The Manson Impact Structure: 40Ar/39Ar Age and Its Distal Impact Ejecta in the Pierre Shale in Southeastern South Dakota

The 40Ar/39Ar ages of a sanidine clast from a melt-matrix breccia of the Manson, Iowa, impact structure (MIS) indicate that the MIS formed 73.8 � 0.3 million years ago (Ma) and is not coincident with the Cretaceous-Tertiary boundary (64.43 � 0.05 Ma). The MIS sanidine is 9 million years older than 40Ar/39Ar age spectra of MIS shock-metamorphosed microcline and melt-matrix breccia interpreted earlier to be 64 to 65 Ma. Grains of shock-metamorphosed quartz, feldspar, and zircon were found in the Crow Creek Member (upper Campanian) at a biostratigraphic level constrained by radiometric ages in the Pierre Shale of South Dakota that are consistent with the 40Ar/39Ar age of 73.8 � 0.3 Ma for MIS reported herein.

The results from 330S indicate not only that crustal thickness variations can fully account for the observed MBA, but also that lateral variations of density within the crust are significant enough that they must be considered when gravity data are interpreted. It follows from these results that other bull's-eye MBAs inferred to exist along many ridge segments on the MAR may also be formed by along-axis crustal thickness variations. Such anomalies are also commonly associated with a shoaling of the axial valley similar to that observed in the 330S area (4,29). The regularity with which these characteristics are being observed suggests that they are a primary feature of the spreading mechanism operating along the slow-spreading MAR.

29.
30. R. Batiza, W. G. Melson, T. O'Hearn, Nature 335, 428 (1988). R. S. Detrick, J. C. Mutter, P. Buhl, l. l. Kim, ibid. 347, 61 (1990). E. M. Parmentier and J. Phipps Morgan, ibid. 348, 325 (1990). J. Lin and J. Phipps Morgan, J. Geophys. Res. Lett. 19, 13 (1992). K. C. Macdonald et al., Nature 335, 217 (1988). P. J. Michael et al., in preparation. J. Phipps Morgan and Y. J. Chen, Nature, in press. S. Carbotte, S. M. Welch, K. C. Macdonald, Mar. Geophys. Res. 13, 51 (1991). J.-C. Semp&r4, G. M. Purdy, H. Schouten, Nature 344, 427 (1990). D. W. Caress, M. S. Burnett, J. A. Orcutt, J. Geophys. Res. 97, 9243 (1992). 31. P. B. Stark and R. L. Parker, Geophys. J. R. Astron. Soc. 89, 997 (1987). 32. The seismic refraction line at 330S was part of the Plume experiment, a multidisciplinary study that included bathymetry, seismics, gravity, magnetics, and dredging. We thank the officers, crew, and scientific complement of the R.V. Thomas Washington for their help in collecting these data. We also thank G. A. Neumann for providing the bathymetry and gravity data before its publication, D. Caress for his two-dimensional ray tracing program, P. An iridium anomaly (1), shock-metamorphosed minerals (2)(3)(4), and relic tektites (5) in Cretaceous-Tertiary (K-T) boundary rocks provide evidence that at least one large extraterrestrial object struck the Earth 64.43 + 0.05 Ma (6), probably in continental rocks in or near North America (3,7). The rocks at two impact structures, Chicxulub on the Yucatan Peninsula and Manson in northwest Iowa, have been examined for evidence that might link them to the K-T impact. The Chicxulub structure (-200 km in diameter) has emerged as the leading candidate (8,9), but the Manson structure (35 km in diameter) has also been proposed as a K-T boundary impact site because of its proximity to K-T boundary rocks containing large and abundant shocked minerals (3,7,10 29 OCTOBER 1993 in the MIS has normal remanent magnetization (12) and, if taken at face value, excludes the possibility that this structure formed at the K-T boundary, which is in reversely magnetized rocks in the upper part of magnetozone 29R (13). In addition, U-Pb ages of shock-metamorphosed zircons from the upper of two thin K-T boundary claystone beds of western North America indicate that their provenance was not Precambrian crystalline basement rocks in the Manson area or sedimentary rocks derived from such rocks (4). In 1992, 12 holes were drilled to explore the MIS and to investigate further the isotopic age of the impact (14). The M-1 hole, on the flank of the central peak, penetrated a high-temperature, melt-matrix breccia layer 40 m thick (15). Examination of core from this rock revealed the presence of rare, chalky-white, microcrystalline feldspar ( Scanning electron microscope analysis showed that these clasts, which differ from the more common shocked and partially melted potassium feldspar (microcline) clasts, generally consist of spherulitic, radially bladed sanidine [K20 (11.4 + 0.7% w/w); 0r68Ab31AnlJ and trace amounts of 729 quartz. The texture and composition of the sanidine clasts imply that they crystallized from a liquid derived from melted, older microcline clasts. If so, the sanidine could yield an isotopic age reflecting its time of crystallization and, thus, the time of the Manson impact.
Although this technique has important advantages relative to the conventional K-Ar method, the ages of unknown samples are relative to ages of neutron fluence-monitor minerals. We calculated or recalculated all relevant ages using the age of a secondary, intralaboratory neutron fluence-monitor mineral [sanidine from the Taylor Creek Rhyolite (Oligocene) of New Mexicol. This sanidine has a 40Ar/39Ar age of 27.92 Ma, relative to a K-Ar age of 162.9 ± 0.8 Ma for a primary fluence-monitor standard SB-3 biotite or 513.9 Ma for an international homblende standard MMhb-1 (20). Ten individual fragments of a Manson M-1 sanidine clast were dated (Table 1), and the weighted mean of 12 laser-fusion 40Ar/ 39Ar ages is 73.83 ± 0.31 Ma (21). This result places an upper limit on the age of the sanidine because a small but significant amount of 40Ar might have been trapped in its lattice as it cooled. To evaluate this possibility, we plotted the analytical data on an inverse correlation diagram ( Fig. 2), resulting in an isochron age for the sanidine of 73.52 ± 0.79 Ma. This age is slightly younger but not statistically different from the weighted mean age and establishes a lower limit for the age of the sanidine and the MIS. The trapped argon component in the sanidine has a 40Ar to 36Ar ratio of 308 ± 31, not statistically different from the ratio of these isotopes (295.5) in atmospheric argon. Given the relatively large amounts of argon obtained for 10 of the individual analyses (Table 1), the marginally higher 40Ar to 36Ar ratio of 308 would reduce the age only slightly from 73.83 0.31 Ma   (22). A physical record of the impact, including tsunami deposits, shock-metamorphosed mineral grains, and altered tektites, might be found in the Pierre Shale (Upper Cretaceous) at a stratigraphic horizon commensurate with our proposed age for the MIS.
We focused our attention on outcrops of the Pierre Shale along the Missouri River in southeastern South Dakota. In this area, the lower part of the Pierre consists typically of 55 m of shale and some marl and includes, in ascending order, the Sharon Springs, Gregory, Crow Creek, DeGrey, and Verendrye members (23). Near Chamberlain, the lower part of the Gregory Member contains the ammonites Baculites gilberti and B. gregoryensis. The ammonites B. scotti and Menuites occur in the upper part of the Gregory within 4 m of the overlying Crow Creek Member. These ammonites are indicative of late middle Campanian age (Fig. 3). The species B. rugosus (late form having a weakly ribbed venter), which is found in black, manganese-rich concretions of the overlying DeGrey Member, is restricted to the zone of Didymoceras cheyennense of early late Campanian age. The Verendrye Member contains B. cuneatus and B. reesidei of late Campanian age. Note that the y axis intercept on the x axis is not zero. Two samples with large error bars (1 u) are of reheated material that yielded only small amounts of argon ( Table 1). The mean square weighted deviates = 1.27. The dashedline inset is an enlargement (x2) of the indicated data.
Thus, the intervening Crow Creek Member, which apparently lacks ammonites and other macrofossils, may be of early late Campanian age and be in the highest part of the Exiteloceras jenneyi zone or in the lowest part of the D. cheyennense zone. Obradovich (24) obtained sanidine 40Ar/ 39Ar ages of 74.83 0.72 Ma, 73.71 ± 0.45 Ma, and 72.32 0.39 Ma for bentonite beds in the B. scotti, E. jenneyi, and B. compressus zones, respectively. The Crow Creek Member, which occurs above the zone of B. scotti and below the zone of B. compressus, must therefore be bracketed by these ages (Fig. 3) and is consistent with our age of 73.8 ± 0.3 Ma for the MIS. Compilations of the paleomagnetic time scale (25) and the biostratigraphic consid- Table 1. Laser-fusion 40Ar/39Ar ages of a sanidine clast from the 11 (24).
(0 en 0 erations given above suggest that the Manson impact occurred during a normal paleomagnetic chron, possibly in chron 33N or in a brief normal in 32R (32R.AN). The Crow Creek Member, which consists of 2 to 3 m of marl, has received considerable attention because of an anomalous lithologic feature (23). This member contains at its base a locally cross-bedded siltstone 15 to 20 cm thick or a concentration of sand grains and rip-up shale clasts as large as 4.5 cm. The sand grains and shale clasts increase in size and abundance from west to east in southeastern South Dakota (23,26). The provenance of the sand grains, some as large as 2.8 mm, in an otherwise shale-dominated sequence has puzzled geologists for more than 40 years. The abrupt appearance of sand granules supports the argument for a sudden change in depositional conditions or source. Just as interesting is the fact that the Crow Creek Member rests disconformably on successively older beds of the Pierre Shale from northwest (Chamberlain) to southeast (Yankton) along the Missouri River toward the MIS. A few kilometers west of Yankton, the Crow Creek overlies about 2 to 3 m of organic-rich shale of the Sharon Springs Member (26).
We searched for mineralogic evidence that is diagnostic of an impact in the basal part of the Crow Creek Member at three places in southeastern South Dakota: a site in Black Dog Township, Lyman County, about 20 km

Age
Member of the same structure obtained by Kunk and Ammon__e zone (Ma) Pierre shale colleagues (1 1) because the Crow Creek ___________ Member of the Pierre Shale contains shock-C__e --Ve-------metamorphosed minerals that were probably ocsanqvow 7232±039 De~ry derived from the MIS. This formation is 7.1n+± .4.s sandwiched between isotopically dated am-_ S/////////~~~~~~monite zones (24), and the inferred numerical w41co 74~~n age for the Crow Creek Member is in accord nu.u with our proposed isotopic age for the MIS. southwest of Chamberlain, and two sites (27) in Yankton County (House of Mary Shrine and an abandoned limestone quarry, 11 and 6 km west of Yankton, respectively). The acidinsoluble residue of the samples from these sites consists chiefly of quartz and minor feldspar and mica. A few percent of the quartz and feldspar (microcline and plagioclase) grains contain multiple intersecting sets of planar lamellae (Fig. 4) identical to those in shocked mineral grains from rocks at known impact structures. Many ofthe shocked quartz grains also exhibit pronounced shock-mosaic texture. Shocked lithic fragments composed of quartz and feldspar also were identified. The presence ofshock lamellae in well-rounded quartz grains suggests that these grains were derived from a sedimentary target rock. A few rounded, shock-metamorphosed zircon crystals were recovered from heavy-mineral concentrates of the insoluble residue of the marl.
The largest shocked mineral grains in the basal Crow Creek marl at the two Yankton County sites were 2.3 mm and 1.7 mm. These sites are about 250 km from the MIS. In contrast, the largest shocked mineral grains from the basal sandstone of the Crow Creek in Lyman County, 200 km west of Yankton County, were only 0.6 mm. The unusually large size of the shocked mineral grains in Yankton County implies that they came from a nearby source, such as the MIS.
Further study of stratigraphic relations in shale directly below the layer of Manson impact ejecta may provide a key to understanding (i) the puzzling absence of three ammonite zones above the zone of E. jenneyi at the Red Bird section in eastern Wyoming (22) and (ii) the origin of regional unconformities above the zone of E. jenneyi and below the Teapot Sandstone Member of the Mesaverde Formation and Pine Ridge Sandstone of the Mesaverde Group in central Wyoming (28). An impact-triggered tsunami could account, in part, for some of these anomalous stratigraphic and biostratigraphic relations.
Our laser-fusion 40Ar/39Ar apparent age of 73.8 + 0.3 Ma for a sanidine clast from a melt layer of the MIS is more rational than incremental-heating 40Ar/39Ar ages of 64 to 65 Ma for shock-metamorphosed microcline from SCIENCE * VOL. 262 * 29 OCTOBER 1993 scribed in (16)(17)(18)(19), including the reactor fluence characteristics, irradiation scheme, and methods for the measurement of corrections for interfering argon isotopes. Fragments (0.2 to 0.4 mm) of sanidine from a clast at the 114.2-m level of the M-1 core were sealed in aluminum foil cups, and the flattened, pancake-like packet was sandwiched between similar packets of the fluence-monitor mineral, sanidine of the Taylor Creek Rhyolite. The packets were arranged in a vertical stack and loaded into a quartz tube that was irradiated (2.4 x 1018 neutrons cm-3 S1-) for  with a 5-W, continuous argon-ion laser. Only one fragment of the sanidine melted to a clear glass bead (93Z0406 , Table 1). Nevertheless, radiogenic argon yieldswere appreciable (1.23 x 10-13 to 5.13 X 10-13 mol). Most fragments were heated with the laser for less than 1 s, but a few were heated for 5 min. Reheating of the sanidine (93Z0408 and 93Z041 1, Table 1) showed that most of the argon was expelled during the initial 1-s heating episode. The neutron fluence within the radiation package was measured by the analysis of five to seven lots of sanidine crystals of each fluence-monitor packet. Each age determination of the sanidine from the M-1 core hole was made by the melting of single fragments (-0.3 mg). For additional internal calibration, 12 K-T boundary tektites from Beloc, Haiti, were irradiated in the same package, and they yielded a weighted mean age of 64.6 ± 0. The Talladega belt is the westernmost crystalline thrust sheet in the southernmost exposed Appalachians and lies between the foreland fold-thrust belt to the northwest and the eastern Blue Ridge to the southeast (Fig. 1 (2)(3)(4), and the controversial Erin Slate that has been argued as either Early Devonian (2) or "probably Pennsylvanian" (5). Metasedimentary rocks in the Talladega belt contain no evidence of polymetamorphism (6). The time of dynamothermal metamorphism has been interpreted to be R Devonian, coeval with Acadian orogenesis, on the basis of K-Ar whole rock ages on slate and the presence of an Early Devonian megafaunal assemblage from chert Jemison Chert) near the stratigraphic top of the sequence (6, 7). An early 20th-century report of "probable" Pennsylvanian plant fossils from the Erin Slate (5), which overlies the Jemison, has been questioned because of the inability of subsequent investigators to replicate previous material. This inconsistency has led to the conclusion that these Carboniferous fossils are exotic (2). In its type area, the Erin Slate is a variably deformed black slate that stratigraphically overlies the Cheaha Quartzite across a gradational contact and underlies the Chulafinnee Schist (Fig. 2). The upper contact with the Chulafinnee is interpreted as a thrust fault (8). The interpretation of the Erin-Chulafinnee contact as gradational and conformable; the discovery of the fossil Veryhachium, a long-ranging (Silurian to Carboniferous) marine acritarch from the Emn Slate; and the correlation of the Chulafinnee with the Jemison Chert (9) have  (8) and (29)