Are Diamictites Impact Ejecta?—No Supporting Evidence from SouthAfrican Dwyka Group Diamictite1

Wolf Uwe Reimold, Victor von Brunn,2 and Christian Koeberl3

Department of Geology, University of the Witwatersrand, Private Bag 3, P.O. Wits 2050,Johannesburg, South Africa

A B S T R A C T

Diamictites, especially those deposited before the break-up of Gondwana in the Late Carboniferous and Permian, haverecently been suggested to represent ejecta deposits from large comet or meteorite impact events. This is in contrastto the commonly held interpretation that these rocks represent glaciomarine sedimentary deposits. To test this contro-versial hypothesis, we carried out a detailed petrographical study of over 75,000 mineral and rock clasts from a largenumber of Dwyka Group diamictite samples from localities covering an extensive part of the Karoo basin in southernAfrica. No definitive evidence of impact-diagnostic shock metamorphic deformation in mineral or lithic clasts fromany of these samples was detected. We conclude, therefore, that to date no unequivocal evidence for an impact originof these diamictites in the South African stratigraphic record has been documented. What is more, the general hypothe-sis that some diamictites in the stratigraphic record could represent impact ejecta is not supported by first-orderobservations of bona fide shock (impact) related phenomena in such rocks.

Introduction

The wider geological community now appears to global Cretaceous-Tertiary (K-T) boundary layer,formed by the 65 Ma old Chicxulub impact eventhave accepted that the impact of extraterrestrial

bodies has been of fundamental importance not (e.g., Kring 1993 and Koeberl 1996 for reviews); the35 Ma Late Eocene ejecta layer, which may be re-only in the evolution of the Solar System, but also

on Earth (e.g., Taylor 1992; Melosh 1989; papers in lated to the 100 km Popigai impact structure inSiberia or to the 90 km Chesapeake Bay impactSharpton and Ward 1990; Silver and Schultz 1982;

and Dressler et al. 1994). About 150 impact struc- structure in North America (Clymer et al. 1996;Langenhorst 1996; Koeberl et al. 1996a); the Latetures are currently known on Earth, and about

three to five additional impact structures are dis- Devonian (Frasnian/Famennian boundary) impactlayer (Claeys et al. 1992; Claeys and Casier 1994);covered or confirmed every year. Most of the

known terrestrial impact structures are either cov- and the Acraman distal ejecta layer in Australia(e.g., Gostin et al. 1986; Williams 1986).ered by post-impact rocks or significantly eroded

and degraded by continuous geological activity and It has also been debated whether several distinctspherule-rich Archean horizons could representsurface processes. Erosional processes are also the

reason why remnants of ejecta blankets have only distal ejecta of large ancient impact events (Loweand Byerly 1986; Glikson 1993) or whether they arebeen preserved around some relatively young im-

pact craters, such as the 15 Ma Ries Crater (Grieve the result of endogenic (volcanogenic or sedimen-tary) processes (Koeberl et al. 1993; Koeberl andet al. 1995) in Germany.

Even fewer layers of distal impact ejecta have Reimold 1995a). Examples are the spherule layersat the stratigraphic interval straddling the Fig Tree-been identified; examples are the well-knownOnverwacht Group contact in the Barberton Moun-tain Land of South Africa (e.g., Lowe and Byerly

1 Manuscript received Dec. 12, 1996; accepted April 27, 1997. 1986; Byerly and Lowe 1994) or from the Hamersley2 Department of Geology, University of Natal, Private Bag Basin of Western Australia (Hassler and SimonsonX01, Scottsville 3209, South Africa.

1995; Simonson and Davies 1996). However, in nei-3 Institute of Geochemistry, University of Vienna, Althan-strasse 14, A-1090 Vienna, Austria. ther case has it been possible to detect first-order

[The Journal of Geology, 1997, volume 105, p. 517–530] 1997 by The University of Chicago. All rights reserved. 0022-1376/97/10505-0008$01.00

517

518 W O L F U W E R E I M O L D E T A L .

Figure 1. Main exposures of the Dwyka Group in southern Africa (after Visser et al. 1990) and sampling sites forthis study. 1: Keetmanshoop; 2: Laingsburg; 3: Matjiesfontein; 4: Umlaas Road Quarry, east of Pietermaritzburg;5: borehole core from John Ross Bridge; 6: Babanango; 7: drilling site north of Vryheid; 8: Mpongoza stream (SwartfoloziValley); 9: Mbizana stream (Itala Valley); 10: Majuba Colliery, Amersfoort; 11: Vereeniging; 12: Matla; and 13: Middel-burg coal mines. The sample of Mozaan diamictite is derived from the vicinity of Klipwal gold mine.

shock metamorphic evidence for impact, and in the related to large-scale impact events, which mighthave triggered the onset of the break-up of Gond-case of the Barberton spherule layers, the chemical

and mineralogical evidence is not in agreement wana.Here we report the results of a systematic searchwith an impact association (Koeberl and Reimold

1995a). for shock petrographic indicators in Dwyka diamic-tite from a variety of localities in South Africa andRecently, Oberbeck et al. (1993a, 1993b) and

Rampino (1992, 1994) suggested that some of the Namibia.extensive breccia horizons, which in the past wereregarded as tillites or diamictites of glaciogenic or The Impact Ejecta–Diamictite/Tilliteglaciomarine origin, could represent impact ejecta Controversydeposits. (Note that the term breccia, commonlydefined as fragmentites with angular clasts [e.g., According to the Glossary of Geology (Bates and

Jackson 1987) the term diamictite is defined as ‘‘ABates and Jackson 1987], is used here in a widersense to describe clast-bearing, matrix-supported comprehensive, nongenetic term proposed by Flint

et al. (1960) for a nonsorted or poorly sorted, non-fragmentites—irrespective of the shapes of clasts—as is customary in planetology [e.g., Stoffler and calcareous, terrigenous sedimentary rock that con-

tains a wide range of particle sizes, such as a rockGrieve 1996].) These publications emphasized thepossibility that the Late Carboniferous to Early Per- with sand and/or larger particles in a muddy ma-

trix; e.g. a tillite or a pebbly mudstone.’’ In contrastmian Dwyka deposits of the Karoo Supergroup, ex-posed over much of the southern African subconti- to this nongenetic definition, the term tillite refers,

according to the same source, to ‘‘A consolidatednent (figure 1), could represent impact ejecta

Journal of Geology A R E D I A M I C T I T E S I M P A C T E J E C T A ? 519

or indurated sedimentary rock formed by lithifica- teristics of diamictites and tillites are similar tothose of debris-flow or mass-flow sediment depos-tion of glacial till, esp. pre-Pleistocene till (such as

the Late Carboniferous tillites in South Africa and its and do not necessarily indicate a glaciogenic ori-gin. However, he conceded that, at least in someIndia),’’ thus restricting the usage of this term to

glacial sediment. cases, a glaciomarine origin for such deposits ap-pears likely, but emphasized that, in the past, sev-In recent years, many diamictites have been re-

garded as being of glaciomarine origin (Von Brunn eral breccia deposits that were later interpreted tobe of impact origin had originally been misidenti-1994; Visser 1994; Eyles and Eyles 1983). For exam-

ple, Visser (1994) concluded that massive diamic- fied as sedimentary strata (for example, the ejectafrom the Ries, Germany, and Puchezh-Katunki,tite beds from the Dwyka Group either had an ori-

gin by rain-out or were the result of subaqueous Russia, impact structures).The identification of an impact origin of a struc-sediment gravity flows. In contrast, Rampino

(1992, p. 99) stated that ‘‘tillites are now generally ture or of rock formations requires the fulfillmentof a number of criteria, of which the presence ofreinterpreted as glaciomarine debris-flow depos-

its,’’ and that ‘‘. . . they are now found to contain shock metamorphic effects is the most important.Shock metamorphism comprises microdeforma-evidence of probable shock-induced deformation,

and are commonly associated with volcaniclastic tion effects and rock deformation resulting fromthe high shock pressures (up to hundreds of GPa),debris.’’ Rampino claimed that ‘‘these deposits co-

incide with extinctions and independent evidence temperatures (up to tens of thousands of °C), andstrain rates (106–108 s21) uniquely associated withof impacts, and are typically the initial sediments

in rift-related basins, close to known or possible hypervelocity impact events. These shock effectsare well known, well documented, and have beenimpact structures.’’

The main arguments presented by Oberbeck et extensively calibrated through decades of shock ex-perimental work (e.g., Stoffler 1972; Grieve et al.al. (1993a) in support of an impact origin of selected

diamictites are: 1996; Reimold 1996; Huffman and Reimold 1996,and references therein). One of the most character-1) Based on current understanding of the impact

cratering flux for this planet during the Phanero- istic shock deformation indicators is provided bythe so-called planar deformation features (PDFs),zoic, a significant proportion of the Earth’s surface

was supposedly covered by thick impact ejecta. Ob- which are formed in many rock-forming and acces-sory minerals. Quartz is most often studied for theerbeck et al. (1993a) estimated that 10% of the

Earth’s surface could have been covered by $10 m presence of shock effects because of its abundancein terrestrial crustal rocks and its strong suscep-of ejecta within the last 2 Ga and up to 2% of the

global surface could have been covered by up to 200 tibility to shock deformation (e.g., Stoffler andLangenhorst 1994; Grieve et al. 1996).m of ejecta. Impact deposits in excess of 3 km

thickness should have been produced locally; dia- Rampino (1994) claimed that his preliminary mi-croscopic studies of thin sections of diamictitesmictites of such thickness were indeed known.

2) Similarities between the textures of impact from various continents had shown that deforma-tion features ‘‘similar to those considered indica-ejecta, such as suevite and Bunte Breccia (the type

locality for these impact breccias is the Ries Crater; tive of shock deformation in quartz might be pres-ent in some diamictites’’ (p. 451). It should bee.g., Engelhardt 1990; Engelhardt et al. 1995) and

diamictites were emphasized. Textural and geolog- noted, though, that such undocumented ‘‘deforma-tion’’ features are in no way indicative of impactical parameters generally used to identify glacio-

genic sediments could also be related to impact processes by the standards required in the study ofimpact cratering (Grieve et al. 1996).breccias.

3) The geological evolution of the Earth’s crust The interpretation of diamictites by Oberbeck etal. (1993a) and Rampino (1994) did not remain un-prior to and around the time of the Gondwana

break-up was discussed with emphasis on the near- challenged (Young 1993; Oberbeck et al. 1993b; LeRoux 1994; Oberbeck et al. 1994). Oberbeck et al.coincidence of extensive magmatic activity (e.g.,

the widespread and voluminous Karoo magmatism) (1994) drew attention to the finding of a shockmetamorphosed granite clast from the Dutch Peakand deposition of the ‘‘diamictites.’’ Oberbeck et

al. (1993a) suggested that the Cape Fold Belt and diamictite of southern Utah (cf. also Horz et al.1994), but they conceded that this observation ofassociated basin structures may be possible rem-

nants of large impact basins remaining after paleo- an isolated clast does not have to be evidence of animpact origin of the host diamictite, as this singlecontinental break-up.

Rampino (1994) noticed that the textural charac- shocked clast could represent impact debris incor-

520 W O L F U W E R E I M O L D E T A L .

porated in the diamictite by secondary sedimen- a considerable section of Dwyka diamictite. To im-prove the statistical value of our microscopic inves-tary processes. An example of such resedimenta-

tion was recently reported to us by H. Henkel of tigation, multiple thin sections were prepared fromselected specimens. Altogether, 74 thin sectionsStockholm (pers. commun. to W.U.R., 1996): At the

beginning of this century, glassy clasts had been were examined in detail.Lithic and mineral clasts in these diamictiterecognized in a tillite deposit near Uppsala in the

Uppland Province (Sweden), more than 200 km samples were carefully scanned with petrographicmicroscope magnifications between 1603 andsouth of the Dellen structure. By tracing glacial de-

bris containing such glass clasts northward, the 4503 for microdeformations that might possibly berelated to tectonic deformation or impact-relatedSwedish geologist Hogboom was able in 1905 to

identify the Dellen structure as their place of ori- shock deformation. Modal analysis, with respect toclast/matrix ratios as well as clast populations,gin. This structure has since been confirmed to be

of impact origin (e.g., Pesonen 1996). were obtained for 15 Dwyka samples covering alllocalities, with the exception of the MpumalangaIn the absence of other sufficiently unambiguous

impact indicators, a search for shock metamor- and Gauteng localities. (It is important to note that,in this study, we use the term ‘‘clasts’’ for particlesphosed material in diamictite remains the only way

to test the hypothesis that some diamictites in the .30 µm in size, i.e., those fragments that could beeasily identified at these microscopic magnifica-stratigraphic record could represent impact ejecta.

To our knowledge, such a comprehensive study of tions. Smaller clastic particles were counted as ma-trix.) Clast population statistics were obtaineddiamictites has not yet been reported in the litera-

ture. However, an investigation by Schmitz et al. for most samples of this study, either by detailedpointcounting on selected thin sections, or by esti-(1994), who studied a sedimentary sequence with

multiple bentonite layers from Gotland (Sweden) mation of clast proportions in all thin sections rep-resenting a given locality. Since only small samplesfor possible shock effects in quartz, is relevant.

Their search did not yield any evidence of shock were available from most sampling sites, popula-tions of clasts .1 cm could not be considered.deformation.

Samples and Analytical Procedures Results

Clast Population. A typical Dwyka diamictiteOur study was focused on specimens of Dwyka dia-mictite of the Late Carboniferous–Jurassic Karoo sample may show some resemblance to suevitic

impact breccias (see figure 2a and b). Most DwykaSupergroup of Southern Africa (SACS 1980; for de-tail on Dwyka diamictites cf. Visser 1990; Visser et diamictite samples consist of generally well-

rounded, but locally (e.g., the drillcore sample fromal. 1990; von Brunn and Talbot 1986; Gravenor andvon Brunn 1987; von Brunn 1994, and references locality 7, figure 1) predominantly angular to sub-

rounded, mineral and lithic clasts set in a nonde-therein). Figure 1 illustrates the regional extent ofDwyka outcrop in the main Karoo basin of Nami- script, phyllosilicate-rich, matrix. In outcrop (fig-

ure 3a), clast sizes vary from barely visible (1–2bia and South Africa and illustrates our sample lo-calities. mm) to over half a meter. Clast size statistics may

be variable from locality to locality, as well asWe studied 25 Dwyka samples from 13 localitiesin Namibia (Keetmanshoop and Orupembe areas), within a single, large study area such as the Umlaas

Road Quarry locality. On thin section scale, clastthe southwestern Cape (Laingsburg-Matjiesfonteinregion), from KwaZulu-Natal Province (localities sizes vary between 30 µm and about 1 cm, and

again, the clast size distributions, at this scale, are4–9), from Mpumalanga Province (Majuba Collierynear Amersfoort town, from a drill core from the variable from locality to locality.

The clast population modal data of the studiedMatla and Middelburg coal mines, near Middelburgtown—localities 10, 12, and 13), and from Gauteng diamictite samples are summarized in table 1. Ma-

trix proportions in Dwyka diamictite are highlyProvince (Vereeniging and Vredefort Dome areas,locality 11). In addition, 28 samples were collected variable, ranging from 30 to 55 vol %. The most

important clast types in the samples from the west-in the Umlaas Road Quarry near Cato Ridge, eastof Pietermaritzburg (locality 4, figure 1), covering ern, southern, and eastern parts of the Karoo basin

(localities 1–6) are quartz, feldspar, and rock frag-an area of about 250 3 250 m from different levelsdown to a maximum quarry depth of about 70 m ments. Their proportions range from 7–13 vol %,

8–15 vol %, and 10–27 vol %, respectively. In these(see description in Gravenor and von Brunn 1987),thus providing a vertical sampling profile through samples plagioclase, as well as K-feldspar and alkali

Journal of Geology A R E D I A M I C T I T E S I M P A C T E J E C T A ? 521

A B

Figure 2. Comparison of a Dwyka diamictite specimen (a) from near Vereeniging (Gauteng)—courtesy of the Geologi-cal Curator of MUSEUMAFRICA, Johannesburg—and (b) impact breccia in the form of a suevite sample from theRoter Kamm impact crater in Namibia. Both samples are about 12 cm wide.

feldspar, are equally important, whereas in the metamorphic clasts, in about 40 vol % matrix. Thegranitoid clast content in this sample is muchsamples from the Vryheid and Mpumalanga regions

(sites 7–10) plagioclase is far more abundant than higher than in all other samples from this north-eastern part of the Dwyka basin. The sample fromother feldspar minerals. In all samples from the

western, southern, and eastern localities, granit- Vereeniging (loc. 11) has mainly quartzite and chertclasts, and the samples from Matla and Middelburgoids form significant clast components. Other min-

eral clast species recognized in these samples in- coal mines (loc. 12 and 13) also have high sedimen-tary clast proportions, mixed with significant, butclude pyroxene, chlorite, carbonate (calcite and

minor dolomite), opaque minerals (mainly magne- not very high, granitoid clast contents. In addition,the Middelburg sample is characterized by a (fortite, minor sulfides), epidote, garnet, muscovite, bi-

otite, zircon, and amphibole, which generally occur this sample suite) unique carbonate clast compo-nent. Samples from Matla and Middelburg alsoas trace components but in individual samples may

constitute a significant, but never large, compo- have a noteworthy granitoid component, whichcomprises mainly granophyric felsic clasts (figurenent. Lithic fragments were predominantly derived

from granitoid sources, but mafic igneous, meta- 4g and h) that could well have been derived fromthe proximal Bushveld granophyre.morphic, and sedimentary rocks presumably were

constituents of source areas of varied lithological Not only the clast populations, but also the grainsize and shape statistics are highly variable forcompositions. Instead of having homogeneous clast

populations, the studied samples display signifi- these samples from the northern and northeasternparts of the study region. Grain size statistics shiftcant lithological variety.

In contrast to the Dwyka samples from other re- significantly, with maxima either occurring around1 to 2 mm or at much smaller grain sizes. In indi-gions, the samples from the Vryheid (loc. 7), Ve-

reeniging (loc. 11), and Mpumalanga (loc. 8–10) lo- vidual samples, mainly angular fragments were ob-served, with rounded clasts mostly occurring in thecalities are all dominated (.40 vol %) by material

from a sedimentary source. Granitoid clasts and larger grain sizes (.2 mm); other samples have amore equilibrated population of angular to roundedmafic igneous materials account for ,1 vol %, al-

though some quartz (and quartzite—annealed clasts.Search for Shock Metamorphism. The main pur-quartz) clasts may be derived from granitoid precur-

sor rocks. One of us (VvB) found significant pose of this study was to scan as many lithic andmineral fragments as possible in all these samplesamounts of large granitoid clasts at Dwyka diamic-

tite exposures in the Vryheid region. for microdeformation occurrence and style. Perthin section, up to several thousand fragments wereThe Majuba Colliery (loc. 10, figure 1) samples

contain a mixture of about 10 vol % granitoid, 31 studied, with an emphasis on quartz and feldspargrains, as these minerals are very susceptible tovol % sediment, 5 vol % volcanic, and ,1 vol %

Figu

re3.

(a)

Abl

ock

ofU

mla

asR

oad

Qu

arry

Dw

yka

diam

icti

tew

ith

typi

cal,

angu

lar

tosu

brou

nde

d,u

pto

abou

t6

cmla

rge,

clas

tsof

vein

quar

tz(l

ower

cen

ter)

,gr

anit

e(u

pper

left

corn

er),

and

quar

tzit

e(b

elow

pen

onri

ght)

.P

en13

cmlo

ng.

(b)

Ase

tof

subp

lan

arm

icro

defo

rmat

ion

feat

ure

sin

aqu

artz

grai

nfr

omD

wyk

adi

amic

tite

from

Kee

tman

shoo

p(N

amib

ia).

Wid

thof

fiel

dof

view

:355

µm;c

ross

edpo

lars

.(c)

Nea

rly,

but

not

perf

ectl

y,pl

anar

flu

idin

clu

sion

trai

lsin

aw

ell-

rou

nde

dqu

artz

clas

tfr

oma

Dw

yka

diam

icti

tefr

omth

eM

atji

esfo

nte

inar

ea.

Wid

thof

fiel

dof

view

:35

5µm

,pa

rall

elpo

lars

.(d

)A

not

her

exam

ple

oflo

wst

rain

-rat

em

icro

defo

rmat

ion

inth

efo

rmof

two

sets

(arr

ows)

ofn

on-p

lan

ar‘‘l

amel

lae’

’in

quar

tzof

the

sam

esa

mpl

eas

in(c

).W

idth

offi

eld

ofvi

ew:3

55µm

,cr

osse

dpo

lars

.

AB

CD

522

Figu

re3.

(Con

tin

ued

)(e

)C

urv

ipla

nar

mic

rode

form

atio

nfe

atu

res

ina

part

iall

yan

nea

led

quar

tzit

iccl

ast

ina

Dw

yka

diam

icti

tefr

omK

waZ

ulu

-Nat

al.W

idth

offi

eld

ofvi

ew:1

.1m

m,c

ross

edn

icol

s.(f

)Kw

aZu

lu-N

atal

Dw

yka

diam

icti

te:p

olym

ict

brec

cia

wit

h‘‘b

asal

tic’

’(m

ark

ed‘‘B

’’),

prob

ably

gran

itoi

d-de

rive

dfe

ldsp

ath

ic,

and

cher

t(m

ark

ed‘‘C

’’)cl

asts

.W

idth

offi

eld

ofvi

ew:

3.5

mm

,cr

osse

dn

icol

s.(g

)A

gran

oph

yre

clas

tin

sam

ple

EG

H27

0fr

omM

atla

coal

min

e;w

idth

offi

eld

ofvi

ew:

1.75

mm

wid

e,cr

osse

dn

icol

s.(h

)A

not

her

mic

roph

otog

raph

ofE

GH

270

show

ing

anot

her

gran

oph

yre

clas

t(l

ower

righ

t)as

wel

las

aty

pica

lly

elon

gate

d,ov

oid

shal

efr

agm

ent

(arr

ow);

wid

thof

fiel

dof

view

:2.

2m

mw

ide,

cros

sed

nic

ols.

EF

GH

523

524 W O L F U W E R E I M O L D E T A L .

Table 1. Modal Compositions of Dwyka Diamictite Samples in this Study

W-S-E loc Vryheid Majuba Matla MiddelburgComponent (loc. 1-6) (loc. 7) (loc. 10) (loc. 12) (loc. 13)

Quartz 7–13 25 25 30 20Feldspar 8–15 ,1 3 2 ⋅ ⋅ ⋅Pyroxene 0–2 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Chlorite 0–1.7 ⋅ ⋅ ⋅ 3 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Carbonate 0–1.7 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ 1 10Opaques ,1.3 1 1 1 2Epidote 0–1.2 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Garnet 0–1.2 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Muscovite 0–1.2 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Biotite 0–1.2 ,0.5 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Amphibole 0–1.2 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Mafic igneous rock 0–4.9 ,1 5 3 5Metamorphic rock 0–2.3 ,1 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅Shale/Siltstone ⋅ ⋅ ⋅ 10 ⋅ ⋅ ⋅ 2 2–5Sandstone/Arkose 0–1.8 2 12 1 ⋅ ⋅ ⋅Quartzile ⋅ ⋅ ⋅ 10 ⋅ ⋅ ⋅ 5 5Chert ⋅ ⋅ ⋅ 20 ⋅ ⋅ ⋅ 2 15Granitoids 10–18 ,1 10 ,3 3Matrixa 47–55 30 40 45 35

Note. All data in volumes %.a All material ,30 µm.

shock deformation. Some of the observed microde- tion features described from quartz in BushveldComplex-related rocks (Buchanan and Reimoldformations are shown in figures 3 and 4.

In many samples, from all localities, quartz 1996; Joreau et al. 1997). The latter authors studiedsuch features in Bushveld samples by transmissiongrains with fluid inclusion trails were observed.

However, in contrast to the appearance of impact- electron microscopy and determined that they rep-resent bands of high dislocation densities. Such mi-derived shock metamorphic decorated PDFs (usu-

ally recognizable as planar trails of fluid or, some- crodeformation is not typical for rocks that haveundergone high-pressure shock deformation. In thetimes, solid inclusions; cf. Leroux et al. 1994; Stof-

fler and Langenhorst 1994), the trails observed in case of Buchanan and Reimold’s (1996; other un-publ. data) work, all localities in the Bushveldquartz clasts from Dwyka samples are never planar

and are generally much wider-spaced than the 2 to Complex from which samples with such deforma-tion could be obtained are closely associated with5 µm spacings characteristic for impact-diagnostic

PDFs (e.g., figure 3c and d ). Lamellar deformation major fault zones. Thus, it appears that such high-dislocation density bands are formed under tec-features in quartz in Dwyka diamictite do not con-

form to the microscopic criteria for bona fide shock tonic strain rates and stresses.Deformation features like those shown in figuresmetamorphic PDFs, as defined by Alexopoulos et

al. (1988), with regard to all their characteristics. 3b to d generally occur in quartz in Dwyka samplesas only one set per deformed grain, with the singleThey are not sharp and distinctly parallel deforma-

tion features. Spacings between individual features exception of the grain shown in figure 3d, whichcontains two sets of widely spaced, non-planar de-generally exceed 5 µm, and the widths of individual

features are on the order of 3 µm. Most importantly, formation features. Undulatory extinction is fre-quently observed in quartz grains. Besides the de-these microdeformations are not planar, and in

most cases they are not continuous across their scribed microdeformation features in quartz, noclasts of other rock-forming minerals display anyhost grains. It is also important that not a single

glass fragment or a fragment resembling devitrified significant microdeformation. Individual plagio-clase grains may be kinked or flexed, and a few feld-glass could be identified. Such glass/devitrified

glass fragments are characteristic of suevitic im- spar grains show intragranular cataclasis. Severalbiotite clasts with kinkbanding were counted.pact breccia.

Some of the subplanar microdeformation fea- However, all of these deformation effects can beproduced under normal tectonic conditions and dotures in quartz clearly represent Bohm lamellae.

Others (e.g., figure 3b) closely resemble deforma- not represent characteristic shock (impact) defor-

Figure 4. (a) Abundant subplanarfluid inclusion trails across twoquartz grains in a quartzite clast inDwyka diamictite sample JH1 fromnorth of Vryheid. Width of field ofview: 1.1 mm, crossed nicols. (b)Dwyka diamictite DLYA from Ma-juba Colliery: Fine- to medium-grained clast population withmainly rounded to subrounded pla-gioclase, microcline, as well asschist (‘‘s’’), sandstone (‘‘ss’’), andmuch quartz; 3.5 mm wide, crossednicols. (c) Typical, irregular, fluid in-clusion trails in a subangular quartzclast of sample Ma-Ca-1 from Matlacoal mine; 900 µm wide, crossed ni-cols.A

B

C

526 W O L F U W E R E I M O L D E T A L .

mation. Finally, a number of boudinaged or flexed metamorphic evidence and found considerablevariability in the clast population. Evidence ofshale fragments were observed.

In summary, none of these observations provide shock metamorphism represents a critical criterionfor the recognition of impact breccias, and it mustany evidence of shock deformation. The observed

defects are most likely the result of tectonic defor- be assumed that impact ejecta would be dominatedby a consistent clast population determined by themation incurred at relatively low strain rates or

were formed during physical diagenesis (such as composition of the relevant target area (it can, how-ever, not be excluded that secondary sedimentarycompaction, e.g., Boggs 1987). They could well

have been produced by subglacial crushing/commi- processes could lead to further admixture of otherlithological components; the current data base onnution beneath a moving ice mass.

This conclusion is further supported by our ob- distal impact ejecta does not allow wide-rangingpredictions as to the effectiveness of secondary pro-servation that the samples from the Laingsburg-

Matjiesfontein area (Western Cape), located just cesses on the lithological composition of an impactejecta breccia). With regard to all Dwyka diamictitenorth of the structurally strongly deformed Cape

Fold Belt (Van Niekerk 1993; Halbich 1983), con- samples investigated so far, absolutely no prima fa-cie evidence in favor of any relation to impact pro-tain a higher abundance of such deformed quartz

and feldspar clasts than the samples from other re- cesses has been found. In conclusion, the Dwykadiamictite cannot, at this point, be considered agions not directly related to orogenic terranes. The

samples from close to the Fold Belt are also tran- candidate for proximal or distal impact ejecta. Alink between the Permo-Triassic (e.g., Thomas etsected by a prominent cleavage. The lowest abun-

dance of deformed clasts was found in the samples al. 1993) Cape Fold Belt and possible impact defor-mation, as purported by Oberbeck et al. (1993a) andfrom the northern and northeastern parts of the

Dwyka basin. The low strain rate microdeforma- Rampino (1994), is speculative.Furthermore, no mass extinction has been docu-tion features observed in our samples from the vi-

cinity of the Cape Fold Belt are most likely either mented around the time of deposition of theDwyka Group, approximately 300 Ma, and no largeinherited from a tectonically deformed source re-

gion or were deformed during post-depositional tec- impact structures of this age are known. Any im-pact structures with an age of about 300 6 50 Matonic overprint. It is probable that some degree of

low strain rate deformation was imparted on mate- are ,40 km in diameter (Grieve et al. 1995), whichis far too small to cause thick ejecta deposits on arials involved in the formation of Dwyka breccia

during erosion and transport; for example, due to subcontinental scale. The assessment of Oberbecket al. (1993a) that impact events on Earth can de-shear stress produced at the base of ice sheets.

According to Oberbeck et al. (1993a) and Ram- posit ejecta layers of up to several kilometers thick-ness must be evaluated in view of known ejecta de-pino (1994), the Cape Fold Belt could represent the

remnant of several large impact structures. To our posits. No preserved ejecta deposits are known fortwo of the largest impact structures on Earth, theknowledge, no evidence has ever been produced

that could support or substantiate such a claim. In- originally 250–300 km wide Vredefort and Sudburystructures. The impact event leading to the forma-deed, such an assertion constitutes a good example

of circular reasoning: the Cape Fold Belt is hypothe- tion of the 200–300 km Chicxulub structure (Yuca-tan, Mexico) resulted in the deposition of a thin (,2sized to be of impact origin, and, therefore, the

Dwyka breccia represents impact breccia—and cm at distances .1000 km) global distal ejectalayer. An about 15 m thick proximal ejecta depositvice versa. In contrast to the views of Oberbeck et

al. (1993a) and Rampino (1994), we conclude that was found in Belize at a distance of about 360 kmfrom the center of the Chicxulub structurethe observed relative enhancement of low strain

rate microdeformation in clasts of Dwyka diamic- (Ocampo et al. 1996). This ejecta deposit (by coinci-dence termed ‘‘diamictite’’ by Ocampo et al. 1996,tite from the northern foothills of the Cape Fold

Belt is most likely linked to tectonic processes. because of its macroscopic appearance) was inter-preted as having originated through a combinationof ballistic and debris flow sedimentation. In viewImplications for a Possible Impact Origin of this evidence from large terrestrial impact eventsof Dwyka Diamictite (i.e., limited ejecta thickness even at the largestknown impact structures) one has to wonder aboutDetailed petrographic analysis of Dwyka diamic-

tite showed that all investigated samples are devoid the size of the impact structures that would lead tothe deposition of kilometer-thick ejecta layers!of any evidence of shock metamorphic deforma-

tion. We have demonstrated the lack of shock However, the result of the present study does not

Journal of Geology A R E D I A M I C T I T E S I M P A C T E J E C T A ? 527

preclude the possibility that future studies of dia- pracrustals) could further decrease the chance toidentify such deformation. Much of the material ofmictites will reveal that certain strata, currently as-

sumed to be of sedimentary origin, could indeed be distal ejecta deposits would only have been sub-jected to—at maximum—low shock pressures ,10the result of impact processes. Shock petrographic

studies of clasts in breccias of presumed sedimen- GPa and would not exhibit characteristic shock de-formation effects such as PDFs, diaplectic glasses,tary origin are still rare. Unfortunately, the concept

of shock metamorphism, and knowledge of the rel- isotropization, or mosaicism (cf. the detailed stud-ies on Ries Crater suevite and Bunte Breccia re-evant criteria, is not yet sufficiently well known

in the general geological community. It has to be viewed by Engelhardt 1990 and Engelhardt et al.1995).emphasized that, if one is aware of them, indicators

of shock metamorphism might be detected in some In contrast to shock pressures of 10 to 60 GPa,few shock experiments with most of the importantdetailed sedimentological or metamorphic studies.

Recent work on breccias from the Red Wing Creek rock-forming minerals were done at lower shocklevels between 1 and 10 GPa. The only known indi-and Newporte structures in North America pro-

vides a point in case, as only detailed petrographical cator of shock deformation in quartz in this regimeare arrays of short extension fractures, such assearches for shock metamorphic indicators pro-

vided sufficient evidence to recognize these brec- those described by Reimold (1990 and referencestherein; cf. also Stoffler and Langenhorst 1994).cias as material of impact origin (Koeberl and Rei-

mold 1995b; Koeberl et al. 1996b). Further experimental work for the major rock-forming minerals at low shock pressures is needed.Another case in point is a recent study of a brec-

cia intersected in a borehole from the western sec- To evaluate the statistical significance of a studysuch as this one is equally difficult, as critical pa-tor of the central uplift structure (the so-called

Vredefort Dome, figure 1) of the Vredefort impact rameters are only ill-defined. We estimate that inthe course of this study a minimum of 75,000 min-structure (Reimold and Gibson 1996). This breccia

horizon lies in a stratigraphic position where a dia- eral and lithic clasts were scanned for the possiblepresence of shock metamorphic microdeformationmictite has been mapped in the past as part of the

Government Reef Group of the Witwatersrand Su- effects. As about 25% of these clasts (compare table1) consist of quartz, an effective number of at leastpergroup. Detailed petrographic study revealed that

the samples from the Vredefort Dome contain a 18,750 quartz grains has been studied. At this stageit is critical at what proportion quartz grains withlarge number of shock metamorphosed quartz

clasts, with planar fluid inclusion trails that repre- characteristic shock metamorphic deformation fea-tures should be expected to occur in an impactsent annealed PDFs as well as some diaplectic

quartz glass. In addition, this breccia carries a com- breccia, for which it is even uncertain whether itwould represent proximal or distal ejecta. On theponent of granitoid-derived clasts, which is absent

in this stratum in other parts of the Witwatersrand basis of the limited data concerning the proportionof moderately to highly shocked clasts in ejectabasin. Contacts between the breccia and adjacent

host quartzite are suggestive of violent breccia em- breccias it appears likely that only 1, or even less,volume % of such ejecta would be shock-deformed.placement: for example, the presence of rip-up

quartzite clasts in the breccia. As the location of It is equally uncertain how much impact melt orimpact glass should be expected to occur in suchthis breccia occurrence—within a confirmed im-

pact structure (Reimold and Gibson 1996)—is breccias, but it is assumed that this value shouldnot exceed, rather be lower than, 1 vol %. Frenchstrongly leading with regard to its origin, it is not

clear whether it represents a diamictite horizon (1990) cited several studies that indicated that im-pact melt rocks from impact structure interiorsoverprinted by shock metamorphic deformation or

an injection of impact breccia that occurred, by co- carry about 1 vol % shocked inclusions. But whatare the figures for clastic impact breccias? In ourincidence, in the stratigraphic position where a dia-

mictite would be expected. case, a value of 1 vol % of shocked particles wouldimply that some 188 grains with typical shock de-A difficulty with regard to investigations into

the possible impact origin of apparently sedimen- formation could have been identified in this study.However, at 0.1 vol %, this figure would be reducedtary strata is the requirement to scan numerous

grains for shock metamorphic deformation in large to only about 19 grains.Because of these inherent uncertainties, we havenumbers of thin sections. Distal impact ejecta (as

opposed to proximal ejecta) do not contain abun- refrained from attempting a probability estimatefor specific levels of confidence—as attempted bydant shock deformed clasts, and post-depositional

reworking (mixing of impact ejecta with local su- French (1990) for his search for shock deformed

528 W O L F U W E R E I M O L D E T A L .

quartzite xenoliths in Rooiberg Group strata of the metamorphic deformation. No evidence for shockmetamorphism was observed. Thus, the assertionBushveld Complex. French determined that sam-

pling of about 455 xenoliths, for an a priori value that the Dwyka diamictite could represent an im-pact breccia deposit remains, to date, unsupportedof 1% clasts with shock deformation, would be suf-

ficient to allow a 99% confidence level for his re- by any concrete evidence.It is, of course, not possible to exclude categori-sult (that no shocked inclusions were present in

these rocks). Clearly our number of observations cally that some individual diamictite horizonscould be impact ejecta horizons. However, the pres-exceeds such an estimate by orders of magnitude.

However, besides the uncertainty about the pri- ent study and the available data do not support thisclaim in general. Any suggestions regarding themary percentage of shocked clasts in impact brec-

cias, the possibility of further dilution by secondary possible impact origin of diamictites/tillites haveto be substantiated by detailed shock-petrographicsedimentary processes after deposition presents an

additional problem for estimation of the reliability studies. Only such studies, which may reveal char-acteristic evidence of shock metamorphic deforma-of studies such as this one.

Clearly this discussion demands, on the one tion, can be used to confirm any links to impactevents.hand, that studies such as the one reported here

need to be continued, and that, on the other, de-tailed shock deformation studies on proximal and A C K N O W L E D G M E N Tdistal impact ejecta need to be taken up. Several GENCOR SA Limited kindly permitted to discussearlier studies by our group (e.g., Reimold et al. the petrographic findings in a breccia occurrence1992, 1994; Koeberl and Reimold 1995b; Koeberl et from the western sector of the Vredefort Dome.al. 1996b) have shown that, if shocked particles are Several samples utilized in this study were kindlypresent, they are readily detected, even in small provided by J. Hancox (University of the Witwaters-numbers, in only a few specimens studied. How- rand), J. N. J. Visser (Free State University, Bloem-ever, all these investigations dealt with impact fontein), and B. Moonsamy and C. Mafiri (Univer-breccias from the interior of impact structures, sity of the Witwatersrand). J. P. Hunt (Universitywhich all are of relatively small size (,9 km diame- of the Witwatersrand) assisted with mineralogicalters). analysis. This work has been supported by the Aus-

trian Fonds zur Forderung der wissenschaftlichenForschung, project No. 8794-GEO (to CK), andConclusions by the Foundation for Research Development,through grants to WUR and VvB. Bevan French,More than 75,000 mineral and rock clasts in

Dwyka diamictite samples were investigated for Sandro Montanari, and an anonymous referee pro-vided helpful reviews.the presence of any evidence of characteristic shock

R E F E R E N C E S C I T E D

Alexopoulos, J. S.; Grieve, R. A. F.; and Robertson, P. B., tite and mass extinctions: Evidence for a Late Dev-onian asteroid impact: Science, v. 257, p. 1102–1104.1988, Microscopic lamellar characteristics of shock-

generated varieties: Geology, v. 16, p. 796–799. Clymer, A. K.; Bice, D. M.; and Montanari, A., 1996,Shocked quartz from the late Eocene: Impact evidenceBates, R. L., and Jackson, J. A., eds., 1987, Glossary of

Geology: Alexandria, Va., Am. Geol. Inst., 788 p. from Massignano, Italy: Geology, v. 24, p. 483–486.Dressler, B. O.; Grieve, R. A. F.; and Sharpton, V. L., eds.,Boggs, S., Jr., 1987, Principles of Sedimentology and Stra-

tigraphy: Columbus, Merrill Pub., 784 p. 1994, Large meteorite impacts and planetary evolu-tion: Geol. Soc. America Spec. Paper, 293, 348 p.Buchanan, P. C., and Reimold, W. U., 1996, Analysis of

deformation lamellae in quartz grains from the Rooi- Engelhardt, W. v., 1990, Distribution, petrography andshock metamorphism of the ejecta of the Ries craterberg Felsite, Bushveld Complex, South Africa, and as-

sociated rocks: Lunar Planet. Sci., v. 27, p. 175–176. in Germany—a review: Tectonophysics, v. 171, p.259–273.Byerly, G. R., and Lowe, D. R., 1994, Spinel from Archean

impact spherules: Geochim. Cosmochim. Acta, v. 58, ———; Arndt, J.; Fecker, B.; and Pankau, H. G., 1995,Suevite breccia from the Ries crater, Germany: Ori-p. 3469–3486.

Claeys, P., and Casier, J.-G., 1994, Microtektite-like im- gin, cooling history, and devitrification of impactglasses: Meteoritics, v. 30, p. 279–293.pact glass associated with the Frasnian-Famennian

boundary mass extinction: Earth Planet. Sci. Lett., v. Eyles, C. H., and Eyles, N., 1983, Glaciomarine modelfor upper Precambrian diamictites of the Port Askaig122, p. 303–315.

———; Casier, J.-G.; and Margolis, S. V., 1992, Microtek- Formation, Scotland: Geology, v. 11, p. 692–694.

Journal of Geology A R E D I A M I C T I T E S I M P A C T E J E C T A ? 529

Flint, R. F.; Sanders, J. E.; and Rodgers, J., 1960, Diamic- ule beds in the Barberton Mountain Land, South Af-rica: No evidence for impact origin: Precamb. Res., v.tite, a substitute term for symmictite: Geol. Soc.

America Bull., v. 71, p. 1809. 74, p. 1–33.———, and ———, 1995b, The Newporte impact struc-French, B. M., 1990, Absence of shock-metamorphic ef-

fects in the Bushveld Complex, South Africa: Results ture, North Dakota: Geochim. Cosmochim. Acta, v.59, p. 4747–4767.of an intensive search: Tectonophysics, v. 171, p. 287–

301. ———; ———; and Boer, R. H., 1993, Geochemistry andmineralogy of Early Archean spherule beds, BarbertonGlikson, A. Y., 1993, Asteroids and early Precambrian

crustal evolution: Earth-Sci. Rev., v. 35, p. 285–319. Mountain Land, South Africa: Evidence for origin byimpact doubtful: Earth Planet. Sci. Lett., v. 119, p.Gostin, V. A.; Haines, P. W.; Jenkins, R. J. F.; Compston,

W.; and Williams, I. S., 1986, Impact ejecta horizon 441–452.———; ———; and Brandt, D., 1996b, Red Wing Creekwithin late Precambrian shales, Adelaide Geosyn-

cline, south Australia: Science, v. 233, p. 198–200. Structure, North Dakota: Petrological and geochemi-cal studies, and confirmation of impact origin: Mete-Gravenor, C. P., and von Brunn, V., 1987, Aspects of Late

Paleozoic glacial sedimentation in parts of the Parana oritics Planet. Sci., v. 31, p. 335–342.Kring, D. A., 1993, The Chicxulub impact event and pos-Basin, Brazil, and the Karoo Basin, South Africa, with

special reference to the origin of massive diamictite, sible causes of K/T boundary extinctions, in Boaz, D.,and Dornan, M., eds., Proceedings of the First Annualin McKenzie, G. D., ed., Gondwana Six: Stratigraphy,

Sedimentology and Paleontology: Washington, Am. Symposium of Fossils of Arizona: Mesa, AZ, MesaSouthwest Museum and Southwest Paleont. Soc., p.Geophys. Union Geophys. Mon. 41, p. 103–111.

Grieve, R. A. F.; Langenhorst, F.; and Stoffler, D., 1996, 63–79.Langenhorst, F., 1996, Characteristics of shocked quartzShock metamorphism of quartz in nature and experi-

ment: II. Significance in geoscience: Meteoritics in late Eocene impact ejecta from Massignano (An-cona, Italy): Clues to shock conditions and source cra-Planet. Sci., v. 31, p. 6–35.

———; Rupert, J.; Smith, J.; and Therriault, A. M., 1995, ter: Geology, v. 24, p. 487–490.Le Roux, J. P., 1994, Impacts, tillites, and the breakup ofThe record of terrestrial impact cratering: GSA Today,

v. 5, p. 193–196. Gondwanaland: A second discussion: Jour. Geology,v. 102, p. 483–485.Halbich, I. W., 1983, Tectonogenesis of the Cape Fold

Belt (CFB), in Sohnge, A. P. G., and Halbich, I. W., eds., Leroux, H.; Reimold, W. U.; and Doukhan, J.-C., 1994, ATEM investigation of shock metamorphism in quartzGeodynamics of the Cape Fold Belt: Geol. Soc. South

Africa Spec. Pub. 12, p. 165–175. from the Vredefort Dome, South Africa: Tectonophy-sics, v. 230, p. 223–239.Hassler, S. W., and Simonson, B. M., 1995, Depositional

history of a resedimented impact layer in the Early Lowe, D. R., and Byerly, G. R., 1986, Early Archean sili-cate spherules of probable impact origin, South AfricaPrecambrian Hamersley Group, Western Australia, in

Montanari, A., and Coccioni, R., eds., Abstracts, The and Western Australia: Geology, v. 14, p. 83–86.Melosh, H. J., 1989, Impact Cratering, A Geologic Pro-Effects of Impacts on the Atmosphere and Biosphere

with Regard to Short- and Long-Term Changes: 4th cess: New York, Oxford Univ. Press, 245 p.Oberbeck, V. R.; Marshall, J. R.; and Aggarwal, H., 1993a,Int. Workshop ESF Sci. Network on Impact Cratering

and Evolution of Planet Earth (Ancona, May 1995), p. Impacts, tillites, and the breakup of Gondwanaland:Jour. Geology, v. 101, p. 1–19.88–89.

Horz, F.; Bunch, T. E.; and Oberbeck, V. R., 1994, ———; ———; and ———, 1993b, Impacts, tillites, andthe breakup of Gondwanaland: A Reply: Jour. Geol-Shocked material from the Dutch Peak diamictite,

Utah: Lunar Planet. Sci., v. 25, p. 565–566. ogy, v. 101, p. 679–683.———; Horz, F.; and Bunch, T. E., 1994, Impacts, tillites,Huffman, A. R., and Reimold, W. U., 1996, Experimental

constraints on shock-induced microstructures in nat- and the breakup of Gondwanaland: A second reply:Jour. Geology, v. 102, p. 485–489.urally deformed silicates: Tectonophysics, v. 256, p.

165–217. Ocampo, A. C.; Pope, K. O.; and Fischer, A. G., 1996,Ejecta blanket deposits of the Chicxulub crater fromJoreau, P.; Reimold, W. U.; Robb, L. J.; and Doukhan,

J.-C., 1997, A TEM study of deformed quartz grains Albion Island, Belize, in Ryder, G.; Fastovsky, D.; andGartner, S., eds., The Cretaceous-Tertiary event andfrom volcaniclastic sediments associated with the

Bushveld Complex, South Africa: Eur. Jour. Miner., other catastrophes in Earth history: Boulder, Col.,Geol. Soc. America Spec. Paper 307, p. 75–88.no. 2/1997 (March 1997), in press.

Koeberl, C., 1996, Chicxulub—The K-T boundary im- Pesonen, L. J., 1996, The impact cratering record of Fen-noscandia: Earth, Moon, and Planets, v. 72, p. 377–pact crater: A review of the evidence, and an introduc-

tion to impact crater studies: Abh. Geol. Bundesan- 393.Rampino, M. R., 1992, Ancient ‘‘glacial’’ deposits arestalt, Wien, v. 53, p. 23–50.

———; Poag, C. W.; Reimold, W. U.; and Brandt, D., ejecta of large impacts: The Ice Age paradox explained(Abs): EOS—Trans. Am. Geophys. Union, Suppl., p.1996a, Impact origin of Chesapeake Bay structure and

the source of North American tektites: Science, v. 99.———, 1994, Tillites, diamictites, and ballistic ejecta of271, p. 1263–1266.

———, and Reimold, W. U., 1995a, Early Archean spher- large impacts: Jour. Geology, v. 102, p. 439–456.

530 W O L F U W E R E I M O L D E T A L .

Reimold, W. U., 1990, The controversial microdefor- Kotnik, B., eds., The Role of Impact Processes in theGeological and Biological Evolution of Planet Earth:mations in quartz from the Vredefort structure, South

Africa: South Afr. Jour. Geol., v. 93, 645–663. Int. Workshop (Postojna, Slovenia), p. 89–90.———, and Langenhorst, F., 1994, Shock metamorphism———, 1996, Impact cratering—A review, with special

reference to the economic importance of impact struc- of quartz in nature and experiment: I. Basic observa-tion and theory: Meteoritics, v. 29, p. 155–181.tures and the Southern African impact crater record:

Earth, Moon, and Planets, v. 70, p. 21–45. Taylor, S. R., 1992, Solar System Evolution—A New Per-spective: New York, Cambridge Univ. Press, 307 p.———, and Gibson, R. L., 1996, Geology and evolution

of the Vredefort impact structure, South Africa: Jour. Thomas, R. J.; von Veh, M. W.; and McCourt, S., 1993,The tectonic evolution of southern Africa: An over-Afr. Earth Sci., v. 23, p. 125–167.

———; Koeberl, C.; Partridge, T. C.; and Kerr, S. J., 1992, view: Jour. Afr. Earth Sci., v. 16, p. 5–24.Van Niekerk, B. N., 1993, Klastemaaksel van diePretoria Saltpan Crater: Impact origin confirmed: Ge-

ology, v. 20, p. 1079–1082. geplooide Dwykagroep, suid en wes van Laingsburg,Kaapprovinsie: Unpub. M.Sc. Thesis, Univ. of the Or-SACS (The South African Committee for Stratigraphy),

1980, Stratigraphy of South Africa, Handbook 8. Part ange Free State, Bloemfontein, 88 p.Visser, J. N. J., 1990, The age of the late Paleozoic glacio-1: Lithostratigraphy of the Republic of South Africa,

South West Africa/Namibia and the Republics of Bo- gene deposits in southern Africa: South Africa Jour.Geol., v. 93, p. 366–375.phutatswana, Transkei and Venda: Pretoria, Gov.

Printer, 690 p. ———, 1994, The interpretation of massive rain-out anddebris-flow diamictites from the glacial marine envi-Schmitz, B.; Jeppson, L.; and Ekvall, J., 1994, A search

for shocked quartz grains and impact ejecta in early ronment, in Deynoux, M.; Miller, J. M. G.; et al., eds.,Earth’s Glacial Record: Cambridge, Cambridge Univ.Silurian sediments on Gotland, Sweden: Geol. Mag.,

v. 131, 361–367. Press, p. 83–94.———, von Brunn, V., and Johnson, M. R., 1990, DwykaSharpton, V. L., and Ward, P. D., eds., 1990, Global catas-

trophes in Earth history: Geol. Soc. America Spec. Pa- Group, in Catalog South Africa Lithostrat. Units, Vol.2: Pretoria, Gov. Printer, p. 15–17.per 247, 631 p.

Silver, L. T., and Schultz, P. H., eds., 1982, Geological von Brunn, V., 1994, Glaciogenic deposits of the Permo-Carboniferous Dwyka Group in the eastern region ofimplications of impacts of large asteroids and comets

on the Earth: Geol. Soc. America Spec. Paper 190, the Karoo Basin, South Africa, in Deynoux, M.; Miller,J. M. G.; et al., eds., Earth’s Glacial Record: Cam-529 p.

Simonson, B. M., and Davies, D., 1996, PGEs and quartz bridge, Cambridge Univ. Press, p. 60–68.———, and Talbot, C. J., 1986, Formation and deforma-grains in a resedimented late Archean impact horizon

in the Hamersley Group of Western Australia: Lunar tion of subglacial intrusive clastic sheets in theDwyka Formation of Northern Natal, South Africa:Planet. Sci., v. 27, p. 1203–1204.

Stoffler, D., 1972, Deformation and transformation of Jour. Sed. Petrol., v. 56, p. 35–44.Williams, G. E., 1986, The Acraman impact structure:rock-forming minerals by natural and experimental

shock processes. I. Behavior of minerals under shock Source of ejecta in late Precambrian shales, SouthAustralia: Science, v. 233, p. 200–203.compression: Fortschr. Miner., v. 49, p. 50–113.

———, and Grieve, R. A. F., 1996, IUGS Classification Young, G. M., 1993, Impacts, tillites, and the breakup ofGondwanaland: A Discussion: Jour. Geology, v. 101,and nomenclature of impact metamorphic rocks: To-

wards a final proposal, in Drobne, K.; Gorican, S.; and p. 675–679.