A Russian record of a Middle Ordovician meteorite shower: Extraterrestrial chromite at Lynna River, St. Petersburg region


Corresponding author. E-mail: anders.lindskog@geol.lu.se


Abstract– Numerous fossil meteorites and high concentrations of sediment-dispersed extraterrestrial chromite (EC) grains with ordinary chondritic composition have previously been documented from 467 ± 1.6 Ma Middle Ordovician (Darriwilian) strata. These finds probably reflect a temporarily enhanced influx of L-chondritic matter, following the disruption of the L-chondrite parent body in the asteroid belt 470 ± 6 Ma. In this study, a Volkhovian-Kundan limestone/marl succession at Lynna River, northwestern Russia, has been searched for EC grains (>63 μm). Eight samples, forming two separate sample sets, were collected. Five samples from strata around the Asaphus expansusA. raniceps trilobite Zone boundary, in the lower-middle Kundan, yielded a total of 496 EC grains in 65.5 kg of rock (average 7.6 EC grains kg−1, but up to 10.2 grains kg−1). These are extremely high concentrations, three orders of magnitude higher than “background” levels in similar condensed sediment from other periods. EC grains are typically about 50 times more abundant than terrestrial chrome spinel in the samples and about as common as terrestrial ilmenite. Three stratigraphically lower lying samples, close to the A. lepidurusA. expansus trilobite Zone boundary, at the Volkhov-Kunda boundary, yielded only two EC grains in 38.2 kg of rock (0.05 grains kg−1). The lack of commonly occurring EC grains in the lower interval probably reflects that these strata formed before the disruption of the L-chondrite parent body. The great similarity in EC chemical composition between this and other comparable studies indicates that all or most EC grains in these Russian mid-Ordovician strata share a common source––the L-chondrite parent body.


Several recent studies have documented high concentrations of extraterrestrial matter, most notably, numerous fossil meteorites (approximately 1–21 cm in diameter), in about 467 Myr old Middle Ordovician marine limestone (e.g., Schmitz et al. 1996, 2001; Tassinari et al. 2004; Schmitz and Häggström 2006; Korochantsev et al. 2009; Cronholm and Schmitz 2010). Although fossil meteorites are yet to be found outside a select few Swedish localities, an anomalous abundance of sediment-dispersed extraterrestrial chromite (EC) grains (63–250 μm) has been shown to occur in coeval strata at several localities in Sweden and also in the Puxi River section in China. This sudden and dramatic increase in extraterrestrial matter in the mid-Ordovician strata most likely reflects the disruption of the L-chondrite parent body in the asteroid belt, which produced swarms of variously sized debris that eventually crossed paths with Earth (e.g., Schmitz et al. 2001, 2003, 2008). In the present study, we try to expand the perspective by reconstructing the temporal changes in EC distribution in the mid-Ordovician section at Lynna River near St. Petersburg in western Russia. This is a “classic” section from which much important sedimentologic and paleontologic knowledge has been gained over the past 120 yr. In Sweden and in the one Chinese section studied so far, EC grains typically become about two orders of magnitude more abundant than in strata below, beginning slightly above the base of the Lenodus variabilis conodont Zone (e.g., Cronholm and Schmitz 2010), in the lower-middle Darriwilian (upper Dw1 stage slice sensui; Bergström et al. 2009). This prominent EC spike continues through strata representing the following 3–4 Myr before the concentrations gradually ebb out again. Even today, about two-thirds of Earthbound L chondrites—equating to about 30% of all identified ordinary chondritic meteorites—bear record of the disruption event by an 40Ar-39Ar gas retention age of 470 ± 6 Ma (e.g., Keil et al. 1994; Korochantseva et al. 2007; Nesvorný et al. 2008). Their cosmic-ray exposure ages are never more than 70 Myr, however, indicating that they come from larger bodies that either survived the event or reaccreted in its aftermath, as cosmic rays only penetrate the outermost meter or so of objects. More than a hundred fossil meteorites have now been found during quarrying of about 467 Myr old mid-Ordovician limestone at Kinnekulle in southern Sweden (B. Schmitz and M. Tassinari, unpublished results). Although the meteorites often retain some relict features, such as recognizable chondrule texture, most original minerals have become substituted principally by calcite, barite, and phyllosilicate pseudomorphs (Thorslund and Wickman 1981; Thorslund et al. 1984; Nyström et al. 1988; Nyström and Wickman 1991; Schmitz et al. 1996; Bridges et al. 2007). Chromite, a spinel-group mineral with the ideal formula FeCr2O4, typically constitutes the sole exception, as it is highly resistant to weathering and, thus, alteration and substitution. It is the dominant oxide in ordinary chondrites (Rubin 1997), typically accounting for approximately 0.25% by weight (Keil 1962), with chromite grains being both more abundant and larger sized in higher petrographic types (Bridges et al. 2007). Chromite of ordinary chondritic origin typically displays distinct chemical composition, with relatively little variation between similar-source grains (Bunch et al. 1967; Snetsinger et al. 1967; Affiatalab and Wasson 1980; Schmitz et al. 2003; Wlotzka 2005; Schmitz and Häggström 2006). As such, H, L, and LL ordinary-chondrite groups have individual, relatively narrow compositional ranges for chromite, but there is some overlap between the groups. Given the arduousness of finding fossil meteorites, extraction and analysis of sediment-dispersed chromite have thus become an alternative, and relatively easy, method to enable evaluation of the ordinary chondritic component in condensed sedimentary rocks (e.g., Schmitz et al. 2003). In the previously studied Middle Ordovician rock sections, EC abundance in the sediments increases from 1–2 EC grains per 100 kg rock to several grains per kilogram of rock; e.g., approximately 3 and 6 grains kg−1, respectively, in the Hällekis and Komstad sections, about 300 km apart in southern Sweden (Schmitz and Häggström 2006; Häggström and Schmitz 2007), and up to 4 grains kg−1 at Puxi River in China (Cronholm and Schmitz 2010). Pre-L. variabilis EC concentrations are comparable to those of similar condensed strata of much younger age, for example, pelagic Paleocene limestone from the Gubbio section in Italy (Cronholm and Schmitz 2007). Most of the sediment-dispersed chromite grains probably stem from weathering and disintegration of unmelted micrometeorites that have passed through the atmosphere affected by little or no melting. This is indicated by high concentrations of solar-derived noble gases in the EC grains (Heck et al. 2008; Meier et al. 2010). The solar wind only affects the outermost nanometers of a (micro-)meteorite, indicating that at least some part of the chromite grains must have shared surface with the meteorite.

The meteorite- and EC-rich Middle Ordovician rock interval encompasses early parts of the Great Ordovician Biodiversification Event (GOBE), which saw the diversification and radiation of the so-called Paleozoic Evolutionary Fauna (see Webby et al. 2004; Schmitz et al. 2008). Rasmussen et al. (2007) found evidence in the Lynna River section and other East Baltic sections, of an important regional turnover and diversification event of brachiopod species coinciding with the base of the L. variabilis conodont Zone. Because this is approximately the stratigraphic level of the first appearance of abundant EC grains in sections in Sweden and China, Schmitz et al. (2008) proposed that an increase in the influx of kilometer-sized asteroids following the L-chondrite disruption event may have stimulated biodiversification during the GOBE, in line with the so-called Intermediate Disturbance Hypothesis (see further references in Schmitz et al. 2008). Similar views were brought forward by Culler et al. (2000).

The aim of this study was to investigate in greater detail if a similar chromite signal is present in the mid-Ordovician Lynna River section. Already, in a pilot study, Korochantsev et al. (2009) confirmed that EC grains at least occur in this section, in beds that are dated to the uppermost Asaphus expansus trilobite Zone of the Kunda Baltoscandian Stage (lower-middle Darriwilian). These authors only studied two small samples, but argued for an inverted stratigraphic distribution trend of the EC grains compared with that seen in the Swedish sections, with the highest concentrations of EC grains at a stratigraphically higher level than in Sweden (see below). From previous studies, it is known that the EC content through the EC-rich interval can vary by a factor of about two to five, most likely reflecting differences in the sedimentation rates of individual beds (Schmitz and Häggström 2006; Cronholm and Schmitz 2010). The observed two orders-of-magnitude difference in EC concentration, however, between every bed deposited before the L-chondrite disruption event and those that formed after, cannot be explained by changes in sedimentation rate alone. To further understand the interplay between the flux of extraterrestrial chromite to the sea surface, sedimentation rate and hydrodynamic mineral sorting processes at the sea floor, and their influence on EC content in individual beds, we here also quantify the relation through the Lynna River section between EC grain abundance and heavy minerals of terrestrial origin, such as ilmenite.

Geologic Setting and Stratigraphy

The Ordovician of Baltoscandia and St. Petersburg Region

The Baltoscandian region (here including westernmost Russia) harbors laterally extensive Ordovician rocks, formed from the sediments of a vast epeiric sea that covered large parts of the paleocontinent Baltica at the time (Lindström 1963; Jaanusson 1972; Nielsen 2004; Dronov 2005; Dronov and Mikuláš 2010). Long-term tectonic stability has maintained a distinct, time-persistent facies zonation in the region, resulting in so-called confacies belts, which record a general deepening of the Baltoscandian paleobasin roughly along a northeast-southwestern transect (Fig. 1A). Terrigenous sediment input was limited, and net deposition rates were most often exceedingly slow, typically only a few millimeters per thousand years (e.g., Lindström 1971; Schmitz et al. 1996; Nielsen 2004). During the Ordovician, Baltica drifted from southern subpolar to subtropical-tropical latitudes, while rotating counterclockwise (Fig. 2; Cocks and Torsvik 2006).

Figure 1.

 A) Map of the Baltoscandian region, with confacies belts (CB) indicated (modified from Mellgren and Eriksson 2010). CCB–Central CB; ECB–Estonian CB; LCB–Lithuanian CB; LT–Livonian Tongue; MB–Moscow Basin; SCB–Scanian CB. Rectangle indicates outline of b. B) Map of the St. Petersburg region, with general geology and notable geographic features indicated (modified from Dronov and Mikuláš 2010). Square indicates outline of c. C) Map of the Lynna River area, with notable geographic features indicated.

Figure 2.

 A reconstruction of Middle Ordovician paleogeography (early Darriwilian, Kundan, about 466 Ma). Baltica was situated between southern temperate and subtropical realms. Note that large parts of most paleocontinents were covered by epeiric seas. This paleogeographic reconstruction was made using the software BugPlates (courtesy of StatoilHydro), developed by Trond H. Torsvik (available at www.geodynamics.no). See Cocks and Torsvik (2006) for further information about Paleozoic paleogeography.

The St. Petersburg region straddles the border between the Baltic igneous shield and the Moscow basin (Dronov 2005; Dronov and Mikuláš 2010). The Ordovician strata of the region comprise some 100–200 m of nearly flat-lying Ordovician rocks. Lowermost Ordovician (Tremadocian) strata mainly consist of unconsolidated clays and quartz sands. The succeeding Ordovician strata are characterized by carbonate rocks, which record a gradual transition from subpolar via temperate (“cool-water”) to tropical (“warm-water”) environment through time, entailing a successive increase in carbonate production. During the early-to-middle Darriwilian, Baltica was positioned at approximately 45 oS, with formation of temperate carbonates (Fig. 2). Ordovician carbonates lie upon an elevated area, called the Ordovician plateau (e.g., Dronov et al. 2002; Dronov and Mikuláš [2010] and references therein). This plateau is bounded to the north by an escarpment called the Baltic-Ladoga Glint (or, simply, the Baltic Glint), resulting in a line of natural Cambrian-Ordovician outcrops (Fig. 1B).

The Ordovician strata of Baltoscandia are subdivided into eighteen regional stages, five of which are in the Middle Ordovician Series (Fig. 3; Bergström et al. 2009; cf. Dronov and Mikuláš 2010). Most of the Ordovician stages are recorded in the St. Petersburg region, but only some aspects of the Volkhov and Kunda stages—the strata of which have been sampled for this study (Figs. 4, 5)—are relevant to discuss here. Each of these stages is considered to represent at least one depositional sequence (e.g., Rasmussen et al. 2009; Dronov and Mikuláš 2010; cf. Dronov et al. 2011). In the St. Petersburg region, Volkhov (BII zone) strata are typified by glauconite-speckled, variably argillaceous bioclastic limestone, ranging from marl to grainstone (Dronov 2005; Dronov and Mikuláš 2010). The succeeding Kunda (BIII) Stage begins with the appearance of the trilobite species Asaphus expansus, and its base is marked by a conspicuous surface of nondeposition that records a significant sea-level fall (estimated to 30–40 m; e.g., Dronov 2005; Dronov and Mikuláš 2010). Kundan strata comprise variably argillaceous bioclastic limestones, often collectively referred to as “orthoceratite limestone.” Unlike in subjacent Volkhovian strata, glauconite is rare, whereas iron ooids may be abundant. The Middle Ordovician deposits in the St. Petersburg region have been interpreted as tempestites formed in a storm-dominated, shallow-marine carbonate ramp environment (e.g., Dronov et al. 2002; Hansen and Nielsen 2003; Dronov and Mikuláš 2010). As such, individual beds may have been laid down in a matter of days or even hours. The beds were then reworked for long periods—often several thousands of years—before subsequent sediment deposition. Graded beds, rich in coarse-grained shell debris, and various storm-related erosional features are typical. These tempestite beds may have substantial lateral extent and some can be traced for hundreds of kilometers along the Baltic-Ladoga Glint (thus providing very precise regional correlation). A sedimentologic west-to-east proximal-to-distal trend is clearly discernible in the region.

Figure 3.

 The Middle Ordovician stratigraphy of the St. Petersburg region (modified from Dronov and Mikuláš 2010).

Figure 4.

 The sampled strata at Lynna River, St. Petersburg region, northwestern Russia, with the Volkhov-Kunda Stage boundary indicated.

Figure 5.

 The sample stratigraphy at the Lynna River locality (cf. Ivantsov 2003; Schmitz and Häggström 2006; Rasmussen et al. 2007, 2009).

The geology of the St. Petersburg region has been actively studied since the early 19th century. A chronicle of significant geologic ventures undertaken in the St. Petersburg region is provided by Dronov and Mikuláš (2010), and a review of the regional Ordovician is provided by Dronov et al. (2005).

The Lynna River Section

Along the Lynna River, a tributary to the larger Syas River, rocks from the mid-Volkhovian (uppermost BIIβ) throughout the mid-Kundan (lowermost BIIIγ) are exposed (Figs. 4 and 5; e.g., Pushkin and Popov 2005; Dronov and Mikuláš 2010). In total, the strata may reach a thickness of approximately 15 m (Volkhovian approximately 3.5 m, Kundan >7.5 m). Rock color may vary on smaller scales, but gray hues dominate the river valley.

Volkhovian strata consist of glauconitic limestones, typically (bioclastic) wackestone and packstone, intercalated with clay/marl (Dronov and Mikuláš 2010; Hansen 2010; cf. regional description above). Glauconite-rich hardgrounds are recurrent. The exposed BIIβ zone (Zheltiaki Member) is characterized by a yellow stain, whereas the BIIγ zone (Frizy Member) is the more typical gray. The boundary between the Volkhov and Kunda stages is marked by a glauconite-impregnated hardground, situated in a strongly glauconite-enriched limestone interval (Fig. 5).

The lower Kundan strata belong to the Lynna Formation, which typically consists of gray, argillaceous limestones, ranging from mudstone to wackestone, intercalated with clay/marl (Dronov and Mikuláš 2010; Hansen 2010). The uppermost 0.5 m of the Lynna Formation differs from the interval below, in being characterized by more massive, reddish limestone beds (see below). The overlying Sillarou Formation begins with a conspicuous, decimeter-thick clay/marl interval, which is subdivided into a lower red and an upper gray part. A discontinuous, highly argillaceous limestone layer is found in between. Gray limestones then resume and continue throughout the overlying strata, in which cephalopod conchs are common (hence the lithology’s informal name, “orthoceratite limestone”). Some uncertainty exists regarding the stratigraphic extent of the Sillaoru Formation at the locality. Iron ooids are scarce throughout.

Sediments forming the strata at Lynna River were deposited in a mid-carbonate ramp environment, below fair-weather wave base (Hansen 2010). This relatively deep setting resulted in a rather complete stratigraphy, with only minor hiatuses, as compared with coeval, more shallow-water, sections toward the west (Rasmussen et al. 2009).

Most previous studies of the Lynna River section have revolved around paleontologic aspects of the rocks, typically with little bearing on the subject in this study (but see, e.g., Hansen and Harper 2003; Hansen and Nielsen 2003; Ivantsov 2003; Pushkin and Popov 2005; Rasmussen et al. 2007, 2009; Rasmussen and Harper 2008; Hansen 2010). Korochantsev et al. (2009) performed a limited investigation of chromite content in the Lynna River strata. Two samples, one from the upper Lynna Formation (6.5 kg) and another from the Lynna-Sillaoru formational boundary (4.02 kg), yielded notable amounts of chromite grains. Chemical analyses showed that a majority of the grains were likely of ordinary chondritic origin. The lower sample also contained two Ni-rich grains. Korochantsev et al. (2009) suggested EC concentrations of 0.6 and 2.9 grains kg−1 for the lower and upper sample, respectively, and noted that an increased EC concentration upward through the stratigraphy is in contrast with EC patterns previously observed at other localities, which typically display decline upward, after an initial EC “spike.” An inverse relationship between chromite and clay content was also observed; as chromite content increased, clay content decreased, however, not at the same rates. The sample levels studied by Korochantsev et al. (2009) roughly correspond to samples Ly1 and Ly4 of this study; however, due to different sample processing methods, results are not directly comparable. Recalculations by the present authors indicate an EC concentration of 4.2 grains kg−1 in the upper sample of Korochantsev et al. (2009).

Materials and Methods

Two separate sets of samples were collected at Lynna River (Fig. 1C), leaving approximately 2 m sample gap in between (Fig. 5). One sample set is centered around the Volkhov-Kunda Stage boundary, which corresponds to the A. lepidurus-A. expansus trilobite Zone boundary and also the Volkhov-Lynna formational boundary (see Ivantsov 2003). The conodont biostratigraphy at Lynna River is not yet well resolved; however, this sample set is most likely within the Baltoniodus norrlandicus conodont Zone, and slightly below the base of the L. variabilis Zone (cf. Schmitz and Häggström 2006; Rasmussen et al. [2007] and references therein). The other sample set covers the succeeding A. expansus-A. raniceps trilobite Zone boundary, coinciding with the Lynna-Sillaoru formational boundary, and probably also the L. variabilis-Yangtzeplacognathus crassus conodont Zone boundary, in the lower-middle Kundan. In total, eight rock samples, in the range of 9.8–14.5 kg and with a total weight of 103.7 kg, were selected for analysis (Table 1). The samples in the lower set (three samples, total weight 38.2 kg) represent specific beds that have been assigned individual numbers by previous workers (see, e.g., Hansen and Harper 2003). The same bed numbers are here used as sample numbers. In the upper set (five samples, total weight 65.5 kg), we assigned our own bed numbers. Both sampled intervals are characterized by mud-, wacke-, and packstone carbonates. Gray strata dominate in the lower sample interval. Ascending through the stratigraphy, the strata of the upper sample interval become increasingly tinted red (oxidized). Glauconitic grains are conspicuously present in the lower sample interval, but are absent/rare in the upper interval. Golden-brown (goethite/limonite?) grains are abundant/common in the uppermost samples (cf. Eriksson et al. 2012). The Ly4 sample level harbors 5–10 cm of poorly consolidated, red clay/marl. Overlying this red clay/marl, the Ly5 sample level constitutes a string of limestone nodules, or rather a discontinuous limestone bed, in turn overlain by a few centimeters of gray clay/marl. Overall, hand samples display varying redox conditions at smaller scale, as inferred by variegated (grayish-reddish) color.

Table 1.   Sediment-dispersed extraterrestrial chromite (EC), other chrome spinel (OC), and ilmenite (sensu lato) in samples from Lynna River (see Fig. 5 for stratigraphic positions of samples). See also Fig. 6.
SampleSample weight kgNo. EC
grains kg−1
No. OC
grains kg−1
No. ilmenite grainsIlmenite
grains kg−1
  1. a Data for unpolished grains.

Total103.7498 11 674 

Samples were first carefully cleaned with regular tap water, and then dissolved in 6 M hydrochloric acid (HCI), following the approach established by Schmitz et al. (2003). All processed samples reacted strongly to HCl treatment, indicating considerable CaCO3 content. The >32 μm HCl residue was then gently stirred in 3.8 M hydrofluoric acid (HF) at room temperature for 10–15 min, after which a few liters of water were added to dilute the acid, and the sediment was allowed to settle for 24 h. Thereafter, the acid solution was gently decanted and water was re-added, once a day, until an acceptable pH was reached. The remaining residue was separated into three size fractions; 32–63, 63–355, and >355 μm. Only the 63–355 μm fraction was studied in detail. Excessive amounts of undissolved glauconitic grains prompted the use of density separation techniques, employing LST Fastfloat heavy liquid, for the HF residues of samples 28 U, 30, and 31 prior to the search for heavy minerals. The dried 63–355 μm fraction of each sample was studied in an optical stereo microscope and suspected chromite and ilmenite (sensu lato, approximately FeTiO3) grains manually extracted. In some sample residues, abundant shiny, opaque iron oxide grains, and carbon-rich (coal?) grains—superficially similar to chromite—introduced considerable difficulty to the search for relevant grains (cf. Alwmark and Schmitz 2007, 2009). All residues were meticulously scanned for grains at least three times, and usually none or only single grains were found during the third search. In previous studies of EC content in sediments, stronger HF (11 M) has been used (e.g., Schmitz and Häggström 2006; Cronholm and Schmitz 2010). As we here wanted to quantify also the abundance of ilmenite, we chose a weaker HF (3.8 M) that generally does not dissolve ilmenite grains >63 μm during the short leaching period used.

All extracted grains were mounted onto carbon tape and chemically analyzed with an Oxford Instruments INCA X-Sight energy-dispersive spectrometer, attached to a Hitachi S-3400N scanning electron microscope (SEM, with 15 kV acceleration voltage, using cobalt as a standard). Mineral types were thus confidently identified. Identified chromite grains were subject to additional (30-second counting live-time) analyses, and individual grain backscatter images and spectra were compiled and stored in a database. Excluding the grains from samples Ly3 and Ly4, which were kept intact for future studies of noble gas content, chromite grains were then cast in epoxy, smoothly polished, and analyzed again (80-second counting live-time). At least three point analyses were performed on each grain, to provide information on possible element zonation and ensure reproducibility of results (typically <2% major-element discrepancy between individual point analyses). Individual grain average compositions were subsequently calculated and compiled into a dataset, which is available in complete, original form as an online supplement to this article. The reliability and accuracy of the chemical analyses have been established in our group over a long time by recurrent analyses of several reference standards, and by comparison of results for the same grains using a Cameca SX-50 electron probe microanalyzer at the Department of Geophysical Sciences, University of Chicago. Potential overlap in emission peaks for, e.g., Ti and V (see Snetsinger et al. [1967] and references therein) has been accounted for, and also data for these elements are reliable. Analytical accuracy was controlled through the USNM 117075 chromite (Smithsonian) reference standard (Jarosewich et al. 1980) and three silicate reference standards (Alwmark and Schmitz 2009).

In this study, EC grains are identified according to compositional characteristics as defined by Schmitz and Häggström (2006), with a revised TiO2 range as introduced by Cronholm and Schmitz (2010): MgO approximately 1.5–4 wt%, Al2O3 approximately 5–8 wt%, TiO2 approximately 1.4–3.5 wt%, V2O3 approximately 0.6–0.9 wt%, Cr2O3 approximately 55–60 wt%, and FeO approximately 25–30 wt%. The notable TiO2 and V2O3 concentrations are considered to be especially important EC indicators (e.g., Bunch et al. 1967; Schmitz and Häggström 2006). Grains that display significant deviation from any of the above ranges, not readily explained by “simple” (e.g., diagenetic) element interchange, are classified as “other chrome spinel” (OC). This, however, does not automatically exclude an extraterrestrial origin (cf. Wlotzka 2005).


Chromite and Ilmenite Content in Samples

Table 1 and Fig. 6 provide an overview of the heavy mineral content in the samples. In summary, the lower sample set yielded a total of only four chrome spinel grains, of which two were interpreted as EC. This corresponds to average concentrations of 0.05 grains kg−1 for both EC and OC grains. Ilmenite grains amassed to nine grains, corresponding to an average concentration of 0.24 grains kg−1. The upper sample set yielded 505 chrome spinel grains, of which 496 were interpreted as EC (note that data from samples Ly3 and Ly4 are based on low-quality analyses of unpolished grains). This corresponds to averages of 7.6 EC grains kg−1 and 0.14 OC grains kg−1. In total, 665 ilmenite grains were found, corresponding to an average concentration of 10.2 grains kg−1 (however, nearly two thirds of these were found in the Ly5 sample alone). In all samples but Ly5, EC grains outnumbered ilmenite grains, and the highest EC concentrations, 10.2 grains kg−1, were found in sample Ly4. A trend of increasing heavy mineral grain concentration, when ascending stratigraphically, is seen in the upper sample set. The average size of extracted chromite grains is estimated to be approximately 90 × 70 μm, with individual axis lengths varying between approximately 40 and 200 μm. As in preceding studies, most EC grains are anhedral and angular, and display an excellent state of preservation, indicating limited time in the sedimentary system (Fig. 7A, C; cf. Alwmark et al. 2010; Cronholm and Schmitz 2010). The EC grains may typically be classified as very angular to sub-angular, with a select few being classified as sub-rounded (Fig. 7B, D), following Powers (1953). OC grains range from being very angular to well-rounded; however, the small number of grains found in this study does not allow any meaningful comparison between grain types. Sixteen EC grains contain notable amounts of NiO (max. 1.26 wt%). Such grains often display enhanced MgO or FeO concentrations, typically at the expense of one another.

Figure 6.

 Heavy mineral content in samples, expressed as grains kg−1. Geologic formations and biozones are indicated. See also Fig. 5 and Table 1.

Figure 7.

 A selection of chromite grains from this study (SEM, backscattered electron images). All scale bars 50 μm. A) Extraterrestrial chromite (EC) grain displaying characteristic angularity and excellent state of preservation (very angular/angular; sample Ly1, EC #72). B) EC grain displaying uncharacteristic rounding (sub-rounded/sub-angular; sample Ly5, EC #105). C) EC grain displaying characteristic angularity and excellent state of preservation (very angular/angular; sample Ly4, EC #37). Note the Widmanstätten-like pattern on the grain surface. D) EC grain displaying uncharacteristic rounding (sub-rounded/sub-angular; sample Ly2Ö, EC #21).

Ilmenite and remaining silicate grains are similar in size to chromite, and display varying levels of roundness, ranging from euhedral to well-rounded. Little can be interpreted from this, however, as most are probably etched by HF. Glauconitic grains, often in the form of fossil steinkerns, are ubiquitous in the 63–355 μm residues from the lower sample set, whereas achromatic, transparent siliciclastic grains (e.g., quartz) dominate in samples Ly1 and Ly2Ö. The Ly3, Ly4, and Ly5 residues are dominated by golden-yellow-colored, iron-rich grains (goethite/limonite?), again often in the form of fossil steinkerns. Biotite grains are typically common, as are iron-oxide and coal-like grains. Chromium-bearing nonspinel grains also occur. Apart from the various mineral(-oid) grains, iron-rich spherules are common in most sample residues.

Chemistry of Chromite Grains

Data from chemical analyses of EC grains are presented in Table 2 and Fig. 8 (full data set in online supplement), and a comparison with grain data from various preceding chromite studies is presented in Table 3 and Fig. 9 (cf. Schmitz and Häggström 2006; Häggström and Schmitz 2007; Cronholm and Schmitz 2010). Overall, EC chemical composition is homogeneous, with little variation within or between grains, nor between samples. Most grains conform within analytical error to the compositional ranges for ordinary chondritic chromite as defined above; however, some deviations for single compounds, most notably in MgO and FeO concentration, have been allowed in some cases, as there are clear element-exchange relationships.

Table 2.   Average chemical composition (compound wt%± 1 s.d.) of sediment-dispersed extraterrestrial chromite (EC) grains in samples from Lynna River. See also Fig. 8.
SampleNo. of grainsMgOAl2O3TiO2V2O3Cr2O3MnOFeONiOZnOTotal
  1. a Data for unpolished grains.

  2. b Not detected.

28U14.37 ± 0.006.25 ± 0.002.40 ± 0.000.75 ± 0.0056.79 ± 0.000.72 ± 0.0026.48 ± 0.001.39 ± 0.00n.d.b99.14 ± 0.00
3012.91 ± 0.003.49 ± 0.002.55 ± 0.000.66 ± 0.0060.87 ± 0.000.69 ± 0.0028.22 ± 0.00n.d.n.d.99.40 ± 0.00
Ly1762.78 ± 1.125.99 ± 0.333.00 ± 0.420.72 ± 0.0859.15 ± 1.350.77 ± 0.2025.81 ± 2.110.02 ± 0.121.14 ± 1.2599.39 ± 0.58
Ly2Ö1132.79 ± 1.306.08 ± 0.492.94 ± 0.420.74 ± 0.0959.33 ± 1.720.81 ± 0.2025.92 ± 2.240.04 ± 0.130.89 ± 0.8899.52 ± 0.61
Ly3a812.48 ± 1.386.48 ± 1.642.34 ± 0.670.63 ± 0.4355.60 ± 5.880.14 ± 0.3729.02 ± 7.46n.d.1.20 ± 2.3697.89 ± 1.62
Ly4a1022.19 ± 1.396.48 ± 1.552.49 ± 0.600.44 ± 0.2854.25 ± 7.060.33 ± 0.4630.48 ± 8.760.01 ± 0.090.91 ± 1.8797.95 ± 1.87
Ly51242.73 ± 0.816.06 ± 0.363.05 ± 0.420.75 ± 0.0959.52 ± 1.280.79 ± 0.2125.76 ± 2.470.00 ± 0.041.26 ± 1.7399.92 ± 0.54
Figure 8.

 Chemical composition of sediment-dispersed extraterrestrial chromite (EC) and other chrome spinel (OC) grains from this study (samples Ly3 and Ly4 exempt), plotted together with average chromite compositions of recent ordinary chondrites (most of which represent observed falls; data from Wlotzka 2005). One unusually FeO-rich OC outlier from sample Ly1 is consistently omitted. See also Table 2.

Table 3.   All-sample average chemical composition (compound wt%± 1 s.d.) of extraterrestrial chromite (EC) grains from this study (samples Ly3 and Ly4 exempt), compared with average compositions determined in previous studies of relict EC and chromite from recent ordinary chondrites (most of which represent observed falls). See also Fig. 9.
Grain source
  1. a Not detected.

315 sediment-dispersed EC grains
  This study
2.77 ± 1.086.04 ± 0.432.99 ± 0.430.74 ± 0.0959.36 ± 1.480.79 ± 0.2025.84 ± 2.300.05 ± 0.131.09 ± 1.37
Chromite from 13 recent H chondrites
  Wlotzka (2005)
3.40 ± 0.186.64 ± 0.411.96 ± 0.290.65 ± 0.0357.1 ± 1.10.88 ± 0.0728.9 ± 0.6n.d. a0.28 ± 0.14
Chromite from 6 recent L chondrites
  Wlotzka (2005)
2.52 ± 0.215.90 ± 0.192.67 ± 0.440.70 ± 0.0656.1 ± 0.80.63 ± 0.0830.9 ± 0.60n.d.0.34 ± 0.06
Chromite from 4 recent LL chondrites
  Wlotzka (2005)
1.85 ± 0.145.52 ± 0.173.40 ± 0.570.67 ± 0.155.8 ± 0.560.51 ± 0.0431.6 ± 0.62n.d.n.d.
39 sediment-dispersed EC grains
  Alwmark et al. (2010)
2.54 ± 0.226.08 ± 0.253.00 ± 0.260.73 ± 0.0556.71 ± 0.480.79 ± 0.1929.20 ± 0.53n.d.0.08 ± 0.21
291 sediment-dispersed EC grains
  Cronholm and Schmitz  (2010)
2.88 ± 0.886.00 ± 0.362.91 ± 0.400.71 ± 0.0857.84 ± 1.140.80 ± 0.2627.40 ± 1.84n.d.1.09 ± 1.12
73 sediment-dispersed EC grains
  Alwmark and Schmitz  (2007)
1.59 ± 1.545.74 ± 0.752.48 ± 0.380.72 ± 0.0557.87 ± 1.091.44 ± 0.4726.83 ± 1.840.05 ± 0.142.35 ± 2.10
274 sediment-dispersed EC grains
  Häggström and Schmitz  (2007)
2.69 ± 1.106.08 ± 0.732.95 ± 0.440.74 ± 0.0756.93 ± 1.290.87 ± 0.2528.74 ± 1.72n.d.0.71 ± 0.84
276 sediment-dispersed EC grains
  Schmitz and Häggström  (2006)
2.58 ± 0.796.07 ± 0.763.09 ± 0.330.75 ± 0.0757.61 ± 1.580.78 ± 0.2027.36 ± 2.63n.d.0.53 ± 0.50
594 EC grains from fossil meteorites
  Schmitz et al. (2001)
2.57 ± 0.835.53 ± 0.292.73 ± 0.400.73 ± 0.0357.60 ± 1.301.01 ± 0.3326.94 ± 3.89n.d.1.86 ± 2.43
Figure 9.

 All-sample average chemical composition of sediment-dispersed extraterrestrial chromite (EC) from this study (samples Ly3 and Ly4 exempt), plotted together with average chromite compositions of recent ordinary chondrites (most of which represent observed falls; data from Wlotzka 2005) and average compositions of relict EC from other studies. See also Table 3.

Averages from various studies of relict EC grains mostly cluster together tightly in Fig. 9; however, compared with average compositions of chromite from recent ordinary chondrites, a shift is consistently seen in plots involving FeO and ZnO. Due to diagenetic influence, these compounds display notable variation in concentration, and, in general, ZnO can be enriched at the expense of FeO in relict EC grains (cf. Schmitz et al. 2001). A “tail” of increasingly altered grains thus protrudes from the main cluster in the ZnO versus FeO plot in Fig. 8. A similar feature may be seen in the Cr2O3 versus MgO+FeO+ZnO plot. The most striking difference between this study and preceding studies is that the all-sample average Cr2O3 value is relatively high, whereas the average FeO value is relatively low. This indicates significant leaching of FeO throughout time, shifting balance toward a higher Cr2O3 concentration. All relict-EC averages lie close to the average composition of chromite from recent L chondrites in the Al2O3-MgO and TiO2-V2O3 plots.

Unlike other compounds, where present, NiO displays highly variable concentration within individual grains. In general, NiO-enrichment appears to be confined to the outermost parts of and cracks in grains, but exceptions occur. Most of the NiO-rich grains are notably vesicular and contain numerous cracks, and some display a sharply defined, but discontinuous, MgO-/FeO-/NiO-enriched rim.

Data from chemical analyses of OC grains are presented in Fig. 8 (full data set in online supplement). OC provenance is not investigated in great detail, but it may be noted that many OC grains have compositions that almost completely fulfill the criteria for EC grain identification, although have at least one compound (typically Al2O3 and/or TiO2) well outside of its typical ordinary chondritic range (see Fig. 8). These grains may be diagenetically altered EC grains, or, simply, other extraterrestrial chromium-bearing spinel grains (cf. Wlotzka 2005; Häggström and Schmitz 2007; Cronholm and Schmitz 2010). Regardless, as very few OC grains were found in this study, small changes in OC numbers have little effect on overall statistics.


The Source of the Sediment-Dispersed EC Grains

This study adds yet another locality to a growing list of sections with EC-enriched strata, showing that there is a globally recognizable dramatic enrichment of EC in sediments dated to the early-middle Darriwilian of the Middle Ordovician. In the Lynna River section, EC grain concentrations in our upper sample set are even a factor two to three higher than in the sections previously studied in Sweden and China. The overall good match between the average chemical composition of the EC grains from this study and those of similar studies, together with the stratigraphic/temporal co-occurrence of EC enrichment, indicates that the grains share a common source––the disruption of the L-chondrite parent body (e.g., Schmitz et al. 2001; Cronholm and Schmitz 2010). As in preceding studies, the all-sample average composition, e.g., for FeO and Cr2O3, of EC grains from this study does not fit entirely with the average chromite composition of either of the recent ordinary chondrite groups, reflecting minor diagenetic alteration of the grains. More stable oxides, however, such as TiO2 and V2O3, plot close to the average of chromite from recent L chondrites. Also, studies of oxygen isotopes (Heck et al. 2010) and the composition of olivine and pyroxene inclusions (Alwmark and Schmitz 2009) in EC grains have shown that most or all grains found in Sweden and China are L-chondritic. Coeval fossil meteorites from Sweden are also L-chondritic, as shown by chondrule size measurements (Bridges et al. 2007), and element chemistry (Schmitz et al. 2001), oxygen isotopes (Greenwood et al. 2007; Heck et al. 2010), and inclusion mineralogy of relict chromite in the meteorites (Alwmark and Schmitz 2009).

The contrastingly low concentration of EC grains in the lower sample set probably represents the background flux of EC grains, preceding the arrival of the first material from the parent-body disruption event. As noted above, the stratigraphic/temporal occurrence of chromite enrichment at Lynna River correlates well with that found at previously studied localities; however, the record of the initial influx of EC from the disruption event remains undocumented somewhere in the strata between our two sample sets. Further studies are planned to locate the exact stratigraphic level at which the EC enrichment ensues at Lynna River. The finding of this level, together with better-resolved conodont biostratigraphy, will enable a more detailed comparison of stratigraphic variations in EC—and other heavy mineral—concentration between Lynna River and other sections.

Chromite and Ilmenite Concentrations, and Their Relationship with the Depositional Environment

By using weaker HF in this study than in previous ones, it has been possible to compare bed-by-bed variations in EC with those in the common terrestrial heavy mineral ilmenite. This shows that there is some covariation between chromite and ilmenite, which is not surprising, as the minerals should behave similarly in sedimentary systems, given their similar density (chromite 4.8 g cm−3, ilmenite 4.7 g cm−3; Wenk and Bulakh 2004) and hardness (chromite 5.5, ilmenite 5–6; Wenk and Bulakh 2004.). Processes of winnowing and selective transport, entailing hydrodynamic concentration and sorting, are likely to produce patterns of covariation among grains with similar characteristics. On this note, the storm-dominated environment represented by the (tempestite) beds at Lynna River may be an important factor behind the unusually high EC concentrations in the upper sample set, as compared with previous studies at other localities. Studies have shown that significant natural sorting may occur during storms, producing more or less concentrated lag/placer deposits, which are most likely to be preserved and further concentrated throughout time if they are deposited below fair-weather wave base (Barrie et al. 1988). Tidal currents can enhance this process. Sturesson et al. (1999) noted a general increase in heavy mineral content in Middle Ordovician strata along a west–east transect of Baltoscandia, which may essentially reflect basinal differences in energy levels.

Although heightened compared with background levels, the EC influx during the time interval represented by the upper sample set is likely to have been rather stable. Thus, the successively increasing EC concentration indicates increasing condensation—i.e., time averaging—toward the top of the sampled strata. This is also indicated by a simultaneous (but not 1:1) increase in ilmenite concentration. The increasing condensation may best be explained by sea-level change and associated changes in sediment deposition. Assigning a direction to this sea-level change, based on heavy-mineral grain data alone, is difficult; however, lithologic change from dominantly gray to more reddish limestone, with increasingly less argillaceous content (cf. Korochantsev et al. 2009), indicates successive relative sea-level rise throughout most of the upper sample set. This interpretation corroborates the high-resolution sea-level curve presented by Rasmussen et al. (2009), which was based on brachiopod assemblages. It may be noted, however, that our grain data can also be interpreted as tracing successive sea-level fall, with increasing weathering and winnowing resulting in higher proportions of heavy minerals. The time averaging that accompanies condensation is likely to have significant impact on biodiversity data and generalized time estimates based on net deposition rate. The anomalous Ly4 sample shows the highest concentration of EC grains in the section (10.2 grains kg−1). Assuming that its lithology is not a result of diagenetic dissolution of carbonate, the sample appears to represent some significant change in the depositional environment. Superficial rock characteristics indicate that shallowing may have occurred between sample levels Ly3 and Ly4. This event level may correspond to the “Täljsten” in Sweden (see Eriksson et al. 2012). Notable presence of heavily rounded, ferruginous (terrestrial) grains in sample Ly4 indicate at least admixture of relatively high-energy sediments; however, this is not reflected in the rounding of chromite grains. Regardless, it may be an indication of winnowing and/or time averaging, probably related to sea-level change. The relatively low EC concentration in sample Ly3 probably reflects a temporarily heightened sedimentation rate, and perhaps some erosion occurred before deposition of the overlying strata (which may in part explain the contrasting features of sample Ly4). In summary, the high amounts of sediment-dispersed EC grains in the upper sample set of this study are best explained by the interplay between an enhanced influx of ordinary chondritic matter and terrestrial-marine processes, such as hydrodynamic sorting and variations in deposition rate.

The consistently low amounts (or absence) of OC grains, as compared with some studies in Sweden (see e.g., Schmitz and Häggström 2006), indicate considerable distance to sources of such grains. Similarly, in the Puxi River section in China, OC grains are very rare through the entire section, and the few grains encountered can be readily recognized by their more rounded appearance compared with the angular (and more abundant) EC grains (Cronholm and Schmitz 2010). The fact that general EC grain distribution patterns can be reproduced among three regions—Sweden, Russia, and China— whereas OC distributions differ, strengthens the interpretation that observed EC enrichments reflect a global rain of L-chondritic (micro-)meteorites (see also Heck et al. 2010; Meier et al. 2010). The intermittently high concentrations of OC grains in Sweden may reflect proximity to coeval subduction-related volcanism associated with the genesis of the Caledonides (Pedersen et al. 1992; see also Sturesson 2003).

NiO-Rich EC grains

Prior to this study, NiO-rich relict EC grains in normal marine strata had only been documented by Korochantsev et al. (2009), although Alwmark and Schmitz (2007) found NiO-containing chromite grains in impact resurge deposits from the mid-Ordovician Lockne crater. The NiO-rich grains in the Lynna River samples may originally have been situated in, or in contact with, the fusion crust of their host meteorite (micrometeorites are essentially all surface, i.e., fusion crust). Sharply defined metal-rich rims may accrete on chromite grains due to melting of surrounding meteorite matrix, as friction heat is produced during descent through Earth’s atmosphere, followed by rapid cooling (cf. Ramdohr 1967; Genge and Grady 1999; Parashar et al. 2010). Meteorite bulk composition is typically rich in Ni and Fe, and melted metals may have interacted with surfaces of chromite, which has higher melting temperatures than most other minerals in a meteorite. A connection to the shock stage of the host body/meteorite is also possible, as NiO-rich grains are typically notably fractured (cf. Alwmark et al. 2011). Alternatively, the grains may have become diagenetically altered. Compounds more typical to be enriched due to diagenetic processes (most notably, ZnO), however, are not particularly abundant in NiO-rich grains, as compared with other grains, and Ni should be rather stable in the environmental conditions present in the sampled carbonate rocks (see Fujiwara and Domae 2004). If NiO-enrichment in pristine EC grains is not a secondary feature that is (as of yet) unique to the Lynna River strata, the reason for a lack of such grains in other studies may be their use of harsher hydrofluoric-acid treatment (11 M; Korochantsev et al. [2009] only used HCl).


The upper sample set of this study, around the A. expansusA. raniceps trilobite (probably also L. variabilisY. crassus conodont) Zone boundary in the lower-middle Kundan, yielded remarkably high amounts of chromite with typical ordinary chondritic composition (average 7.6 EC grains kg−1), as of yet outnumbered only by impact-related deposits––in all but one sample, EC is the most common heavy mineral in the acid-insoluble residues. In contrast, the lower sample set, around the A. lepidurusA. expansus trilobite Zone boundary, corresponding to the Volkhov-Kunda boundary, does not record any increase in EC grains (average 0.05 grains kg−1). This may simply reflect that these strata were deposited slightly before the disruption of the L-chondrite parent body.

Consistently high and partly harmonizing chromite and ilmenite (sensu lato) numbers in the upper sample set probably reflect the combined effect of an enhanced influx of extraterrestrial matter and effective sediment winnowing in the relatively high-energy depositional environment represented by the Lynna River strata. The concentration of heavy minerals, in particular ilmenite, increases when ascending stratigraphically, probably tracing sea-level rise.

Yet again, it is shown that an enhanced concentration of EC grains may be found in lower-middle Darriwilian strata. This rather predictable EC presence corroborates the prevailing view that there was an increased influx of extraterrestrial matter, following the disruption of the L-chondrite parent body around 470 Ma. Future studies may determine at exactly which stratigraphic level, and in what manner, this EC enrichment initiates at Lynna River.


Acknowledgments— We thank Carl Alwmark, Mats Eriksson, Sanna Holm, Matthias Meier, My Riebe, and Rainer Wieler for valuable discussions, and Li Stenberg and Mario Tassinari for assistance in the field. The study was supported by grants to Birger Schmitz from the European Research Council and the Swedish Research Council. The reviewers, John Bridges and Alan Rubin, are thanked for their comments and suggestions, which contributed to improving the overall quality of this article.

Editorial Handling— Dr. John Spray