A new type of solar-system material recovered from Ordovician marine limestone

From mid-Ordovician ∼470 Myr-old limestone >100 fossil L-chondritic meteorites have been recovered, representing the markedly enhanced flux of meteorites to Earth following the breakup of the L-chondrite parent body. Recently one anomalous meteorite, Österplana 065 (Öst 65), was found in the same beds that yield L chondrites. The cosmic-ray exposure age of Öst 65 shows that it may be a fragment of the impactor that broke up the L-chondrite parent body. Here we show that in a chromium versus oxygen-isotope plot Öst 65 falls outside all fields encompassing the known meteorite types. This may be the first documented example of an ‘extinct' meteorite, that is, a meteorite type that does not fall on Earth today because its parent body has been consumed by collisions. The meteorites found on Earth today apparently do not give a full representation of the kind of bodies in the asteroid belt ∼500 Myr ago.

T he ordinary chondrites are the dominant type of meteorite falling on Earth today, representing about 85% of the total flux. They are divided in three subgroups, H (42% of ordinary chondrite falls), L (47%) and LL (11%), depending on their iron and metal content. In the 1960s, it was realized that the L chondrites show a major peak in K-Ar degassing ages at B500 Myr ago 1,2 , an age recently better constrained to B470 Myr ago 3 . These meteorites apparently underwent a massive collisional event at this time, involving the breakup of their parent body. In contrast, the typical degassing age of most meteorite types goes back to the young solar system, that is, 43.5 Gyr ago. In marine limestone that formed on Earth around 470 Myr ago, there is an evidence for an at least two orders of magnitude increase in the flux of L-chondritic meteorites and micrometeorites (for a review, see ref. 4). The prime evidence is the more than 100 fossil L chondrites recovered during the industrial quarrying of Ordovician limestone in southern Sweden 4,5 . The meteorites are almost completely replaced by secondary minerals, but spinel grains, and often also petrographic textures, are preserved. Mineralogical, chemical, isotopic and petrographic studies have shown that mid-Ordovician fossil extraterrestrial material is indeed of L-chondritic origin 4 . Recently, one 8-cmlarge meteorite was found, containing spinel grains very different from all the other meteorites 6 . On the basis of 21 Ne cosmic-ray exposure ages of its spinel grains this meteorite, Ö sterplana 065, previously referred to as the 'mysterious object', appears to have been liberated as a meteoroid at the same time as the other fossil meteorites [6][7][8] . It may thus represent a piece of the impactor that broke up the L-chondrite parent body 6 . The breakup of this body and the following enhanced flux of extraterrestrial material, including km-sized asteroids, coincide with marked evolutionary changes in Earth's invertebrate life, the so-called Great Ordovician Biodiversification Event 9 . Unravelling why the L-chondrite parent body broke up at this time, may lead to a better understanding of possible large-scale astronomical perturbations affecting both Earth and the solar system 4 .
Ö sterplana 065 was preliminarily classified as 'winonaite-like', based on oxygen isotopic and elemental composition of its spinels and by comparison with a winonaite clast in the Villabeto de la Peña ordinary (L6) chondrite 6,10 . In order to further clarify the origin of Ö st 65 here we apply chromium isotopic analyses of whole-rock samples of Ö st 65 plus three fossil L chondrites (Ö sterplana 018, 029 and 032, or informal names Ark 018, Gol 001 and Bot 003, respectively), as well as clast and host material from Villabeto de la Peña. We also perform Cr-isotopic analyses of separated chrome-spinel grains from two recent meteorites, the winonaite NWA 725 and the L6 ordinary chondrite Lundsgard. In addition, we present refined oxygen three-isotopic analyses for the chrome spinels of Ö st 65, and new data on the appearances and distribution of relict minerals in Ö st 65. We show that Ö st 65 has no documented analogue among the known meteorites that have fallen on Earth in recent time. The 54 Cr values of Ö st 65 are similar to those of ordinary chondrites, whereas oxygen isotopes show its affinity with some rare primitive achondrites.

Results
Isotopic results. The fossil meteorite Ö st 65 is shown in Fig. 1, and its stratigraphic position in the Thorsberg quarry, together with previously established 21 Ne cosmic-ray exposure ages 6,7 , is presented in Fig. 2. Our new e 54 Cr and D 17 O data are compared to similar data for many of the major groups of meteorites known today in Fig. 3. For the detailed results of our isotopic analyses, see Tables 1 and 2. The spinel grains from Lundsgard and NWA 725 yield e 54 Cr values that are the same as previously analyzed L6 chondrites and winonaites, respectively. The samples of the fossil L chondrites yield e 54 Cr results in agreement with an ordinary chondrite origin, that is, in the range À 0.34 to À 0.24. This indicates that whole-rock analyses of fossil meteorites provide reliable Cr-isotopic results. This is in line with previous elemental analyses of whole-rock fossil L chondrites showing that they have retained most of their original chromium 11 . The Villabeto de la Peña host and clast give e 54 Cr values confirming the previous classifications based on oxygen isotopes alone 10 . The two whole-rock samples of Ö st 65 yield e 54 Cr values of À 0.26 to À 0.23, which are within the uncertainties of previously measured ordinary chondrites 12 . The new oxygen-isotope analyses confirm previous D 17 O values for Ö st 65 spinels just below the terrestrial fractionation line. The data plot with the winonaites, acapulcoites and lodronites, clearly separated from the chromites of the typical L-chondrite fossil meteorites. Although the oxygen isotopes support a relationship to the winonaites (as suggested in ref. 6), acapulcoites or lodronites, the e 54 Cr results for Ö st 65 rule out such a relationship. Our combined e 54 Cr and D 17 O analyses of Ö st 65 suggest that it originated from an asteroid of a kind not represented in our collections of recent meteorites (Fig. 3).
Mineralogy and petrology. Mineralogical and petrological data support the conclusion that Ö st 65 represents a so far unknown type of meteorite. Our data rule out that Ö st 65 is a breccia with clasts of, for example, winonaite and ordinary chondrite origin. The relict texture is unbrecciated and no relict chondrule texture is observed. We have dissolved several fragments from different parts of Ö st 65, and have not been able to recover one single grain of the type of chromite that is so common in equilibrated ordinary chondrites. Such chromite (4ca. 30 mm) has a narrow element compositional range, making it easy to identify 4 . Our studies of polished sections show that the type of chrome-spinel grains we used for oxygen isotopic analyses are distributed throughout the meteorite. There are no MgAl spinels (45 mm) in Ö st 65. The only type of inclusion found in the chrome spinels from Ö st 65 has a Ti-and Cr-rich composition (similar to olkhonskite). These inclusions are quite common in the spinels of Ö st 65 ( Fig. 4a), but no silicate inclusions have been found although searched extensively. The chrome-spinel grains are sometimes associated with 10-30 mm large rutile grains (Fig. 4b). To our knowledge there is no known meteorite with the chrome-spinel-rutile assemblage that we observe in Ö st 65. The majority of the chrome-spinel grains in Ö st 65 show abundant planar deformation features (Fig. 4c), something not observed in the many fossil L chondrites. The planar features appear similar to shock lamellae in chromites from shock melt veins in recent meteorites 13 . The Ö st 65 spinels are also heavily fractured (Fig. 4a,c), which provides independent evidence for severe shock. Chromites of recently fallen L chondrites show increasing levels of fracturing with increasing shock stage 14 .

Discussion
The meteorite Ö st 65 accompanies 4100 L chondrites recovered in the sediments of the Swedish quarry. Although single random meteorites are possible, one has to consider that Ö st 65 represents on the order of one per cent of the meteorites that have been found on the mid-Ordovician sea floor. This indicates that Ö st 65 may represent one of the dominant types of meteorites arriving on Earth 470 Myr ago. Although we cannot rule out that a meteorite similar to Ö st 65 may eventually be found among meteorites falling on Earth today, we can be quite certain that the Ö st 65 type does not represent one per cent of the 52,600 classified recent meteorites. At the most there could be a few overlooked specimens.
The discovery in the Ordovician record of a meteorite type that is apparently not falling on the Earth today has important implications. For example, it is possible that the asteroid that produced Ö st 65 no longer exists, and there is no source for such meteorites today. The asteroid belt has been evolving through collisions over the history of the solar system, and many of the original asteroids have undoubtedly been destroyed. The record of fossil meteorites on Earth (or on other planetary bodies) provides the only evidence for their former existence and the only way to investigate the collisional evolution of the solar system. The diversity of meteorite parent bodies even relatively late in solar-system history, B500 Myr ago, may have been significantly greater than today. We know, for example, that remnants of Earth building blocks are not present in our meteorite collections. Such meteorites are extinct now but clearly built the Earth in the distant past.
The cosmic-ray exposure age of Ö st 65 is B1 Myr and is within the range of those of the fossil L chondrites (Fig. 2) [6][7][8] . This associates Ö st 65 with the breakup of the L-chondrite parent body. The high level of shock metamorphism in the chromespinel grains (Fig. 4c)   Cosmic-ray exposure ages for some of the mid-Ordovician fossil meteorites. Meteorite cosmic-ray exposure ages increase with stratigraphic height, in accordance with sedimentation rates as established from conventional biochronology 7,8 . The data indicate that all the fossil meteorites, including Ö st 65, originate from a breakup event that took place about 200 Ka before the lowermost beds in the quarry formed 6 . The details and the analytical background of the 21 Ne exposure ages are discussed in ref. 6. In the figure the Ö sterplana 065 meteorite is represented by its 'field' name, Gla3 003, at the top of the section studied. Our fossil meteorites obtain official names, approved by the Meteoritical Society, following the convention of naming meteorites after the locality at which they have been found, that is, in this case Ö sterplana. But in order to emphasize that meteorites in different beds fell at different times they also receive an informal 'field' name. For Ö st 65 the name Gla3 003, implies that it is the third meteorite found in the bed Glaskarten 3.
asteroid while only disrupting the L-chondrite parent body into several smaller bodies and debris. These smaller bodies continue to feed meteorites, but there is nothing left to provide Ö st 65 meteorites to Earth today. We do not know whether the Ö st 65 parent body was one of the original asteroids or already a fragment of a larger object. Iron meteorites come from parent bodies that experienced melting and differentiation that separated the Fe-Ni metal from the silicate components. Trace-element data indicate that iron meteorites may originate from up to 60 different parent bodies 15 . With a few potential exceptions, we have not been able to identify the silicate fractions that accompanied the Fe-Ni metal in the original parent bodies. The silicates were probably stripped from tougher iron cores of the asteroids early in solar-system history. The iron cores of these parent bodies were much more resistant to collisional degradation, whereas the outer silicate parts have been consumed over time in this process 16 . Because Ö st 65 does not exhibit chondrite texture and has other characteristics that do not match most chondrites, the Ö st 65 parent body might represent some of this missing silicate. Winonaite meteorites have very similar oxygen-isotope compositions to the silicate inclusions in IAB and IIICD iron meteorites, implying that they may have originated on the same parent asteroid 15,17 . The 'forensic' power of paired D 17 O-e 54 Cr isotope systematics 18,19 could be applied to silicate and oxide inclusions of many types of iron meteorites in order to identify their genetic link with other planetary materials. Studies of fossil meteorites in the Earth sedimentary record may identify other classes of 'extinct' meteorites, some of which could be linked to known classes of iron meteorites.
The Ö st 65 meteorite is significant because it demonstrates that B500 Myr ago we may have had different meteorites falling to Earth than what we see today. Some may be extinct, whereas others, such as the L chondrites, still fall on Earth. Apparently there is potential to reconstruct important aspects of solar-system history by looking down in Earth's sediments, in addition to looking up at the skies 4 .    After dissolution, the sample solutions were dried down and treated with concentrated HNO 3 and 6 N HCl to eliminate fluorides generated during the dissolution processes. Chromium was removed from the sample matrix utilizing a 3-column chemistry procedure described in ref. 20. High-precision isotope ratios were measured in the purified Cr fractions using a Thermo Triton Plus thermal ionization mass spectrometer at the University of California at Davis. The Cr fractions were loaded onto outgassed tungsten filaments after being mixed with an Al-boric acid-silica gel activator solution. Each filament was loaded with 3 mg of Cr and four filaments were loaded for each sample for a total load of 12 mg. Each four filament set was bracketed with filaments loaded with the NIST SRM 979 Cr standard. The signal intensity for each filament was set to 10 V ( ± 15%) for 52 Cr with 10 11 O resistors. Each filament analysis consisted of 1,200 ratios with 8 s integrations times. A gain calibration was made at the beginning of each filament, and the baseline was measured and the amplifiers were rotated between each block of 25 ratios. The mass fractionation was corrected using an exponential law and a 50 Cr/ 52 Cr ratio of 0.051859 (ref. 21). The 53 Cr/ 52 Cr and 54 Cr/ 52 Cr ratios are expressed in e-notation (that is, parts per 10,000 deviation from the NIST SRM 979 Cr standard).

Methods
Oxygen isotopic analysis. Individual chromite grains, 50-100 mm in diameter, separated from Ö st 65, were mounted in resin in 0.25-inch-diameter stainless steel bullets. Somewhat larger grains of Stillwater chromite standard were mounted in similar bullets. All grains were ground and polished to give a flat cross section at least 50 mm across. The grains were imaged with secondary electrons and measurement spots were selected to avoid cracks, surface imperfections, inclusions and so on. The spots were marked using the scanning electron microscope so that the spots could be seen by scanning ion imaging to accurately position the beam for measurement.
Oxygen-isotope compositions were measured using the University of Hawai'i Cameca ims-1280 ion microprobe. The detailed methodology is provided in ref. 22. Here, we provide details specific for these measurements. A 1 nA Cs þ primary ion beam, focused to B10 mm, with a total impact energy of 20 keV was used. Spots were presputtered for 60 s using a 5-mm raster, after which the raster was reduced to 3 mm for measurement. Masses 16 O -, 17 Oand 18 Owere measured simultaneously in multicollection mode. 16 Oand 18 Owere measured by multicollector Faraday cups with low-mass resolving power (MRP B2,000), while 17 Owas measured using the axial monocollector electron multiplier with MRP B5,600, sufficient to separate the 16 OHinterference. The relatively high 17 Ocount rate produces a loss in gain due to aging of the first dynode of the electron multipler. To minimize this effect, the 17 Osignal was measured for only 4 s in each cycle, after which the beam was deflected into a monocollector Faraday cup for 10 s (procedure originally proposed by Kita et al. 23 ). Signals for 16  Measured data were corrected for detector background and deadtime (electron multiplier). The background corrections for the Faraday cups were made using a smoothed fit to the backgrounds measured during presputtering. A tail correction was made for the 16 OHinterference. The contribution of the 16 OHtail at mass 17 O was estimated from measurement of the low-mass tail of the 17 O peak to be B6 Â 10 -6 of the 16 OHinterference. The height of the 16 OHinterference was measured at the end of each measurement. Even with our protocol to minimize changes in the gain of the electron multiplier, the gain did change with time. We checked the pulse-height distribution at the beginning of each day and set the gain to our 'standard' condition. We did not check the gain again during the day, as we have learned that the act of checking the multiplier gain introduces a transient instability in the multiplier gain that compromises the data. The gain of the monocollector electron multiplier during the day was monitored by measuring the Stillwater chromite standard for five-six measurements at the beginning of each day, after five-six measurements of the unknown, and at the end of the day. The Stillwater measurements permitted us to model and account for the gain drift in the multiplier during each measurement day.
The uncertainties for the individual measurement data reported in Table 2 include the counting-statistical uncertainties from the individual measurements, the standard deviation of the standard measurements (including uncertainties in the drift correction in the case of 17