integrated age (an age calculated by samples yield consistent ages for laser spot and stepped summing all gas released, equivalent to a K-Ar age) from two heating, and support a revised age for this impact event of 377 samples, one of which yielded a plateau (358.3 ± 4.8 Ma, 2 ± 2 Ma (95% confidence limits). Thus, the Siljan case Laser Ar dating of melt breccias from Siljan 605 provides further evidence for the need to cautiously interpret Late Eocene (including Chesapeake Bay and Popigai), it will existing geochronological results on impact breccias. The key be much more difficult to obtain the same level of constraint to obtaining good age data for impact melt rocks is detailed on impact structures suggested to be of similar age to Siljan petrographic and chemical characterization of samples. In the (e.g., Charlevoix 357 ± 15 Ma, Woodleigh 364 ± 20 Ma, and Siljan samples, characterization of clast content and state of Flynn Creek 360 ± 20 Ma). Craters of Devonian age often alteration were crucially combined with Ar-Ar analysis of a have poorly constrained ages; more detailed work is required suite of samples, which also assessed the effects of alteration to improve the geochronology, before we can determine if and likelihood of clast-derived extraneous argon. However, they form a significant cluster. In addition, it is unlikely that well-preserved impact melt samples are rare and dating these these events of relatively minor magnitude, even if they had important terrestrial events continues to provide a challenge occurred as a cluster of events, would have resulted in a major to isotope geochronology. global extinction event, such as that at the Frasnian/ The revised Siljan age (377 ± 2 Ma) does not correspond Famennian boundary. with the previously accepted stratigraphic age for the Frasnian/Famennian boundary (364 Ma, Gradstein and Ogg Acknowledgments–Sharon Turner carried out the XRF 1996). Thus, any discussion of whether or not this impact analyses and Lyn Whitfield and Henja Czekanowska event can be correlated with any of the known catastrophic provided expert drafting and photographic support. Sample events in the Late Devonian period (Sandberg et al. 2002) Si-3 was kindly provided by Dr. Jan Olov Nystrˆm of the would be rendered invalid. However, the recent revision of Museum of Natural History, Stockholm. CK is supported by the geological time scale (Gradstein et al. 2004; Gradstein and the Austrian Science Foundation (FWF). SCS acknowledges Ogg 2004) has resulted in the curious situation that the new NERC fellowship NER/I/S/2002/00692 and SPK Siljan age falls within errors of the newly recommended age acknowledges funding from the Leverhulme Trust. Critical for the Frasnian/Famennian boundary at 374.5 ± 2.6 Ma. The reviews by Philippe Claeys and Birger Schmitz, as well as new boundary is based partly on a reappraisal of the Devonian editorial comments from Alex Deutsch, are much time scale using new U-Pb zircon ages from the Devonian appreciated. This is University of the Witwatersrand Impact Appalachian Basin in the USA (Tucker et al. 1998). This case Cratering Research Group Contribution No. 85. shows the extreme difficulty in tying absolute ages and biostratigraphic boundaries particularly in older events (e.g., Editorial Handling—Dr. Alexander Deutsch Deutsch and Sch‰rer 1994). Achieving the close control which has been achieved for the K/T boundary may simply REFERENCES not be possible in older sequences. Reliance will have to be placed more upon obtaining short term climate change Åberg G. and Bollmark B. 1985. Retention of U and Pb in zircons from shocked granite in the Siljan impact structure, Sweden. signals. Earth and Planetary Science Letters 74:347–349. Siljan may have originally been as large as 85 km BodÈn A. and Eriksson K. G., editors. 1988. Deep drilling in diameter (Henkel and Aaro 2005), but could Siljan have crystalline bedrock, volume 1: The deep gas drilling in the Siljan generated detectable global catastrophe and mass extinction? impact structure, Sweden and astroblemes. Berlin: Springer- Reimold and Koeberl (2002) discussed evidence that a strong Verlag. 364 p. Bottomley R. J., York D., and Grieve R. A. F. 1978. 40Ar-39Ar ages relationship between a large impact and global environmental of Scandinavian impact structures: I. and Siljan. extinction event only exists for the Cretaceous/Tertiary Contributions to Mineralogy and Petrology 68:79–84. boundary event at Chicxulub, an impact structure that Bottomley R. J., York D., and Grieve R. A. F. 1990. 40Argon-39Argon measures approximately 180 km in diameter. Several impact dating of impact craters. Proceedings, 20th Lunar and Planetary structures with diameters around or just below 100 km, Science Conference. pp. 421–431. Claeys P. and Casier J.-G. 1994. Microtektite-like glass associated including the Chesapeake Bay structure (85 km, age 35.5 Ma) with the Frasnian-Famennian boundary mass extinction. Earth at the eastern seaboard of the United States (Poag et al. 2004), and Planetary Science Letters 122:303–315. Manicouagan in Canada (100 km, age 214 Ma), and Popigai Collini B. 1988. Geological setting of the structure. Deep in Siberia (100 km, age 35.7 Ma), have not been related to drilling in crystalline bedrock, volume 1: The deep gas drilling in major global extinction events. the Siljan impact structure, Sweden and astroblemes, edited by BodÈn A. and Eriksson K. G. Berlin: Springer-Verlag. 364 p. Based on the currently defined impact flux for the Deutsch A. and Sch‰rer U. 1994. Dating terrestrial impact events. Phanerozoic (e.g., Hughes 2000; Schmitz and Peucker- Meteoritics 29:301–322. Ehrenbrink 2001), an impact event of comparable magnitude Ellwood B. B., Benoist S. L., El Hassani A., Wheeler C., Crick R. E. (producing craters in the 65–85 km diameter range) would 2003. Impact ejecta layer from the mid-Devonian: Possible have taken place at a likely rate of 1 per 10–20 million years. connection to global mass extinctions. Science 300:1734–1737. Fredriksson K. and Wickman F. E. 1963. Meteoriter. In Svensk Thus, the presently known cratering record of the Late naturvetenskap, edited by Lundholm B. Stockholm: Swedish Devonian and Early Silurian period is clearly incomplete. Natural Science Research Council. pp. 121–157. Although there appears to be a cluster of impacts during the Gibson R. L. and Reimold W. U. 2005. Shock pressure distribution 606 W. U. Reimold et al.

in the Vredefort impact structure, South Africa. In Large mass extinctions on the Earth. In: Catastrophic events and mass meteorite impacts III. Boulder, Colorado: Geological Society of extinctions: Impacts and beyond, edited by Koeberl C. and America. pp. 329–350. MacLeod K. G. Boulder, Colorado: Geological Society of Gold T. 1987. Power from the Earth. London: J. M. Dent & Sons Ltd. America. pp. 667–678. 208 p. Reimold W. U. and Koeberl C. 2002. Petrography and geochemistry Gold T. 1988. The deep earth gas theory with respect to the results of a deep drill core from the edge of the Morokweng impact from the Gravberg-1 well. In Deep drilling in crystalline structure, South Africa. In: Impact markers in the stratigraphic bedrock, volume 1: The deep gas drilling in the Siljan impact record, edited by Koeberl C. and Martinez-Ruiz F. Heidelberg: structure, Sweden and astroblemes edited by BodÈn A. and Springer-Verlag. pp. 271–292. Eriksson K. G. Berlin: Springer-Verlag. pp. 18–27. Reimold W. U., Gibson R. L., Koeberl C., and Dressler B. O. 2005. Gold, T. and Soter S. 1980. The deep-earth gas hypothesis. Scientific Economic ore deposits in impact structures and their geological American 242:154–161. setting. In: Impact tectonics, edited by Koeberl C. and Henkel H. Gradstein F. M. and Ogg J. G. 1996. A Phanerozoic time scale. Berlin: Springer-Verlag. pp. 479–552. Episodes 19:3–4. Renne P. R., Swisher C. C., Deino A. L., Karner D. B., Owens T. L., Gradstein F. M. and Ogg J. G. 2004. 2004— and DePaolo D. J. 1998. Intercalibration of standards, absolute Why, how, and where next! Lethaia 37:175–181. ages and uncertainties in 40Ar/39Ar dating. Chemical Geology Gradstein F. M., Ogg J. G., Smith A. G., Bleeker W., and Lourens L. J. 145:117–152. 2004. A new geologic time scale, with special reference to Renne P. R., Reimold W. U., Koeberl C., Hough R., and Claeys P. Precambrian and Neogene. Episodes 27:83–100. 2002. Critical comment on ‘K-Ar evidence from illitic clays of a Grieve R. A. F. 1988. The formation of large impact structures and Late Devonian age for the 120 km diameter Woodleigh impact constraints on the nature of Siljan. In Deep drilling in crystalline structure, Southern Carnarvon Basin, Western Australia’ by I. T. bedrock, volume 1: The deep gas drilling in the Siljan impact Uysal et al. Earth and Planetary Science Letters 201:221–232. structure, Sweden and astroblemes edited by BodÈn A. and Rondot J. 1975. Comparaison entre les astroblemes de Siljan, Suède, Eriksson K. G. Berlin: Springer-Verlag. pp. 328–348. et de Charlevoix, Quebec. Bulletin of the Geological Institutions Hallam A. and Wignall P. B. 1997. Mass extinctions and their of the University of Uppsala 6:85–92. In French. aftermath. Oxford: Oxford University Press. 320 p. Sandberg C. A. and Morrow J. R. 1988. Role of conodonts in Henkel H. and Aaro S. 2005. Geophysical investigations of the Siljan deciphering and dating Late Devonian Alamo impact impact structure: A review. In: Impact tectonics, edited by megabreccia, southeastern Nevada, USA (abstract). Koeberl C. and Henkel H. Berlin: Springer-Verlag. pp. 247–283. Proceedings, Seventh International Conodont Symposium. pp. Hode T., von Dalwigk I., and Broman C. 2002. A hydrothermal 93–94. system associated with the Siljan impact structure, Sweden— Sandberg C. A. and Warme J. E. 1993. Conodont dating, biofacies, Implications for the search for life on Mars. Astrobiology and catastrophic origin of Late Devonian (early Frasnian) Alamo 3:271–289. breccia, southern Nevada (abstract). Geological Society of Hughes D. W. 2000. A new approach to the calculation of the America Abstracts with Programs 25:77. cratering record of the Earth over the last 125 ± 20 Myr. Monthly Sandberg C. A., Ziegler W., Dreesen R., and Butler J. L. 1988. Late Notices of the Royal Astronomical Society 317:429–437. Frasnian mass extinction: Conodont event stratigraphy, global Juhlin C. and Pedersen L. B. 1987. Reflection seismic investigations changes, and possible causes. Proceedings, First International of the Siljan impact structure, Sweden. Journal of Geophysical Senckenberg Conference and 5th European Conodont Research 92:14,113–14,122. Symposium. pp. 263–307. Juhlin C. 1991. Scientific summary report of the Deep Gas Drilling Sandberg C. A., Morrow J. R., and Ziegler W. 2000. Possible impact Project in the Siljan ring structure. Swedish State Power Board origin of the enigmatic early Late Devonian Amˆnau breccia, U(G) 1991/14. 357 p. Rheinisches Schiefergebirge, Germany (abstract #3020). Kelley S. P. and Gurov E. 2002. Boltysh, another end-Cretaceous International Conference on Catastrophic Events and Mass impact. Meteoritics & Planetary Science 37:1031–1043. Extinctions: Impacts and Beyond. Kenkmann T. and von Dalwigk I. 2000. Radial transpression ridges: Sandberg C. A., Morrow J. R., and Ziegler W. 2002. Late Devonian A new structural feature of complex impact craters. Meteoritics sea-level changes, catastrophic events, and mass extinctions. In & Planetary Science 35:1189–1201. Catastrophic events and mass extinctions: Impacts and beyond, Koeberl C. 1993. Instrumental neutron activation analysis of edited by Koeberl C. and MacLeod K. G. Boulder, Colorado: geochemical and cosmochemical samples: A fast and proven Geological Society of America. pp. 473–487. method for small sample analysis. Journal of Radioanalytical Schmitz B. and Peucker-Ehrenbrink B., editors. 2001. Accretion of and Nuclear Chemistry 168:47–60. extraterrestrial matter throughout Earth’s history. New York: McGhee G. R., Jr. 1996. The Late Devonian mass extinction: The Kluwer Academic/Plenum Publishers. 492 pp. Frasnian/Famennian crisis. New York: Columbia University Schmitz B., Haggstrom T., and Tassinari M. 2003. - Press. 303 p. dispersed extraterrestrial chromite traces a major asteroid Ogg J. G. 2004. Staus of divisions of the international geologic time disruption event. Science 300:961–964. scale. Lethaia 37:183–199. Steiger R. J. and J‰ger E. 1977. Subcommission on geochronology: Over D. J., Conaway C. A., Katz D. J., Goodfellow W. D., and Convention on the use of decay constants in geo- and Gregoire D. C. 1997. Platinum group element enrichments and cosmochronology. Earth and Planetary Science Letters 36:359– possible chondritic Ru:Ir across the Frasnian-Famennian 362. boundary, western New York State. Palaeogeography, Svensson N. B. 1971. Probable meteorite in central Palaeoclimatology, Palaeoecology 132:399–410. Sweden. Nature 229:90–92. Racki G. and Koeberl C. 2004. Comment on “Impact ejecta layer Svensson N. B. 1973. Shatter cones from the Siljan structure, central from the mid-Devonian: Possible connection to global mass Sweden. Geologiska Foreningens I Stockholm Forhendlingar extinctions.” Science 303:471. 95:139–143. Rampino M. R. 2002. Role of the galaxy in periodic impacts and Therriault A. M., Grieve R. A. F., and Reimold W. U. 1997. Original Laser Ar dating of melt breccias from Siljan 607

size of the Vredefort structure: Implications for the geological Von Dalwigk I. and Kenkmann T. 1999. The Siljan impact structure: evolution of the Witwatersrand Basin. Meteoritics & Planetary New constraints for a diameter reconstruction (abstract). Science 32:71–77. Proceedings, 23rd Nordic Geological Winter Meeting. p. 24. Tucker R. D., Bradley D. C., Straeten C. A. V., Harris A. G., Ebert Wang K., Orth C. J., Attrep M. A. Jr., Chatterton B. D. E., Hou H., J. R., and McCutcheon S. R. 1998. New U-Pb zircon ages and the and Geldsetzer H. H. J. 1991. Geochemical evidence for a duration and division of Devonian time. Earth and Planetary catastrophic biotic event at the Frasnian/Famennian boundary in Science Letters 158:175–186. South China. Geology 10:776–779. Turner S. P., Kelley S. P., Hawkesworth C. J., and Mantovani M. Warme J. E., Morgan M., and Kuehner H. 2002. Impact-generated 1994. Magmatism and continental breakup in the South Atlantic: carbonate accretion lapilli in the Late Devonian Alamo breccia. High precision 40Ar-39Ar geochronology. Earth and Planetary In Catastrophic events and mass extinctions: Impacts and Science Letters 121:333–348. beyond, edited by Koeberl C. and MacLeod K. G. Boulder, Uysal I. T., Golding S. D., Glikson A. Y., Mory A. J., and Glikson M. Colorado: Geological Society of America. pp. 489–504. 2002. K-Ar evidence from illitic clays of a Late Devonian age for Wickman F. E., Blomqvist N. G., Geijer P., Parwel A. V., Ubisch H., the 120 km diameter Woodleigh impact structure, central and Welin E. 1963. Isotopic constitution of ore lead in Sweden. Carnarvon Basin, western Australia. Earth and Planetary Arkiv för Mineralogi och Geologi 3:193–257. Science Letters 192:281–189.