Wednesday, February 09, 1994
                           SESSION 1:  PERSPECTIVES
                                9:00 - 12:30 PM

Chair(s):  K. Burke


Herbert T. D.   Fischer A. G.*   D'Hondt S. L.
   Cyclochronologic Approaches to the KT Event

The KT boundary crisis and associated events occurred within
Polarity Chron 29R, occupying only a small fraction of that
interval. Milankovitch-band cyclicity recorded in pelagic
sediments promises to provide time-resolution of events
within this Chron, down to the ca. 20,000 year level. As yet
only a few of many sedimentary sequences have been analyzed
but we present consonant data from the South Atlantic (Rio
Grande Rise, Walvis Ridge, Herbert and D'Hondt) and from
Spain (Kate and Sprenger). All sequences show precessional
signals (low-carbonate/high-carbonate couplets), and some
show a bundling of these into ca. 100 ka eccentricity
cycles, some of which show grouping into 400 ka eccentricity
cycles.

Unless the "boundary clay" represents a full 100,000 year
eccentricity cycle, Polarity Chron 29R, beginning after the
second or third precession in an eccentricity cycle, lasted
for about 600,000 years. The Ir event and main extinction
occurred after the third or fourth precession in the third
succeeding eccentricity cycle, about 320,000 years after the
magnetic reversal. The main boundary event thus falls near
the middle of the polarity chron; its seeming upward
displacement in stratigraphic plots of pelagic carbonates
results from a drastic (factor of 2-3) diminution of
sedimentation rates, owing to a crisis-induced drop in
carbonate production.

Cyclostratigraphic analysis of more KT boundary sections is
required to check and refine these results. A
cyclostratigraphic answer to the duration of the "boundary
clay" requires sites at which the 400 ka cyclicity is
developed above as well as below the boundary. A general
cyclochronology of the KT episode should provide timing of
associated isotope excursions (preceding positive and
succeeding negative ^13C anomalies) and biotic events (first
and last occurrences of species, recurrences of disaster
forms).


Jablonski D.*
   Mass Extinctions:  Persistent Problems and New Directions

Few contest that mass extinctions have punctuated the
history of life, or that those events were so pervasive
environmentally, taxonomically, and geographically that
physical forcing factors were probably involved. However,
consensus remains elusive on the nature of those factors,
and on how a given perturbation (impact, volcanism, sea-
level change, oceanic anoxic event) could actually generate
the observed intensity and selectivity of biotic losses. At
least two basic problems underlie these long-standing
disagreements: (1) difficulties in resolving the fine
details of taxon ranges and abundances immediately prior to
and after an extinction boundary and (2) the scarcity of
simple, unitary cause-and-effect relations in complex
biological systems.

(a) Detailed stratigraphic patterns: Local outcrops and
cores are the ultimate source of the data used to analyze
mass extinctions, but the pitfalls to taking local data at
face value are still little appreciated, resulting in
massive overinterpretation of paleontological patterns. The
accumulation of sediments and fossils is discontinuous, and
environmental change is the rule, so that temporal gaps and
stepped extinctions are inevitable on some scale in any
local sequence. Further, reworking and time-averaging mix
fossils from successive intervals; radiocarbon dates on
shells collected from surface sediments and forams in core
tops indicate a time-averaging window of 103-104 years in
marine shelf and deep-sea sediments alike [1]. Statistical
protocols are available to test for artificial extinction
steps and to place confidence limits on stratigraphic ranges
[2,3], but to date these have seen little use in the mass
extinction literature: nearly all workers wants to take
their data at face value and as virtually devoid of local
overprint. Clearly, the answer to this problem is to test
patterns against an appropriate null hypothesis.

(b) Inferring cause from effect: One fundamental obstacle in
linking extinction patterns and hypothesized forcing factors
resides in the nature of complex systems: nonlinearities,
thresholds, and elaborate feedbacks often rule out the
reconstruction of simple cause-and-effect cascades. The same
forcing factor might have radically different effects
depending on the state of the system at the time of
perturbation, and several alternative forcing factors might
produce the same biotic response. The survival of an
evolutionary lineage during a mass extinction, for example,
could be because it (1) lived in a habitat that was not
stressed, (2) possessed a physiology or life habit that
allowed it to survive in a stressed habitat, (3) was so
widespread that its range includes a locality that provided
a refuge, and so on. Urgently needed is not another catalog
of potential reasons for survival or extinction, but the
development of protocols for testing the alternatives. This
is partly a matter of constructing large, robust databases
amenable to statistical analysis, and coming to grips with
the need to integrate local and synoptic databases. With
sufficient tuning, the long list of potential KT killing
mechanisms can indeed account for virtually any conceivable
extinction pattern; therefore, while the mere fact of
observed selectivity clearly is no argument against impacts,
neither is it an argument in their favor. The time for
consistency arguments is past: hypothesized extinction
mechanisms need to generate unique predictions on the
timing, selectivity (taxonomic, biogeographic, ecologic), or
other biotic patterns to advance beyond the status of
plausible alternatives.

Selectivity is still a neglected area of study for mass
extinctions, and even negative results will be important
here. Selectivity is played out, if at all, at lower
taxonomic levels (families and genera), biogeographically,
or ecologically--and the relevant parameters are largely
absent from the synoptic databases [4]. Consider the problem
of multiple causation for one general pattern: the
apparently higher extinction intensities in the tropics. The
question is whether the greater losses of shallow-water late
Devonian corals relative to deep-water genera, or of
symbiont-bearing KT corals and rudist bivalves relative to
nonsymbiotic corals and non-rudist bivalves, occur because
(1) tropical biotas in general are fragile, perhaps because
their species are adapted to a narrow range of climatic and
other conditions; (2) reef communities in particular are
such a tightly woven network of biological interactions that
the initial removal of the same proportion of species as
were lost at high latitudes could be more disruptive; (3)
tropical biotas contain a large proportion of extinction-
prone endemics, so that losses are high here owing to
biogeographic structure; (4) the favored habitat of reef
communities, low-sedimentation, and low-nutrient shallow-
water platforms or ramps, is itself easily disrupted [5,6].
Some support exists for (3) and (4). First, among-province
variation in mass extinction intensities within latitudinal
belts tend to be positively related to the proportion of
endemic genera in the pre-extinction biota [7]. Second, a
global analysis of end-Cretaceous extinction in marine
bivalves found that tropical settings outside of the
carbonate platforms suffered no greater losses than did
extratropical faunas [8]; the reported latitudinal gradient
in plankton extinction intensities may have a similar basis.
The statistical dissection of alternative mechanisms offers
considerable promise for improving our understanding of
extinction mechanisms and biological consequences, both for
the Big Five mass extinctions and for the smaller extinction
maxima that occur throughout the Phanerozoic.

The initial physical and biological conditions must play a
role in the biotic response to a perturbation. Perhaps, for
example, impacts or volcanism at times of low relative sea
level yield greater taxonomic losses than the same event at
times high sea level when the thermal inertia and other
ameliorating effects of shallow seas are prominent. This
frequent suggestion has yet to be modeled rigorously, let
alone tested empirically, but may help to explain why some
geologically detectable impacts evidently had negligible
biotic effects. The global biota or even individual taxa may
vary in relative extinction vulnerability through time. For
example, it may take time to accumulate a new crop of
extinction-prone taxa after a major extinction event has
removed all but the hardiest lineages. Conversely, lineages
may evolve in directions that make them less vulnerable to
successive perturbations. Such biotic lags and long-term
shifts might explain waiting times between extinction events
[9, but see 10], and the differential responses of
individual taxa to successive extinctions [7].

The evolutionary impact of mass extinctions is another
active area of research. Many of the biotic replacements
once thought to represent competitive victories over
inferior lineages now appear to have been mediated by major
extinction events, even though most species extinctions in
the fossil record, probably >90%, occur outside of the five
major extinction events [11]. Mass extinctions have such
profound biological consequences because they bite deep into
standing diversity and disrupt background selection regimes,
not because they account for most species terminations.
Traits that favor survival during mass extinctions need have
little correlation with those that enhance survival and
diversification during background times, so mass extinctions
can have unpredictable and lasting evolutionary effects
[5,7]. Not only do mass extinctions remove taxa and
adaptations well-suited to the background regimes that
represent the great bulk of geologic time, they create
ecological and evolutionary opportunities by removing
incumbant, dominant taxa and enabling other taxa to
diversify in the aftermath of the extinction event. On the
other hand, mass extinctions do not completely reset the
evolutionary clock: many major evolutionary and ecological
trends transcend even the Big Five events (e.g., the
modernization of marine communities, the rise of flowering
plants, and of predatory neogastropods). We need to
understand survivors and rebounds as well as victims and
ecosystem collapse.

References: [1] Kidwell S. M. and Behrensmeyer A. K. (1993)
Paleont. Soc., Short Courses in Paleontology, 6. [2]
Gilinsky N. L. and Signor P. W. (1991) Analytical
paleontology, Paleont. Soc., Short Courses in paleontology,
4, Knoxville, Univ. TN. [3] Sepkoski J. J. Jr. and Koch C. F
(1994) in Global Bio-Events and Event-Stratigraphy (O. H.
Walliser, ed.), Berlin, Springer. [4] Sepkoski J. J. Jr.
(1990) GSA Spec. Paper, 247, 33-44. [5] Jablonski D. (1986)
in Patterns and Processes in the History of Life (D. M. Raup
and D. Jablonski, ed.), Berlin, Springer, 313- 329. [6]
Jablonski D. (1994) Phil. Trans. Roy. Soc. London, in press.
[7] Jablonski D. (1989) ibid., B325, 357-368. [8] Raup D. M.
and Jablonski D. (1993) Science, 260, 971-973. [9] Stanley
S. M. (1990) Paleobiology, 401-414. [10] Sepkoski J. J. Jr.
(1989) J. Geol. Soc. London, 146, 7-19. [11] Raup D. M.
(1991) Paleobiology, 17, 37-48.


Alvarez W.*   Asaro F.   Claeys P.   Grajales-N. J. M.   Montanari A.   Smit J.
   Developments in the KT Impact Theory Since Snowbird II

At the second Snowbird conference, entitled "Global
Catastrophes in Earth History: An Interdisciplinary
Conference on Impacts, Volcanism, and Mass Mortality," held
in October 1988, much of the discussion focused on criteria
for choosing between the competing explanations--impact or
volcanism--to account for the KT boundary mass extinction.
Other participants argued that the paleontological record
does not support a rapid mass extinction, and that neither
impact nor volcanism was needed to explain the
biostratigraphic observations.

At Snowbird II, the proponents of volcanism struggled to
explain the physical characteristics of the KT boundary
stratum (shocked quartz, spherules, anomalous iridium), but
pointed with confidence to the Deccan Traps as the volcanic
center responsible for the extinction. In contrast, the
proponents of impact could reasonably explain the physical
features, but had no candidate crater of anywhere near
sufficient size.

As members of the proimpact group, we present the following
account of our perception of developments since Snowbird II,
well aware that supporters of volcanism will see the matter
differently.

Back at the time of the first Snowbird conference, Signor
and Lipps [1] showed that an abrupt extinction studied with
insufficient sampling of the fossils would appear gradual.
Two major animal groups previously said to have died out
gradually have recently been studied with heavy sampling and
careful statistics (ammonites [2] and dinosaurs [3]) and
their disappearences at the KT boundary are now seen to be
indistinguishable from a sudden extinction. These cases are
clear examples of the Signor-Lipps effect, and weaken the
case for a gradual KT extinction.

The strongest improvement in the volcanist position has come
from isotopic dating of the Deccan Traps, which yielded an
age of 64.96 +- 0.11 Ma [4]. However, this date comes from a
late intrusion cross-cutting the lavas, and earlier ages of
68.53 +- 0.16 Ma and 68.57 +- 0.08 Ma were measured on early
volcanic alkalic complexes. Although the main pulse of
volcanism may have occurred close to the KT boundary at 65
Ma, the Deccan Traps seem to have begun erupting well before
the extinction, making a causal relationship less likely.
Nevertheless, the near coincidence in timing is striking,
keeping alive the possibility of some connection between the
Deccan volcanism and the KT boundary event.

It is probably fair to say that the case for explaining the
physical characteristics of the KT stratum by volcanic
processes is no stronger than it was in 1987. Around the
time of Snowbird II, volcanists were reporting the
occurrence of anomalous Ir, spherules, and shocked quartz in
more than one horizon across the KT boundary at Gubbio
which, if true, would falsify the derivation of these
features from a single large impact, and better permit their
attribution to volcanic sources. A blind sample
investigation was organized to test whether these features
are confined to the KT boundary bed alone. Although the
results of the blind experiment have not yet been published,
the results seem less critical now because of the
recognition, in the interim, of a strong candidate crater in
the subsurface of the Yucatan Peninsula, larger than any
other impact crater known on Earth, and apparently dating
precisely from the time of the KT boundary.

About the time of Snowbird II, the unusual clastic beds at
the KT boundary of the Brazos River in Texas began to be
interpreted as tsunami deposits [5], suggesting that the KT
impact site might be located in the Gulf of Mexico-Caribbean
region. Persistent searching [6] yielded various candidate
sites, one of which--Chicxulub--is now widely favored as the
KT impact crater.

The Chicxulub crater was located by PEMEX geophysicists in
the 1950s [7], drilled as a petroleum prospect, and
abandoned when it was found to contain melt rocks that were
interpreted as andesites. Penfield and Camargo [8] proposed
in 1981 that it was an impact crater, but the suggestion lay
dormant for 10 years, until Hildebrand et al. [9] proposed
that Chicxulub was the long-sought KT impact site.

Recently located samples from the old PEMEX wells include
breccias and crystalline rocks compatible with an impact
origin but very difficult to explain by volcanic processes
[10]. Single-crystal ^40Ar/^39Ar dating places the melt rock
at 65.0 Ma [11, 12].

If Chicxulub is indeed a very large impact crater dating
from KT boundary time, it should be surrounded by extensive
deposits of impact ejecta and tsunami deposits. This is
indeed the case, and the expected deposits have been located
in many outcrops from the southern U.S. states of Alabama,
Mississippi, and Texas, through the eastern Mexican states
of Nuevo Leon, Tamaulipas, and Vera Cruz, and the State of
Chiapas in southeastern Mexico, as well as in Haiti and in
DSDP Leg 77 sites 536 and 540, between Yucatan, Cuba, and
Florida. Many of these outcrops contain bubble-rich
spherules with tektitelike morphology in a basal layer,
overlain by a clastic bed with sedimentological
characteristics strongly indicative of deposition by a
large, complex tsunami event [13]. They are constrained by
biostratigraphy to lie precisely at the KT boundary (unless
the boundary is defined so as to force them to be slightly
older [14]).

Many of the tsunami outcrops discussed above contain intense
enrichments of the element iridium above background, a
characteristic found worldwide at the KT boundary in more
than 100 sites. The Ir distribution at Arroyo el Mimbral in
Tamaulipas [15] shows several maxima near the top of the
3.2-m-thick tsunami deposit, in contrast to marine sections
elsewhere that have a main Ir peak near the bottom of a
typically 1-cm KT boundary clay bed. Multiple peaking of the
Ir and interlayering of different sized sediment fractions
like sandstone and siltstone suggest that tidal-wave seiche
activity was not over at the time the iridium peaks were
being deposited, and so the latter may have occurred within
a few days after the impact. This does not mean, a priori,
that the Ir could not have been deposited somewhat earlier,
since it could have been remobilized by the tsunami action.
Calculations have shown, however, that a bolide large enough
to cause the Chicxulub Crater would contain enough Ir to
supply that found worldwide at the KT boundary, and that
impact dynamics would distribute it worldwide. Although
other sources for the KT Ir are possible, the Chicxulub
impact is sufficient to produce all that has been observed,
and to explain its stratigraphic distribution.

Glass has been recovered from the Beloc outcrop in Haiti
[16-18] and the Mimbral outcrop in Tamaulipas [15]. Two
varieties are common: a black glass of roughly continental
crust (or andesitic) composition, and a yellow glass rich in
calcium. The Haiti and Mimbral glasses give single-crystal
40Ar/39Ar ages of 65.0 Ma, exactly the same value as found
for the Chicxulub melt rock. Isotopic studies by Blum and
Chamberlain [19] demonstrated that the Haiti glass (1) has a
strong variation in oxygen isotopic composition that rules
out a volcanic origin, and (2) has an isotopic and major-
element composition that can be explained by melting of
silicate basement and carbonate cover. In a follow-up study,
Blum et al. [20] found an essentially perfect isotopic match
between the Haiti glass and the Chicxulub melt rock.

Developments since Snowbird II have thus strengthened the
case for the impact explanation for the KT boundary. An
impact crater as large as [9] or larger than [21] any other
on Earth has been discovered in the Yucatan Peninsula and
radiometrically dated as of exactly KT boundary age. Ringing
the Chicxulub Crater are clastic deposits apparently due to
a giant tsunami and rich in tektite-shaped objects, some
still containing glass. The glass has the isotopic
characteristics of impact melt, matches the chemistry of
Chicxulub, and gives radiometric dates of KT boundary age,
and the clastic deposits lie at the biostratigraphic KT
boundary. It would be hard to imagine a more dramatic
confirmation of the essential prediction of the KT impact
theory, that an appropriate crater should be found.
Nevertheless, some recent authors totally reject this view
[22-25]. These last papers represent the skepticism required
for testing any scientific interpretation, but even though
the testing is not yet complete, the impact theory must be
considered to have been notably strengthened since Snowbird
II. It is now essential to explore the Chicxulub Crater with
the drill.

Dedication: This review is dedicated to the memory of S. V.
Margolis. Stan contributed greatly to the developments
described here, and we all miss him.

References:  [1] Signor P. W. and Lipps J. H. (1982) GSA
Spec. Paper 190, 291-296. [2] Ward P. D. et al. (1991)
Geology, 19, 1181-1184. [3] Sheehan P. M. et al. (1991)
Science, 254, 835-839. [4] Basu A. R. et al. (1993) Science,
261, 902-906. [5] Bourgeois J. et al. (1988) Science, 241,
567-570. [6] Hildebrand A. R. and Boynton W. V. (1989) GSA
Abstr. with Progr., 21, A371; Hildebrand A. R. and Boynton
W. V. (1990) LPSC XXI, 512-513; Hildebrand A. R. and Boynton
W. V. (1990) GSA Abstr. with Progr., 22, A280; Hildebrand A.
R. and Boynton W. V. (1990) Eos, 71, 1424-1425; Hildebrand
A. R. and Penfield G. T. (1990) Eos, 71, 1425. [7] Cornejo
A., and Hernandez A. (1950) Bol. Asoc. Mex. Geol.
Petroleros, 2, 453-460. [8] Penfield G. T. and Camargo Z. A.
(1981) Soc. Expl. Geophys. Tech. Progr., Abstr., Biogr., 51,
37. [9] Hildebrand A. R. et al. (1991) Geology, 19, 867-871.
[10] Cedillo E. et al., this volume. [11] Swisher C. C. III
et al. (1992) Science, 257, 954-958. [12] Sharpton V. L. et
al. (1992) Nature, 359, 819-822. [13] Smit J. et al., this
volume. [14] Smit J. et al., this volume. [15] Smit J. et
al. (1992) Geology, 20, 99-103. [16] Izett G. A. et al.
(1990) U.S. Geol. Surv. Open-File Rept. 90-635, 1-31. [17]
Sigurdsson H. et al. (1991) Nature, 349, 482-487. [18]
Maurrasse F. J.-M. R. and Sen G. (1991) Science, 252, 1690-
1693. [19] Blum J. D. and Chamberlain C. P. (1992) Science,
257, 1104-1107. [20] Blum J. D. et al. (1993) Nature, 364,
325-327. [21] Sharpton V. L. et al. (1993) Science, 261,
1564-1567. [22] Lyons J. B. and Officer C. B. (1992) EPSL,
109, 205-224. [23] Jehanno C. et al. (1992) EPSL, 109, 229-
241. [24] Keller G. et al. (1993) Geology, 21, 776-780. [25]
Stinnesbeck W. et al. (1993) Geology, 21, 797-800.


Sigurdsson H.*
   Environmental Consequences of Volcanic Eruptions and Meteorite Impacts

No abstract available.


Ward P. D.*
   Anatomy of the KT Extinction Event

No abstract available.


Olsen P.*
   Vertebrate Behavior Under Crisis:  The Fossil Record

No abstract available.





                          Wednesday, February 09, 1994
                   SESSION 2: CHICXULUB AND OTHER KT IMPACTS
                                 2:30 - 6:00 PM

Chair(s):  G. Ryder
           G. Izett


Camargo A.*
   Historical Overview of the Chicxulub Crater

No abstract available.


Sharpton V. L.*   Marin L. E.   Schuraytz B. C.
   The Chicxulub Multiring Basin: Evaluation of Geophysical Data, Well Logs,
   and Drill Core Samples

In 1981, Penfield and Camargo [1] reported concentric
gravity and magnetic anomalies over a region in northernmost
Yucatan where unusual occurrences of crystalline rocks had
been reported from wells drilled into the carbonate
platform, and they proposed that these observations could
signal a buried impact structure. Over the last few years,
as eyes turned toward the Gulf of Mexico region in search of
the KT impact site, support has mounted for this Yucatecan
crater, now known as the Chicxulub structure [2-9]. Although
uncertainties still exist as to its exact size and how it
affected the global biosphere, it is now clear that the
Chicxulub impact crater is the source of the worldwide
ejecta layer associated with the Cretaceous-Tertiary
extinction event [2,6,8,10,11]. Here we present observations
from geophysical data, well logs and drill core samples that
indicate the Chicxulub structure is a multiring impact basin
~300 km diameter, and propose that it is the sole source of
the impact debris laid down at KT boundary layer.

Lithologic Evidence of Impact. Three Petroleos Mexicanos
exploratory wells are located near the center of the
concentric gravity anomalies over the Chicxulub structure
(Fig. 1) and all intercepted breccias and crystalline
silicate rocks at depths of ~1 km below sea level [e.g., 12,
13]. Samples from the Yucatan 6 and Sacapuc 1 wells include
suevite, a polymict, clastic-matrix breccia (Fig. 2)
analogous to impact-produced breccias observed at other
terrestrial craters such as the Ries Crater in Germany. The
Chicxulub suevite contains clasts of silicate basement and
platform rocks showing abundant and unequivocal evidence of
shock metamorphism [6], including: (1) planar deformation
features in quartz, feldspar, and zircons, indicating
dynamic pressures up to 23 GPa; (2) shock mosaicism in
quartz and feldspar grains; (3) diaplectic glasses such as
maskelynite (30-45 GPa); (4) fused mineral glasses,
particularly of alkali feldspar, indicating shock pressures
above 45 GPa, and (5) impact melts of whole rock composition
(>60 GPa).

Below the deposits of suevite (Fig. 2) are thick sequences
of impact melt and melt-matrix breccia [6], similar to the
clast-free and clast-rich melt zones observed at other
meteorite impact craters such as Manicouagan, Quebec. Melt
samples show ubiquitous evidence of chemical and thermal
disequilibrium and super heating as is characteristic of
impact melt rocks. Undigested basement clasts retain
indications of shock metamorphism but are usually
overprinted with subsequent thermal effects such as
recrystallization and annealing. Elevated siderophile
element abundances in some melt rock samples [6,7,9] we have
studied indicate a meteoritic component nonuniformly
distributed throughout the melt sheet.

Outside the central region of buried crystalline rocks in
the Chicxulub structure (Fig. 2) wells intercepted thick
sequences of breccia composed mainly of anhydrite and
carbonate clasts from the Mesozoic platform sequence. These
breccia deposits are similar to the Bunte Breccia at the
Ries Crater in that they represent primarily the uppermost
target lithologies and show modest evidence for shock
metamorphism.

The Size and Morphology of the Chicxulub Impact Basin.
Penfield and Camargo [1] originally recognized two
concentric zones in gravity anomaly and aeromagnetic data:
an inner zone of ~60 km diameter characterized by a gravity
high and high frequency magnetic anomalies approaching 1000
nT, and an outer gravity trough with low amplitude (5 to 20
nT) magnetic anomalies. Primarily on the basis of these
patterns Hildebrand et al. [2] proposed that the Chicxulub
crater was a double-ring, or peak-ring basin with a rim
diameter of ~180 km.

More recently, however, the gravity data over the Northern
Yucatan have been reprocessed [14] and additional concentric
patterns outside the original two have been identified
(Figs. 1 and 3). The outermost ring is located at a diameter
of ~300 km. Several arguments have been presented suggesting
that this ring corresponds to the modified rim of the
Chicxulub multiring basin [14]. Of particular importance is
the observation that gravity values increase abruptly and
significantly at the inner flank of the ~200 km diameter
ring (Ring 3 in Figs. 1 and 3), previously interpreted as
the rim crest. Such steep gradients define the edge of a
major mass deficiency in the basin center. We believe that
this is more consistent with the deep transient crater
boundary than with the rim crest, which is the outer edge of
a broad zone of modest and shallow deformation located well
outside the deep transient crater. The fourth ring is
probably not an exterior ring (i.e., located outside the
topographic rim), as observed at some large multiring basins
on other planets, because it has a detectable gravity
expression and because of its association with the edge of
the basin's circular gravity low.

There are additional observations that also indicate that
Chicxulub may be ~300 km in diameter. Well log and sample
analysis data have been used to compile the stratigraphic
correlations in the vicinity of the Chicxulub basin shown in
Fig. 2. These data indicate that the boundary between the
platform sediments and the crystalline basement is depressed
~750 m in wells within 150 km of the basin center. Other
major stratigraphic boundaries are depressed as well. In
addition, thick sequences of breccia are observed in each of
these wells. The Ticul 1 well, closest to the intermediate
ring (Ring 3 in Figs. 1 and 3) contains the thickest breccia
unit, even though previously published reports based only on
drill cuttings show no breccias at all [e.g., 12,13].

Major element chemistry [2,5,6] and isotope analysis [11]
indicate that the Haitian impact glass spherules and the
Chicxulub melt rocks are identical and could be formed by a
mixture of 94% silicate basement and 6% platform carbonate
rocks [11]. Given an average thickness of ~2 km for the
platform cover in this region [12], and allowing for
vaporization of carbonates near the impact point, silicate
basement would have to be melted to a minimum depth of 15 to
20 km to provide these proportions. To the degree that depth
of melting constrains the minimum depth of the excavation
cavity d(sub)exc then the transient crater diameter D(sub)t
can be estimated [15] from the equation d(sub)exc ~= 0.1
D(sub)t or, for Chicxulub, D(sub)t ~= 150 to 200 km. The
final crater diameter D resulting from the collapse of the
transient cavity follows from the relationship D(sub)t ~=
0.5 to 0.65 D [15]. This predicts D ~= 300 km for the
Chicxulub crater, in good agreement with the stratigraphic
data from well logs and the geophysical data.

Finally, crater studies show that even at the smallest
possible final rim crest diameter of 170 km, the Chicxulub
basin would not likely be a peak ring basin. Planetary
datasets show a clear progression in the morphological
classes of large craters with increasing size and that the
transition diameter from one class (such as peak ring basin)
to the next (multiring basin) depends strongly on surface
gravity. The planet whose surface gravity and other
properties are most similar to Earth is Venus. Evaluation of
over 1000 impact craters detected by the Magellan space
craft reveal that double ring, or peak ring basins, are
constrained to diameters between 40 km and 110 km [16]. All
larger craters are multiring basins, i.e., basins
characterized by 3 or more concentric rings. This, coupled
with the evidence above, suggests that the previous model of
Chicxulub as a ~180-km diameter peak ring basin is
incorrect.

Consequently, we believe the interpretation most consistent
with all available information is that the Chicxulub
structure as a ~300-km-diameter multiring basin (Fig. 4)
similar to the largest impact landforms observed on the
Moon, Mercury, and Venus [14]. Such events are extremely
rare; there is only one other impact basin of comparable
size produced in the inner solar system within the last
billion or so years: the 280-km Mead Basin on Venus.

The Chicxulub Multiring Basin and the KT Boundary. Melt
rocks within the Chicxulub crater have experienced varying
levels of hydrothermal alteration and albitization [17];
however ^40Ar-^39Ar determinations on relatively pristine
melt rock samples from the center of the basin indicate a
crystallization age at or very near the KT boundary [6,10].
Evaluation of the magnetization of these and other samples
show that they cooled during an episode of reversed
geomagnetic polarity, consistent with a KT boundary age
[6,18]. Analysis of the Rb-Sr, O, and Nd-Sm isotopic systems
confirm a chemical link between the Chicxulub melt rocks and
the impact glasses contained in the KT boundary deposits at
Beloc, Haiti [11]. Unmelted breccia clasts, representing the
silicate basement impacted by the Chicxulub basin forming
event, are medium- to high-grade continental crust similar
to the lithic clasts observed within KT boundary sediments
from around the world [6,19].

Krogh et al. [8] recently determined U-Pb ages of zircon
xenocrysts from a Chicxulub breccia sample are ~545 Ma, with
a small percentage showing a 418 Ma age. These 545 Ma old
zircons match the crystallization age of those within the
upper member (magic layer, fireball layer, etc.) of the KT
doublet in the Raton Basin of New Mexico. This provides
compelling evidence that the Chicxulub impact event is
responsible for both the shocked minerals observed in the
upper member of the KT boundary sequence and the glass rich
lower member. Consequently all evidence to date points to a
singular, extremely energetic, and deadly, impact event at
the KT boundary.

References: [1] Penfield G. T. and Camargo-Z A. (1981) Soc.
Explor. Geophys. 51st Ann. Meeting, Tech Prog. 37. [2]
Hildebrand A. R. et al. (1991) Geology, 19, 867. [3] Pope
K.O. et al, (1991) Nature ,351, 105. [4] Quezada-M. J. M. et
al. (1992) LPSC XXIII, 1121. [5] Kring D. A. and Boynton W.
V. (1992) Nature 358, 141. [6] Sharpton V. L. et al. (1992)
Nature 359, 819. [7] Koeberl C. et al. (1993) Geochim.
Cosmochim. Acta, submitted. [8] Krogh, T. et al.(1993)
Nature, submitted. [9] Schuraytz B. C. and Sharpton V. L.,
this volume. [10] Swisher C. C. et al. (1992) Science, 257,
954. [11] Blum J. D. et al. (1993) Nature, 364, 325. [12]
Lopez-Ramos E. (1993) Geologia de Mexico, UNAM Press. [13]
Weidie A. E. (1985) Geology of Yucatan Platform, in Geology
and Hydrogeology of the Yucatan, (W. C. Ward, et al, eds.),
NOGS Publications. [14] Sharpton V. L. et al. (1993)
Science, 259, 1564. [15] Melosh H. J. (1989) Impact
Cratering, Oxford Press. [16] Schaber G. C. and Sharpton V.
L. (1993) Nature, 362, 503. [18] Urrutia-Fucugauchi J. et
al.(1993) Tectonophysics submitted. [19] Sharpton V. L. et
al. (1990) Geol. Soc. America Spec. Paper, 247, 349.

Fig. 1, which appears here in the hard copy, shows surface
units, ring locations and wells in the vicinity the
Chicxulub multiring impact basin. See Sharpton et al. [6]
for details.

Fig. 2, which appears here in the hard copy, shows
stratigraphic correlations across the Chicxulub basin based
on well logs and sample analysis. Wells are arranged in
distance from the basin center. The samples we have studied
are identified by the numbered arrows. Other lithological
assignments are based on visual inspection of cores updating
previous published and unpublished reports. Ring
designations refer to the three most prominent rings in the
gravity analysis of Sharpton et al. [6].

Fig. 3, which appears here in the hard copy, shows gravity
profiles taken across the Chicxulub basin at 10 degrees
intervals of azimuth. D (sub)t and D (sub)a refer to the
transient crater diameter and final or apparent crater
diameter resulting from this analysis. Profiles are offset
vertically by 10 mgal; the annotated vertical axis shows the
0 mgal value for the designated profiles, beginning from the
bottom with N to S. Dashed profile is due E to West. The
value of the center point of each gravity profile is 10.4
mgal.

Fig. 4, which appears here in the hard copy, shows a
schematic model of the Chicxulub impact basin that we derive
from this analysis. This simplified cross section shows the
general configuration of the crater but does not consider
erosion; erosion at the time of impact could rearrange the
upper crater units significantly and reduce crater
topography. Faults and unit boundaries are simplified.


Cedillo P. E.   Claeys P.   Grajales-N. J. M.   Alvarez W.
   New Mineralogical and Chemical Constraints on the Nature of Target Rocks at
   the Chicxulub Crater

The Chicxulub Crater melt rocks are being found to display
striking mineralogical and chemical features. The most
important mineralogical and textural features are related to
(1) partial melting of the rock and the preexisting minerals
(quartz, plagioclase, and anhydrite), and (2) textural
features developed in pyroxenes, feldspars, and magnetite
that provide clues to the crystallization behavior of the
melts. Chemical analyses of minerals and whole rocks display
a wide range of compositions. Core samples used for the
present study come from the Yucatan-6 (Y-6) and Chicxulub-1
(CH-1) wells.

Preexisting minerals exhibit numerous features indicative of
partial melting. In some shocked quartz grains, this process
has destroyed lamellar features, while in others the
lamellae were only partially destroyed. Quartz xenoclasts
show a rim of melted and recrystallized quartz surrounded by
a reaction corona of pyroxenes, similar to those found in
the Manicouagan impact melts [1]. Plagioclase crystals show
sievelike melting textures consisting of a network of glassy
material following crystallographic directions. Textural
evidence also supports partial melting of anhydrite; the
physical-chemical conditions under which this process took
place are not completely understood.

Skeletal crystals of pyroxene, feldspar, and ilmenite
provide clues to the crystallization behavior of the melts.
The pyroxene is found as small crystals (10 micrometers),
some only partially crystallized, which indicates
supercooling. It is also common to find larger crystals of
pyroxene (70 micrometers) as reaction coronas around quartz.
This texture shows competition for silica in the melt. The
chemical composition of pyroxenes is characterized by high
calcium content; the analyses plot outside the normal
pyroxene quadrilateral, to the calcic side of normal
diopside. Although X-ray diffraction lines of melt pyroxene
are similar to those of diopside, its composition is closer
to fassaite. The abundance of Ca-rich pyroxenes in the melt
rocks is strong evidence for the participation of platform
carbonate sediments in the genesis of melts at Chicxulub.

Feldspars were formed from glass derived from the melting of
basement rocks. The chemical composition ranges from albite
to andesine, supporting a variety of sources for the glass,
as pointed out earlier [2-6]. Chemical analyses of glasses
show the same variation in composition as their associated
plagioclases, indicating that they have a common origin.

The basement rocks were the source of silica for the
pyroxenes and plagioclases, as shown by xenoclasts found in
the melts of Y-6 [4] and by the abundance of shocked quartz
grains of igneous and metamorphic origin found in KT
boundary sediments [7]. The ilmenite shows skeletal growth
indicative of supercooling, which is also supported by the
absence of exsolution.

In summary, the mineralogy indicates that two groups of
rocks (platform carbonates and basement) were involved in
the genesis of melt rocks at Chicxulub. The relative
contribution of the two sources determines the final
composition of melt rocks. On the other hand, the abundance
of xenoclasts in the melt rocks played an important role in
the thermal behaviour of the melts, favoring a greater
cooling rate. This is the case of melts found in the samples
of Y-6, in contrast to those of Chicxulub-1. Work in
progress is focused on the chemical composition of glasses
in order to constrain the importance of target rocks as
source for melt rocks, vaporization products, and tektites
of the KT boundary.

Acknowledgments: The authors wish to thank the permanent
support of PEMEX and the Instituto Mexicano del Petroleo. We
also express our gratitude to the late Dr. S. V. Margolis,
who participated in the microprobe analytical work at the
beginning of the investigation.

References: [1] Floran R. J. et al. (1978) JGR, 83, 2737-
2759. [2] Swisher C. C. III et al. (1992) Science, 257, 954-
958. [3] Cedillo-Pardo. E. et al. (1992) GSA Abstr. with
Progr., 24, A333. [4] Quezada J. M. et al. (1992) LPSC
XXIII, 1121-1122. [5] Sharpton V. L. et al. (1992) Nature,
359, 819-821. [6] Blum J. D. and Chamberlain C. P. (1992)
Science, 257, 1104-1107. [7] Owen M. R. et al. (1990) GSA
Spec. Paper, 247, 343-347.


Ivanov B. A.*   Badukov D. D.   Yakovlev O. I.
   Shock Degassing of Sedimentary Rocks Due to Chicxulub Impact: Hydrocode
   Simulation

We present the progress report of the work concerned the
scenario of the Chicxulub impact event. The work includes
both 2D hydrocode simulation of the impact and study of
geochemistry of shocked rocks.

The specific feature of the target geology at the Chicxulub
site is the presence of 2 to 3 km thick layer of sedimentary
rocks over the crystalline basement. High contents of
anhydrite (CaSO4) may causes the production of SO2 gases due
to shock heating of rocks [1]; CO2 production from
carbonates also may impact the terrestrial environment at
the KT boundary [2].

Our computer simulation shortly reported earlier [3] is now
in progress to model various stages of the Chicxulub event.
All simulations use the supposed rim diameter of the crater
of 180 to 300 km. To create such a crater the silicate
asteroid with a diameter of 10 to 20 km would strike Earth
at the most probable for asteroids impact velocity of 20
km/s [4]. Variations of asteroid's size, composition, and
impact velocity give an intrinsic uncertainty for all
estimates of a single event. An oblique impact may result in
addition complex phenomena [5].

To make first simple estimates for the future geochemical
modeling we calculate the vertical impact of a cylindric
asteroid into the double layer target. Such a geometry
allows us to use the relatively simple SALE code [6]. The
problem to use this code for the Chicxulub event is to
calculate the shock wave propagation in the target with a
top 3 km layer, which is thin in comparison with a
projectile diameter (10 to 20 km). To resolve the shock
front in the hydrocode with an artificial viscosity we need
to have at least 10 rows of cells to describe the
sedimentary layers. So the cell dimension is limited within
the range of a few hundreds of meters. In this case we need
to cover the projectile with several tens of cells. This
mean that the shape of the projectile may play an important
role in the close to projectile zones.

We run several variants of the code varying the equation of
state of target layers. Tillotson EOS has been used with a
minor modification in the expansion range. To model dynamic
properties of the sedimentary layer we used constants for
wet tuff and limestone, and aluminum and granite constants
for the crystalline basement cited by Melosh [7]. The first
output of the simulation is the estimate of a sedimentary
rock volumes compressed above a given shock pressure. The
geometry of isobars has been also mapped. For all our
calculations the volume of sedimentary material shocked up
to pressures 30 to 60 GPa (the typical for the post-shock
thermal decomposition of carbonates) and higher vary in the
range 500 to 1500 km^3. These values may be used as an upper
limit for all scenario of the consequent effect of impact
produced gases to the terrestrial environment.

It is important to note that more than half of this volume
is the "plug" of sediments compressed by a projectile. This
material will be released to the atmosphere relatively
lately in a mixture with a projectile material and may be
with crystalline material. Other part of a highly shocked
sediments may be involved in the early high speed ejecta. So
the scenario of transport of impact-produced gases to the
atmosphere and stratosphere may have several consequent
stages, which should be investigated separately.

Amount of gaseous phases formed by impact devolatilization
of sulfur-bearing minerals such as anhydrite and gyps should
depend on two causes, which act in opposite ways: (1) a
shock-induced heating and (2) an effect of reverse reactions
after decompression and devolatilization. The degree of the
thermochemical decomposition can be obtained as a function
of amount of post-shock specific internal energy provided
that: (1) the reaction of the decomposition takes place only
after decompression, (2) the system is thermodynamically
closed at the ambient pressure of 1 bar, and (3) all
processes in the system are in equilibrium. According to our
calculation incipient outgassing of anhydrite starts by 90
GPa and is completed by 155 GPa. Violation of an
equilibrium, for example due to fast expansion, may change
simple equilibrium estimates.

One must consider this estimation of as a maximum one
because of high ability of new formed CaO for reaction with
SO3. This reverse reaction can take place in cooling ejecta
plumes. On the other hand the completeness of the reverse
reaction can be restricted due to removal of the gas from
plumes and slow rate of this heterogeneous reaction.
Laboratory shock experiments by high parameters can give
constrains on the suggestion. This work is now in progress.

At the next step of the work we plan to use calculated
evolution of various elements of shocked materials to
estimate the possible kinetics of chemical reactions for the
better constraint of the residual composition of impact-
produced gases.

References. [1] Brett R. (1992) GCA, 56, 3603-3606. [2]
O'Keefe J. D. and Ahrens T. J. (1989) Nature, 338, 247-249.
[3] Pope K. O. et al. (1993) LPS XXIV, 1165-1166. [4] Roddy
D. J. et al. (1987) Int. J. Impact Eng., 5, 525-541. [5]
Schultz P. H. and Gault D. E. (1990) GSA Spec. Paper, 247,
239-262. [6] Amsden A. A. et al. (1980) LA-8095, Los Almos,
NM, 100. [7] Melosh H. J. (1989) Impact Cratering, Oxford
Press, 245.


Takata T.*   Ahrens T. J.
   Numerical Simulation of Impact Cratering at Chicxulub and the Possible
   Causes of KT Catastrophe

The chicxulub crater, located on the Yucatan peninsula is
considered as the most likely site of the impact crater,
which was formed by an extraterrestrial object that appeared
to have caused the catastrophic extinction at the KT
boundary [1]. Recent investigations of the gravity field in
this area extended the possible diameter of the crater from
180 km, which had been previously estimated in Hildebrand et
al. [1] to 260-340 km [2]. In order to create a 260-340 km
diameter crater, the impacting object diameter required is
20 km to 30 km, if the meteorite impacted at a velocity of
20 km/s. The impact of the meteorite of this diameter, that
is 2 to 4 times of scale height of the Earth atmosphere,
could result in global effects on the ecology, environment,
and trigger climate changes that may last ~10^4 years.

Geological investigations indicate that the Chicxulub crater
inferred to have been excavated shallow sea platform
containing 2-4 km thick carbonate and evaporitic rocks
(mainly limestone, dolomite, and gypsum) overlying the
basement rock of andesite at KT time (65 Myr) [3]. This
topographical and geological environment for the Chicxulub
crater is expected to be a typical feature of KT impact
craters. This is because of the high ratio of sea surface to
the continental surface, and the warmer climate of
Cretaceous age. These will result in thick carbonate rock
sequences. The production and release of the water vapor and
CO2 from the terrestrial surface to the atmosphere is
expected during an impact event. The increase of the water
vapor and CO2 in the atmosphere may have resulted in
increasing the global temperature of the atmosphere via the
green house effect [4,5].

Moreover the enormous amount of fine dust ejected from the
crater into the atmosphere and its subsequent global
dispersal in the stratosphere could have possibly decreased
solar insolation inducing a short-duration decrease in
global temperature as first proposed by Alvarez et al. [6].
Partial destruction of the ozone layer in the upper
atmosphere may have also been a factor in extinction of some
species. The possible scenarios including breakdown of the
usual annual hemispheric climate cycle due to the
combination of the above mechanisms have been discussed for
a decade. What mechanisms caused extinction of what
organisms have not been conclusively demonstrated. One
approach to this complex problem is to examine the mechanics
of impact cratering in the Chicxulub crater lithology and
estimate the possible environmental or climatic
consequences.

The impact cratering process on the lithology of the
Chicxulub impact site was investigated by a numerical
simulation. We used the Smoothed Particle Hydrodynamics
(SPH) numerical technique [7]. The SPH method is a free
lagrangian method, and suitable for the highly distorted
motion of materials. All the materials are expressed by
particles that satisfy the hydrodynamic equations and are
treated like N-body problems. In our simulations,
axisymmetric geometry is applied in order to expand the
region taken into account in the calculations.

In order to simulate the response of the lithology at KT
time around the Chicxulub crater site in our code, we
assumed a target with basement rock of andesite, overlaid by
3 km thick-carbonate rock layer, and an atmosphere. A
Tillotson equation of state for limestone is applied for
carbonate, and an ideal gas with constant specific heat
ratio is used for the atmosphere. All the target materials
including atmosphere is initially stabilized and is in
hydrostatic equilibrium. The radius of meteorite is assumed
to be 12.5 km. The advantage of the SPH method is that it
can represent the transfer of energy and interaction of
materials among impacting meteorite, the surface materials,
and the atmosphere. Numerical simulations have been employed
since publication of Alvarez's hypothesis [8,9], however,
the uniqueness of our calculation is that we include an
atmosphere in addition to the carbonate systems.

Figures 1 and 2 show position of materials and particle
velocity at 16 seconds after the impact. We can observe that
the ejecta curtain extends to more than one scale height of
the atmosphere and the mixing of carbonate, basement rocks,
and impactor in the ejecta within the atmosphere. Drilling
data from the center of Chicxulub crater demonstrated the
melt sheet was observed beneath a 200-400 m thick breccia
layer. The results of the calculation show that
approximately 10^1 km thick melt sheet is produced. The
total mass of melt is approximately 20 times the impactor
mass.

The energy transfer to the atmosphere takes place via two
mechanisms: (1) Direct impact of the bolide on the
atmosphere accounts for very few percent of the atmospheric
heating and (2) The work done on the atmosphere by the
moving ejecta curtain transfers the bolide energy to the
atmosphere. When the ejecta curtain of target materials
starts to develop after the impact of meteorite, the total
amount of energy transfered to the atmosphere is less than
0.5%, that is, ~20^22 J. However the ejecta transfers an
enormous amount of energy into the atmosphere due to the gas
drag [8]. As a result of the impact, carbonate rock takes up
approximately 10% of the impact energy, 5% of the initial
energy was left in the meteorite as the internal heating,
and the rest of the energy is transfered to the basement
rock.

The amount of CO2 production by the shock-induced
decarbonation of the carbonate rocks is also estimated. The
decarbonation of calcite was observed experimentally to
occur at 20 GPa [10]. Young carbonate rocks should have had
a certain degree of porosity because they were recently
deposited. In the laboratory, shock experiments using 50%
porous chalk induces devolatilization of 90% of the
available CO2 at 10 GPa [11]. Therefore, we assumed that all
the carbonate pressurized to more than 10 GPa devolatilizes.
By determining the mass of carbonate exposed to >10 GPa, the
production of CO2 by Chicxulub cratering is found to be
10^16-10^17 kg. This mass of CO2 is 10 to 100 times the CO2
inventory in the present atmosphere. This is a greater
increase in CO2 than obtained previously by O'Keefe and
Ahrens [5].

The mixing ratio of impact to target material ejected into
the atmosphere is approximately 0.01 to 0.02. This value
agrees with this ratio (0.01 to 0.1) found in the worldwide
platinum-rich element ratio at the KT boundary [8]. We
calculate that all the meteorite material ejected to the
atmosphere is melted or vaporized.

In conclusion, the production of the enormous amount of CO2,
dust of terrestrial and extraterrestrial origin was
confirmed in our numerical simulation of Chicxulub crater.
The climate system should have been affected by the impact-
induced H2O and CO2 and dust on various time scales. Heating
of the atmosphere probably produces higher water and CO2
vapor pressures [4,5]. If a water cloud expanded globally
this may have increased the albedo of the Earth on a several
month time scale. These albedo-temperature feedback systems
may have been disordered by abrupt changes of constituents
of atmosphere.

Acknowledgements: We appreciate the use of the CRAY Y-MP at
JPL for these calculations.

References: [1] Hildebrand A. R. et al. (1991) Geology, 19,
867-872. [2] Sharpton V. L. et al. (1993) Science, 261,
1564-1567. [3] Swisher et al. (1992) Science, 257, 954-958.
[4] Emiliani C. et al. (1981) EPSL, 55, 317-334. [5] O'Keefe
J. D. and Ahrens T. J. (1989) Nature, 228, 247-249. [6]
Alvarez et al. (1980) Science, 208, 1095-1108. [7] Gingold
R. A. and Monaghan J. J. (1977) Month. Not. Astr. Soc., 181,
375. [8] O'Keefe J. D. and Ahrens T. J. (1982) GSA Special
paper, 190, 103-120. [9] Roddy D. J. et al. (1987) J.
Impact. Eng., 5, 525-541. [10] Lange M. A. and Ahrens T. J.
(1983) LPSC 14, 419-420. [11] Tyburczy J. A. and Ahrens T.
J. (1986) JGR, 91, 4730-4744.

Fig. 1. Particle position during impact Chicxulub at t=16
[s] after the impact. The symbols of the particles, bullet,
triangle, +, and times, represent impactor, basement rock,
carbonate rock, and atmospheric particles, respectively. The
impactor size is 25 km and the impact velocity is 20 km/s.

Fig. 2. The velocity field of the impact cratering at the
same time of Fig. 1.


Robin E.*   Gayraud J.   Froget R.
   On the Origin of Regional Variations in Spinel Compositions at the KT
   Boundary

Introduction: Spinels with morphologies and compositions
quite similar to meteoric spinels (the term "meteoric" that
we use here is defined in reference [1]) have been found
throughout the world at the Cretaceous-Tertiary (KT)
boundary. Their presence and their high abundance is a
strong argument in favor of a major extraterrestrial
accretion event at the end of the Cretaceous. These minerals
display a wide variety of morphology (octahedral, dendritic,
skeletal, cruciform,...) and composition [2] which likely
reflect differences in their formation conditions [1-4].
Here we discuss on the cause of the regional variations in
spinel compositions and on the implications to the
understanding of the KT event.

Samples and Procedure: We have searched for meteoric spinels
at 18 KT sites: Sopelana and Caravaca in Spain, Bidart in
France, Stevn Klint in Danemark, Gosau in Austria, Petriccio
and Nago in Italy, El Kef in Tunisia, Brazos River in Texas
(USA), Beloc in Haoti, Mimbral in Mexico, DSDP site 536 in
the Gulf of Mexico, DSDP sites 577 and 465 respectively in
North and South Pacific Ocean, DSDP site 524 and 527 in
South Atlantic Ocean and ODP site 761C and 752 in the Indian
Ocean. Spinels were extracted from sediment samples with a
magnetic separator and recovered on a 0.4 5m nuclepore
filter (see reference [5] for a description of the method).
The filters were then mounted in epoxy and polished for
detailed scanning electron microscopy and electron
microprobe analysis of individual spinels (> 2 micrometers).

Results and Discussion: Meteoric spinels are found at nearly
all sites except at Stevn Klint, Brazos River and DSDP site
465 where their concentration is below detection limit (<
0.1/mg). At other sites, the abundance of spinels varies
from ~ 0.3/mg at Mimbral up to 3000/mg at ODP site 761C.
Their occurrence is generally associated with the maximum of
the Ir anomaly, in coincidence with a sharp decrease of the
carbonate fraction. Results of the average composition of KT
spinels are given in the table except for ODP site 752, DSDP
sites 524 and 527 for which spinels were not analyzed so
far. Several remarks can be done from these results
considering the formation conditions of meteoric spinels
[1]:

KT spinels have high Ni content (>1%) and high Fe^3+/Fe
(sub)total ratio (>70%) showing that they were formed from
undifferentiated material in a very oxidizing environment
[1,3]. These formation conditions are very constraining and
allow to rule out some possible sources for KT spinels. They
cannot derive from terrestrial magmas which are strongly
depleted in Ni and evolve under extremely low oxygen
fugacity [1,3]. Direct condensation of spinel crystals
during continental cratering events is very unlikely because
of the reducing conditions that prevailed in the impact
vapour cloud [4]. The only process known to form spinels
with such a composition is ablation and oxidation of
meteoritic debris in the atmosphere [3]. Spinels from
Pacific and Indian Oceans have an Fe^3+/Fe (sub)total ratio
>90% which require for their crystallization an oxygen
fugacity >0.21 atm or higher than the one existing at sea
level [1]. Although no definitive explanation exists today
[1], such highly oxidized spinels could be indicative of
oceanic impacts [4,6].

Large variations in spinel compositions are observed from
site to site, even for nearby sites. For example, there is
virtually no overlap in the chromium content of spinels
between Italian sites (Petriccio, Nago) and Tunisian site
(El Kef) (see figure 1). Variation in the Cr content of
spinels is also observed between Sopelana (Spain) and Bidart
(France) which are distant by only 200 km. It implies that
very heterogeneous formation conditions prevailed on a local
geographical scale. At the other hand, numerous cosmic
debris containing Al-rich spinels of quite homogeneous
composition are found over at least 2x10^7 km^2 in the
Pacific Ocean [7]. This suggests that more than one object
struck the Earth at the end of the Cretaceous. The largest
ones in the Pacific to account for the wide dispersion area
of Al-rich spinels and smaller ones over Europe and North
Africa to account for the variation in spinel composition on
a local scale.

Spinels from Pacific and Indian Oceans are always more
oxidized than those from European and African sites (see
figure 2). Spinel from the Carribean area seems to have an
intermediate oxidation state though spinel from DSDP site
536 are highly oxidized. The same tendancy is observed for
their Cr content: spinel from Pacific and Indian Oceans have
lower Cr content than those from European and African sites.
Only spinels from Italy have a relatively low Cr content.
The fact that the Cr content and the Fe^3+/Fe (sub)total
ratio of meteoric spinels is controlled by the oxygen
fugacity [1] suggests that spinels from Pacific and Indian
Oceans were formed under higher oxygen fugacities than those
from european and african sites. This is consistent with the
hypothesis of large objects impacting in the Pacific and
Indian Oceans and smaller ones in Europe and North Africa.
Indeed, large objects decelerate at low altitudes and thus
at high oxygen partial pressures while small objects
decelerate at higher altitudes and thus lower oxygen partial
pressures. If so, it suggests a preferential arrival of
cosmic bolides in the hemisphere containing the Pacific and
Indian Oceans.

Conclusions: (1) The high abundance of meteoric spinels in
most KT samples investigated here confirms that a major
cosmic accretion event occurred at the end of the
Cretaceous. (2) The wide range of spinel composition
observed from site to site indicates that several objects
have impacted the Earth at this epoch rather than a single
one. Multiple objects could derive from a single disrupted
bolide. (3) The observed variation of the Fe3+/Fetotal ratio
of KT spinels from Pacific to Europe suggests that the
largest objects would have impacted in the Pacific and
Indian Oceans and smaller ones in Europe and North Africa.
This would also explain the homogeneous spinel composition
over large area in the Pacific and the heterogeneous spinel
composition on a local scale in Europe and North Africa.

References: [1] Gayraud J. et al. (1994) this volume. [2]
Kyte F. T. and Smit J. (1986) Geology, 14, 485-487. [3]
Robin E. et al. (1992) EPSL, 108, 181-190. [4] Robin E. et
al. (1992) LPS XXIII, 1161-1162. [5] Robin E. et al. (1991)
EPSL, 107, 715-721. [6] Margolis S. V. et al. (1991)
Science, 251, 1594-1597. [7] Robin E. et al. (1993) Nature,
363, 615-617.

Table 1, which appears here in the hard copy, shows Aberage
compostion of KT spinels.

Fig. 1, which appears in the hard copy, shows a histogram
showing the spinel chromium content.

Fig. 2, which appears in the hard copy shows a histogram
showing the Fe^3+/Fe (sub)tot of KT spinels.



                              POSTER PRESENTATIONS

Schuraytz B. C.   Sharpton V. L.
   Siderophile Element Distribution in Chicxulub Melt Rocks: Forensic Chemistry
   on the KT Smoking Gun

The magnitude and duration of catastrophic environmental
effects resulting from collision of an extraterrestrial body
with Earth are, to first order, proportional to the energy
released on impact. This energy can be estimated from
knowledge of the characteristics of the projectile, the size
and morphology of the impact structure, and the amount,
distribution, and composition of the ejecta. Significant
progress toward piecing together the KT puzzle has come from
recent studies of distal and proximal ejecta deposits [1-7]
and recognition of the ~300 km diameter Chicxulub impact
basin as the KT source crater [8-13]. The character of the
projectile in terms of its size and correlation with known
meteorite types, however, is still uncertain [14-16]. In
order to constrain better the chemical fingerprint of the
projectile that may have triggered the KT extinctions, we
report the results of analyses on samples of impact melt
rocks from Chicxulub.

Samples of impact melt rock from drill core intervals Y6-
N17, Y6-N19, and C1-N10 have major element and lithophile
trace element abundances characteristic of average
continental crust [11,17], and most elements fall within the
compositional range of impact glasses from Haiti [3]. The
crustal character of Chicxulub melt rocks and their
compositional similarity to proximal KT ejecta are supported
by S, Nd, and O isotopic analyses [6].

Instrumental neutron activation analyses (INAA) reveal that
several melt rock fragments from C1-N10 and Y6-N19 have Ir
abundances well above those typical of continental crust
[11]. Measured Ir abundances in five samples range from 2.5
+ 0.5 ppb to 13.5 + 0.9 ppb. These anomalous concentrations
are within the range of Ir enrichments in melt rocks from
other terrestrial impact structures [15,18], and suggest
that the Ir is of extraterrestrial origin. Iridium was below
our detection limit in six other samples, with 2s upper
limits ranging from <1.5 ppb to <2.9 ppb. No measurable Ir
was found in our sample of Y6-N17 (2 sigma upper limit <15
ppb Ir; 8.1 + 1.8 ppb Au), confirmed by results of [19] for
their analyses of samples from the same core interval. While
some variation in Ir concentration among melt rocks is not
unusual, it has been suggested that if Ir is detected in one
sample, Ir enrichments are generally found in all samples
within a single melt sheet [15]. Consequently, it is
possible that all our samples contain enhanced levels of Ir,
albeit below our detection limit of ~1 ppb. In addition, the
Ir variations we observe reflect sampling on a relatively
small scale, and although the differences in measured Ir
abundances are significant analytically, they may not be
representative of variations among larger samples integrated
over the entire melt sheet.

Iridium variations among the different samples of Y6-N19 may
reflect in part, the heterogeneous lithologic character of
this melt rock breccia [11]. The nine, centimeter-size
fragments were taken from an 18 cm length of core to sample
specifically the variety of melt clasts and melt matrix
types.  However, significant Ir variations were also
observed between ~100 mg and ~60 mg splits of finely-ground
powder from a single fragment of Y6-N19-R (8.5 + 0.5 ppb and
13.5 + 0.9 ppb, respectively). In contrast to the
lithological heterogeneity of Y6-N19, two samples taken from
different biscuits within the C1-N10 core interval exhibit
similar, uniform intergranular textures, virtually free of
unmelted clasts. Thus, the difference in Ir abundance
between C1-N10-1 (6.0 + 0.7 ppb) and two splits of the same
powder from C1-N10-2 (<2.2 ppb and <1.7 ppb) cannot be
attributed to macroscopic lithologic differences.
Furthermore, a split of our C1-N10-1 powder analyzed by a
different laboratory yielded an Ir concentration of 15 ppb
[17]. Additional evidence for the heterogeneous distribution
of an extraterrestrial component in C1-N10 is indicated by
significant variations in Os abundance and Re-Os isotopic
ratios [17].

Chromium and the siderophile elements Co and Au also show
variations within and among some of the melt rock fragments,
however, inter-element correlations are not clear-cut for
the limited number that also contain Ir. There appears to be
a crude positive correlation of Co with Ir, and inverse
correlations of Cr with Co and Ir, but these trends are not
consistently maintained for all samples. The analytical
uncertainties for Ni, As, and Sb are too large to resolve
differences among the samples for these elements.

Figure 1 shows the Cr and siderophile element distribution
patterns relative to CI chondrites based on average measured
abundances within the C1-N10 and Y6-N19 core intervals.
Common to both intervals is the depletion in Ni relative to
the more refractory siderophiles, and progressive enrichment
of the more volatile elements with chalcophilic tendencies.
Although different in detail, these melt rock patterns bear
an overall resemblance to that of the KT boundary fish clay
at Stevns Klint [14], providing additional constraints on
impactor type and the partitioning of projectile material
between the target and the ejecta. For example, arguments
against a chondritic projectile stem mainly from the low
Ni/Ir ratio observed at Stevns Klint [14-16]. The KT
boundary layer in the Raton Basin also appears to have a
sub-chondritic Ni/Ir ratio [20]. If the low Ni/Ir ratio
common to Chicxulub melt rocks and both continental and
marine KT sections is also characteristic of the projectile,
then arguments favoring an Fe or ureilite meteorite might be
strengthened [15]. However, indications of hydrothermal
alteration and the potential for fractionation in the melt
rocks [21], possible post-depositional mobility of Ir in
sediments [22], as well as evidence for fractionation of
siderophile elements in impact melt specimens relative to
the projectile where the projectile composition is known
[23], suggest that attempts to characterize the impactor at
this juncture would be premature.

Small-scale variations, particularly between aliquots of a
single powdered sample, suggest that the Ir is concentrated
in a trace phase that is not uniformly disseminated in the
melt rock. Our initial report on the discovery of Ir in the
Chicxulub melt rocks noted that the higher Ir concentrations
seemed to correlate with the abundance of opaque, highly
reflective phases [11]. No metal grains have yet been
identified in the melt rock. The dominant opaque mineral is
magnetite, however, petrographic examination of thin
sections from C1-N10 and Y6-N19 reveal the presence of <1%
sulfides, ranging in size from 10 to 300 mm and distributed
irregularly in the melt matrix. Significant enrichments of
siderophile elements have been found in association with Ni-
Fe sulfide particles in melt rocks at the E. Clearwater
Impact structure [24,25]. In the absence of discrete metal,
sulfides would seem to be likely carrier phases for Ir and
other siderophile elements.

Work is in progress to separate the sulfides and other
phases to search for the Ir by INAA. Preliminary
characterization of the sulfides by electron microprobe
indicates that most of the grains are pyrite, with
subordinate chalcopyrite. Random analyses of the pyrite
revealed significant enrichments in Co and Ni, with Co
concentrations up to 6 wt%. Figure 2 shows an example of
pyrite from C1-N10, which suggests that the Co and Ni were
enriched by oscillatory growth zoning. Although both
siderophiles appear to be enriched in the same general
regions, close inspection reveals that the zones of highest
Ni concentration (1.1 wt%) are relatively depleted in Co
(0.6 wt%), whereas the highest Co (3.9 wt%) corresponds to
0.2 wt% Ni.  The best estimates of the average
concentrations in this pyrite are 0.7 wt% Co and 0.5 wt% Ni.
These results confirm the concentration of siderophile
elements in a heterogeneously disseminated trace phase, and
indicate that some siderophile element fractionation has
occurred. The paragenesis of the pyrite, and whether or not
this phase contains the iridium requires further
investigation. Clearly, more evidence must be gathered
before the composition and caliber of the KT bullet can be
ascertained.

References: [1] Sigurdsson H. et al. (1991) Nature, 353,
839-842. [2] Izett G. A. et al. (1991) Science, 252,1539-
1542. [3] Koeberl C. and Sigurdsson H. (1992) GCA, 56, 2113-
2129. [4] Smit J. et al. (1992) Geology, 20, 99-103. [5]
Alvarez W. et al. (1992) Geology, 20, 697-700. [6] Blum J.
D. et al. (1993) Nature, 364, 325-327. [7] Krogh T. E. et
al. (1993) EPSL, 119, 425-429. [8] Hildebrand A. R. et al.
(1991) Geology, 19, 867-871. [9] Kring D. A. and Boynton W.
V. (1992) Nature, 358, 141-144. [10] Swisher C. C. et al.
(1992) Science, 257, 954-958. [11] Sharpton V. L. et al.
(1992) Nature, 359, 819-821. [12] Sharpton V. L. et al.
(1993) Science, 261, 1564-1567. [13] Sharpton V. L. et al.,
this volume. [14] Kyte F. T. et al. (1980) Nature, 288, 651-
656. [15] Palme H. (1982) GSA Spec. Paper, 190, 223-234.
[16] Kyte F. T. and Wasson J. T. (1982) GSA Spec. Paper,
190, 235-242. [17] Koeberl C. et al. (1993) GCA, in press.
[18] Grieve R. A. F. (1982) GSA Spec. Paper, 190, 25-37.
[19] Hildebrand A. R. et al. (1993) LPS XXIV, 657-658. [20]
Izett G. A. (1990) GSA Spec. Paper, 249, 100. [21] Schuraytz
B. C. and Sharpton V. L. (1993) Nature, 362, 503-504. [22]
Colodner D. C. et al. (1992) Nature, 358, 402-404. [23]
Mittlefehldt D. W. et al. (1992) Meteoritics, 27, 361-370.
[24] Palme H. et al. (1979) Proc. LPSC 10th, 2465-2492. [25]
Grieve R. A. F. et al. (1980) Contrib. Mineral. Petrol., 75,
187-198. [26] Wasson J. T. and Kallemeyn G. W. (1988) Phil.
Trans. R. Soc. Lond. A, 325, 535-544. [27] Wasson J. T.
(1985) Meteorites, W. H. Freeman and Co., New York, 267.

Fig. 1, which appears here in the hard copy, shows
siderophile element and Cr abundance ratios in Chicxulub
melt rock samples relative to CI chondrites [26]. Elements
along the abscissa are arranged in order of decreasing
nebular condensation temperature [27]. Shown are the average
and range of two samples from C1-N10, and the mean and
standard deviation of nine samples from Y6-N19. Rhenium and
Os data for C1-N10 from [17]. Also shown are the mean and
standard deviation of six samples of the KT boundary fish
clay at Stevns Klint [14].

Fig. 2, which appears here in the hard copy, shows
photomicrographs of pyrite grain in thin section from C1-
N10. (a) Backscatter electron image of pyrite with minor
silicate inclusions surrounded by porous matrix of quartz
and feldspar. Slightly darker-gray grain in saddle along the
upper right margin of the pyrite is magnetite. (b) X-ray map
of Co. (c) X-ray map of Ni.


Sigurdsson H.   Smith S.   D'Hondt S.   Carey S.   Espindola J.-M.
   Crystals, Lithics, and Glassy Ejecta at the KT Boundary: Implications for
   Lithology of the Crust at the Impact Site

Ejecta from bolide impact may contain melt quenched to glass
droplets as well as rock fragments and fractured mineral
grains from the impact terrane. In distal sections of the
ejecta deposit, an impact origin for glass and shocked
minerals may generally be established, whereas the origin of
lithics and unshocked crystal fragments found in association
with impact glass is uncertain, due to possible
contamination from other sources. In addition to 1 to 8 mm
diameter glassy ejecta spherules, the 0.5 to 1 m thick KT
boundary impact ejecta layer in the Beloc pelagic carbonate
sediment formation in Haiti contains silicate mineral
fragments and rock fragments that may provide clues about
the nature of the Earth's crust at the impact site. Because
of the monotonous and relatively pure lithology of the
enclosing upper Createceous and lower Tertiary carbonate
sediments, the Haiti exposures of the KT boundary layer
provide an opportunity to detect mineral ejecta present in
only trace amounts, in addition to the abundant impact glass
[1,2]. New studies of the chemistry of glass spherules from
the Haiti deposit (Fig. 1) further define the mixing trend
already documented between high-Ca yellow glass and
high-silica black glass spherules [1,3]. The high-silica
glasses with SiO2 >62 wt% define a separate trend in terms
of most oxides, such as FeO (Fig. 2), which may reflect
chemical range in the crustal source region. New analyses
show that S content of yellow high-Ca glasses ranges up to
1.25% SO3 [4].

Isolated mineral fragments in the ejecta layer include
amphibole, quartz, clinopyroxene, and plagioclase. Amphibole
is a relatively common and well-preserved crystal component
in the ejecta layer, and is present at all levels, except in
the interval from 1 to 30 cm above the base. The amphiboles
are up to 0.5 rnm in size, and occur exclusively as angular,
euhedral crystals, whose shape is defined by {110} cleavage
planes (Fig. 3). These amphiboles are magnesio-hornblendes
to tschermakitic hornblendes, and define a coherent
compositional trend, with a relatively narrow range from
(Mg/Mg+Fe) = 0.70 to 0.80 (Amph. in table). Variation or
zoning within individual crystals is small, in the range
(Mg/Mg+Fe) = 0.02 to 0.07. The amphiboles are relatively
high in A12O3 (7-12 wt%); with total Al cations in the range
1.3 to 1.45 in the structural formula, these amphiboles
could have equilibrated at total pressure of 3 to 4 kbar (9
to 12 km; [5]). Shocked quartz crystals up to 0.55 mm in
size have already been documented from the Haiti KT section
[6,7]. Fractured plagioclase is very rare and occurs as
solitary and angular crystal fragments. The crystals are of
andesine-labradorite An(sub)5OAb(sub)48Or(sub)2, and
relatively unzoned (Plag-1 in table; An(sub)47-52). Biotite
crystal flakes are very rare.

Solitary clinopyroxene and plagioclase crystals with a
reticulate glass coating are also present. The euhedral
clinopyroxene has the composition
Wo(sub)44En(sub)44Fs(sub)12 (Cpx in table), is 0.7 mm in
length and coated with glass. It contains a rhyodacite glass
inclusion (Glass incl. in table), indicating a magmatic or
volcanic origin of the mineral. However, the glass coating
on exterior of the crystal (Glass-cpx in table) is dacitic
in composition and virtually identical to the common impact
glass spherules found at the KT boundary in Haiti (Haiti
glass in table). The subhedral to euhedral 0.7 mm
plagioclase is of An(sub)47.2Ab(sub)51.6Or(sub)1.2 mean
composition (Plag-2 in table), zoned from An(sub)51 core to
An(sub)44 rim, and also coated with dacitic glass. A single
1.3 mm highly vesicular pumice fragment is also present. In
addition, the KT boundary deposit in Haiti contains minerals
that are authigenic in origin. They include common crystals
of euhedral to subhedral barite, up to 1 mm in diameter,
euhedral and bipyramidal crystals of perfectly formed clear
quartz, and high-Ca-K phillipsite zeolites. Lithic fragments
of graywacke are present but rare in the ejecta deposit, up
to 1.5 mm in size. They consist dominantly of relatively
poorly sorted sedimentary rock, containing large (0.5 mm)
and rounded quartz grains in a silty to clayey matrix.

Studies of mineralogy of upper Cretaceous and lower Tertiary
sediments from DSDP Legs 4 and 15 at ten sites in the
Caribbean provide information on sedimentation of silicate
minerals elsewhere in the region at KT boundary time [8,9].
In addition to authigenic minerals in the latest
Mastrichtian-early Paleocene central Caribbean carbonate
sediments, such as chert, clinoptilolite, barite, these
studies also record grains of hornblende, plagioclase, and
clinopyroxene.

If the silicate mineral assemblage in the Haiti deposit is
of ejecta origin, as indicated by impact glass coating on
clinopyroxene, then it may indicate a crustal section
including both calcalkaline volcanics and possibly mid- to
upper-crustal amphibolites, overlain by greywacke,
evaporites and carbonate sediments. In terms of comparison
with geologic terrane in the Yucatan region, the greywacke
lithic fragments in the Haiti deposit may correlate with the
Jurassic Todos Santos Formation, which is over 400 m thick
in well Y-1 [10,11]. Similarly, the silicate mineral
assemblage in the Haiti deposit may be derived from the
Yucatan crystalline basement rocks of Paleozoic age, which
in the same well include felsic metaporphyry,
crystalloblastic quartz chlorite schist, and rhyolite [10-
12].

References: [1] Sigurdsson H. et al. (1991) Nature, 349,
482-487. [2] Sigurdsson H. et al. (1991) Nature, 353, 839-
842. [3] Sigurdsson H. et al. (1992) EPSL, 109, 543-560. [4]
Chaussidon M. et al. (1994) this volume. [5] Johnson M. C.
and Rutherford M. J. (1989) Geology, 17, 837-841. [6]
Hildebrand A. R. and Boynton W. V. (1990) Science, 248, 843-
847. [7] Izett G. A. (1991) JGR, 96, 20879-20905. [8]
Donnelly T. W. and Nalli G. (1973) Init. Rep. DSDP. XV,
929-961. [9] Donnelly T. W. (1973) Init. Rep. DSDP. XV,
969-987. [10] Bass M. N. and Zartman R. E. (1969) EOS, 50,
313. [11] Viniegra O. F. (1971) Am. Assoc. Petrol. Geol.
Bull., 55, 478-494. [12] Donnelly, T. W. et al (1990)
Geology of North America, H, Geol. Soc. Amer., 37-76.

Table 1, which appears here in the hard copy, shows silicate
minerals and glass in the KT boundary ejecta layer.

Fig. 1. Compositional trends in KT boundary glassy ejecta
spherules, illustrating the smooth compositional trends of
A12O3 and CaO, from high-Ca yellow glasses to the more
abundant highsilica glasses. In total, the dataset
represents analyses of 140 spherules. The compositional
range from 44 to 62 wt% SiO2 can be modeled as a simple
binary mixing trend [1], whereas the highsilica glasses
reflect a more heterogenous source. Curves define best-fit
regression lines through the data, with correlation
coefficient of 0.992 for CaO and 0.91 for Al2O3.

Fig. 2. Two distinct trends of FeO content in KT boundary
impact glasses. The trend of decreasing Fe content with
increasing silica in the high-silica glasses may reflect
heterogeneity in the source region, whereas the trend
between high-Ca glasses (44 to 54 wt% SiO2) and high-silica
glasses can be modeled as a mixing trend.

Fig. 3. Scanning electron micrograph of a crystal of
amphibole from the KT boundary ejecta layer. The 400 micron
euhedral crystal is entirely defined by {110} cleavage
planes and formed by breakage. Mineral chemistry is
consistent with derivation from 9 to 12 km depth in the
crust.


Hough R. M.   Sigurdsson H.   Franchi I. A.   Wright I. P.   Pillinger C. T.
Gilmour I.
   Carbonate Derived Gases in Haitian KT Boundary Glass Spherules

Glass spherules, thought to be tektites from Haiti, have
been analyzed previously for their mineralogy and chemical
composition to identify their origin and mode of formation
[1].

They occur in various colors, dependent upon the composition
of the original target rock, and in many cases are seen to
contain bubbles. To investigate these spherules and the
nature of any gas phase present in the bubbles, several dark
brown glasses have been analyzed for their C content and
isotope composition, using stepped combustion analysis and
static mass spectrometry [2]. Furthermore, brown and yellow
spherules were analyzed for O isotope composition using
laser fluorination and conventional dynamic gas source mass
spectrometry.

Petrographic analysis using a standard microscope
illustrated the presence of "bubbles" within the dark-brown
glass spherules. By simply submerging a fragment in
refractive index oil, placing a glass slide above it, and
applying pressure it was possible to release the gas in the
bubbles. Since on no occasion did the oil move to fill the
void left when the gas was released it can be concluded that
the bubbles were at greater than atmospheric pressure and
may represent a volatile component such as CO2 or
hydrocarbon. Bubbles were also released simply through
fracturing glass fragments that didn't display any signs of
bubbles.

In most cases spherules analyzed for C were treated as whole
entities, but one was broken into fragments for the purpose
of replication. Individual fragments were found to contain
0.2 wt% C two components of different isotopic compositions.
The first component had a release temperature of 350-400
degrees C with a delta^13C of -22 permil, the second
component was released at higher temperatures, 500-600
degrees C and had a delta^13C value of -6.3 permil. The
isotopic composition and release temperature of the first
component strongly suggests that it is organic and probably
represents surface contamination of the sample. To try and
reduce contamination, other spherule fragments were
pretreated with 0.1 M chromic acid to remove organic and
carbonate components. Analyses of cleaned fragments
indicated a variable carbon content from 0.005 to 2.6 wt% C,
but again with two isotopically different components. The
first with a peak delta^13C of -0.8 permil released between
420 and 460 degrees C and the second, a delta^13C of -19.0
permil released between 550-600 degrees C (Fig. 1). Levels
of low-temperature organic contamination were greatly
reduced in the cleaned spherules.

The component with a delta^13C of -19 permil is present in
most of the spherule fragments. A component with a similar
combustion temperature and delta^13C has been encountered in
KT residues containing nanodiamonds [3]. There is currently
no information available confirming its identity, but it
does not appear to be surficial or an oxidizable organic.

The C released between 420-460 degree C is present in most
spherule fragments and has a consistently heavy isotopic
composition (maximum delta^13C of -0.8 permil). This
isotopic composition is remarkably similar to those of
marine carbonates [4], but isotopically heavier than present
day atmospheric CO2 (-7 permil).

If the bubbles are indeed CO2 then one likely origin is that
they represent CO2 released from target rock carbonates by
the impact [5]. Identification of these C components by
future work may reveal a possible source and mode of
formation for the spherules and will also clarify the effect
of the internal bubbles upon the compositions.

Dark-brown spherules selected for O isotope measurements
were broken into fragments to allow repeat analyses on the
same spherule. Due to the smaller size of the yellow
spherules they were analyzed whole. The dark-brown spherules
have delta^180 values between +7.1 and +9.5 permil and
delta^170 between +3.7 and +5.0 permil. The yellow spherules
have a delta^180 of +13.0 permil and a delta^170 of +6.8
permil.

Both types of spherule fall on the terrestrial fractionation
line (TFL) on a plot of delta^170 vs. delta^180 (Fig. 2).
The O isotope data suggests no obvious input from an
extraterrestrial impactor into the glasses but it should be
noted that only a small number of spherules have been
analyzed.

Heterogeneities seen in the C data for the dark-brown
spherules seem to be reflected in the O data with variations
between fragments of the same spherule and between whole
spherules.

References: [1] Koeberl C. and Sigurdsson H. (1992) GCA, 56,
2113-2119. [2] Yates P. D. et al. (1992) Chem. Geol., 101,
81-91. [3] Gilmour I. et al. (1992) Science, 258, 1624-1625.
[4] Faure G. (1986) Wiley & Sons, New York, 496-499. [5]
Sigurdsson H. et al. (1991) Nature, 353, 839.

Fig. 1, which appears here in the hard copy, shows the
stepped combustion profile of a dark-brown glass spherule
from Haiti.

Fig. 2, which appears here in the hard copy, shows the
oxygen isotopic composition of glass spherules.


Evans N. J.   Ahrens T. J.   McInnes B. I. A.   Gregoire D. C.
   New Evidence for Primary Fractionation of Ruthenium and Iridium in the
   Chicxulub Ejecta Cloud

Introduction: Platinum-group element (PGE; Ru, Ir, Pt, Pd,
Ir, plus Au, included in the term PGE for the sake of
brevity) analysis of the Cretaceous-Tertiary (KT) boundary
fireball layer (terminology of [1]) has revealed that the
mean Ru/Ir ratio of the many marine sites studied (1.67 +-
0.38) is close to the chondrite ratio (1.48; [2]) whereas
the value for nonmarine sites (0.76 +- 0.26) is not [3,4]. A
positive correlation between the Ru/Ir ratio of globally
distributed KT sites and 65Ma paleo-distance from the
Chicxulub structure was also observed [4]. This trend
suggested that temperature-dependant, primary fractionation
of PGE occurred in the ejecta cloud during condensation of
vaporized projectile material. However, this previous work
could not negate the dependance of the paleo-distance-Ru/Ir
ratio correlation on environment of deposition because all
the marine sites studied were far from Chicxulub (Europe and
New Zealand) and all the nonmarine sites were in North
America. This work presents additional Ru and Ir analysis of
the fireball layer from marine KT sites proximal to the
impact structure at Chicxulub, Yucutan, other North American
nonmarine KT sites and DSDP cores. With this new data we
assess the dependance of the previously observed trend on
depositional environment and suggest a simple mechanism for
primary fractionation of PGE, prior to their deposition in
the fireball layer.

Methods and Results: Ruthenium and Ir in KT boundary samples
were measured by isotope dilution inductively coupled mass
spectrometry (ICP-MS) and a complete description of the
digestion and analytical procedures have been published
elsewhere [3,4]. Precision for the technique as percent
standard deviation for multiplicate analysis of USGS diabase
standard W-1 are 6% (Ru) and 0.7% (Ir).

The abundances of Ru and Ir for all KT sites is given in
Table 1. The Ru/Ir ratio and nature of the fireball layer
samples analyzed from the following sites have been
previously published [4]; Petriccio, Italy; Knappengraben
and Elendgraben, Austria; Stevns Klint, Denmark; Agost,
Spain; Woodside Creek, New Zealand; Raton Basin, Colorado;
Red Deer Valley, Alberta; Morgan Creek, Saskatchewan and
Lance Creek, Wyoming. New data is presented here for DSDP
577, DSDP 596, Brazos River, Texas and Beloc, Haiti.

DSDP Samples: Typical Ir contents for the boundary in DSDP
577 are 5 ppb [5] over an order of magnitude higher than the
value obtained here. A possible explanation for this
discrepancy may be that the boundary sample analyzed is at
the upper limit of the boundary interval, which is smeared
out by bioturbation over 10 cm in this core [5,6]. Since no
Ru values have previously been presented for this section,
there is no basis for comparison of the present Ru data and
no way to estimate how sampling the outer region of the
boundary interval might affect the Ru/Ir ratio. However, we
have both Ru and Ir data for the entire 10 cm boundary
interval in DSDP 596 and the Ru/Ir ratio only varies by +-
0.06 throughout (unpublished data).

The PGE-rich KT section in DSDP 596 is smeared over at least
10 cm by bioturbation [7] and the values in Table 1
represent a mean for this interval. The physical mixing of
the sediment does not seem to have affected the Ru/Ir ratio
since, as noted above, it varies little over the 10 cm
interval.

Brazos River, Texas and Beloc, Haiti: There is significant
controversy over the location of the KT boundary at the
extensively studied Brazos River site [8-10], however the
observed Ru and Ir data indicate that the degree of
fractionation of Ru from Ir during physical reworking
processes is insignificant and we conclude that the Brazos
River Ru/Ir ratio reflects as primary a value as that for
the other sites analyzed. The fireball layer sample from
Beloc, Haiti (HKB92-56, obtained from A. Hildebrand) is a
medium gray claystone 1.5 cm thick.

There is some uncertainty regarding the composition of the
boundary sample we analyzed for Lance Creek (collected by
others) in that it might include both the underlying
kaolinitic claystone and the fireball layer. This would have
the effect of raising the Ru/Ir ratio because the ratio in
the kaolinitic claystone is always higher than that of the
associated fireball layer for any given site [4]. For this
reason, the Lance Creek value has been excluded from the
calculation of regression coefficients (Fig. 1).
Table 1 gives the paleo-distances (65 Ma) from each site to
the Chicxulub impact structure and the corresponding Ru/Ir
ratios. In Fig. 1 we have plotted all of the Ru/Ir ratios
for the fireball layer worldwide against the paleo-distance
from Chicxulub. In addition to sites analyzed in our
laboratory, data for Chancet Rocks and Neddles Point, New
Zealand from [11] have been included.

Discussion: Previous work has shown that although all the
PGE in the KT fireball layer are somewhat redistributed
laterally and vertically by post-depositional processes [11-
13], Ir and Ru are not significantly fractionated from each
other during redistribution and are the most immobile PGE in
the sedimentary environment [14,4]. The Ru/Ir ratio is
therefore the most useful PGE interelement parameter for
providing insight into primary fractionation processes. The
Ru/Ir ratio of marine and nonmarine KT fireball layer
samples are statistically distinct and nonmarine sites have
nonchondritic ratios. In addition, the fact that all the
marine sites previously analyzed were far from the Chicxulub
impact site and all nonmarine sites were located in North
America resulted in a positive correlation between Ru/Ir
ratio and paleo-distance to Chicxulub. This trend cannot be
explained by input of mantle-derived PGE or addition of PGE
from a simultaneous impact event [3]. A primary mechanism
for fractionation, which operates to change the Ru/Ir ratio
of nonmarine, North American KT sites from chondrite prior
to their deposition in the fireball layer, is therefore
indicated. However, more data was needed to definitively
eliminate the possible dependance of the correlation on
factors related to environment of deposition. The addition
of data from the marine sites, Brazos River and Beloc, which
both plot in the region dominated by nonmarine sites (Fig.
1), negates the dependance of the trend on differential
remobilization associated with differences in the
depositional environment.

In addition, samples of DSDP cores from the Pacific
represent an environment not previously investigated for
Ru/Ir ratios and provide convincing evidence that the
environment of deposition is independent of the observed
trend. The greater then 1000 degrees C difference in the
condensation temperatures of Ir (5017 degrees C) and Ru
(3900 degrees C)[15] leads us to propose that fractionation
of Ir from Ru during condensation of the projectile-rich
component of the ejecta may have occurred in the cooling
ejecta cloud. The first condensates would have a low Ru/Ir
ratio because Ir condensed from the cloud while Ru remained
in the vapor. Settling of early-formed condensates close to
the impact site could explain the low Ru/Ir ratios of North
American nonmarine KT sites (Table 1, Fig. 1). Early
fractionation of Ir from the ejecta cloud combined with
later incorporation of Ru results in later condensates
having higher Ru/Ir ratios [4].

Chemical analysis of ejecta collected after a 6.4km/s
experimental hypervelocity impact of a Fe-Ni-PGE alloy
projectile into a Mo target indicates that temperature-
dependant fractionation of PGE occurs in the ejecta during
an impact event (Evans and Ahrens, unpublished data).
Further study of these experimental results will determine
whether this fractionation occurred during melting-quenching
or vaporization-condensation processes.

Fig. 1. Paleo-distance (km) from KT sample sites studied to
Chicxulub, Yucatan vs. Ru/Ir ratio. 1. Raton Basin,
Colorado; 2. Lance Creek, Wyoming (not included in
regression coefficient calculation); 3. Morgan Creek,
Saskatchewan; 4. Frenchman River, Saskatchewan; 5. Hell
Creek, Montana; 6. Red Deer Valley, Alberta; 7. Stevns
Klint, Denmark; 8. Agost, Spain; 9. Petriccio, Italy; 10.
Knappengraben, Austria; 11. Elendgraben, Austria; 12.
Woodside Creek, New Zealand; 13. Brazos River, Texas; 14.
Beloc, Haiti; 15. DSDP 577; 16. DSDP 596; 17. Chancet Rocks,
New Zealand; 18. Needles Point, New Zealand. Sites 17 and 18
are data of [11].

References: [1] Hildebrand A. R. and Boynton W. V. (1990)
Science, 248, 843-847. [2] Anders E. and Grevesse N. (1989)
GCA, 53, 2363-2380. [3] Evans N. J. et al. (1993a) GCA, 57,
3737-3748. [4] Evans N. J. et al. (1993b) GCA, 57, 3149-
3158. [5] Michel H. V. et al. (1985) Init. Reports DSDP, 86
(G. R. Heath et al. eds.), 533-538. [6] Smit J. and Romein
A. J. T. (1985) EPSL, 74, 155-170. [7] Zhou L. et al. (1991)
Geology, 19, 694-697. [8] Bourgeois J. et al. (1988)
Science, 241, 567-570. [9] Keller G. (1989) GSA Bull., 101,
1408-1419. [10] Montgomery H. et al. (1992) EPSL, 109, 593-
600. [11] Tredoux M. et al. (1989) J. Geol., 97, 585-605.
[12] Schmitz B. (1988) Geology, 16, 1068-1072. [13] Izett G.
A. (1990) Geological Society of America Special Paper 249,
100. [14] Cousins C. A. and Vermaak C. F. (1976) Econ.
Geol., 71, 287-305. [15] Weast R. C. (1989) CRC Handbook of
Chemistry and Physics, 69th ed. [16] Schmitz B. (1985) GCA,
49, 2361-2370.


Chaussidon M.   Sigurdsson H.   Metrich N.
   Sulfur Isotope Study of High-Calcium Impact Glasses from the KT Boundary

High-Ca yellow glasses form a volumetrically minor component
of the impact spherules found in the KT boundary deposit in
Haiti, but these glasses contain valuable information about
lithology of the geologic terrane at the impact site. The
high-Ca glasses define a mixing trend with abundant high-
silica glasses, which led to the hypothesis that they are
derived from fusion of Ca-rich sediment in presence of
silicate melt [1]. Two different sources of Ca have been
proposed for the yellow glasses, either anhydrite-rich
evaporitic terrane, as suggested by their shigh S content
and elevated delta^34S (+13.2 +- 4 per mil, one sample) [2],
or platform carbonates as suggested by their high delta^18O
values (+13.6 +- 0.8 per mil, 3 samples) [3]. These two
hypotheses have important and radically different
implications for atmospheric and other environmental effects
following bolide impact: either global cooling due to
sulfate aerosol formation [4], or greenhouse effect because
of increased CO2 concentration in the atmosphere. Because of
the correlations observed between S and CaO in the yellow
glasses [2], we conducted a detailed S isotopic study of 20
high-Ca yellow glass particles in order to constrain sources
of S and Ca.

Sulfur isotopes were determined with ims 3f ion microprobe,
using an O-primary beam and analyzing secondary S-ions at a
mass resolution of ~3500 [5,6]. Instrumental mass
fractionation was determined (alpha(sub)instr = 0.9547 +-
2.8 per mil) on high-Ca synthetic glass standards with 0.8
to 1.2 wt% SO3, dissolved mainly as sulfate and having
delta^34S of +24.4 per mil. Analytical uncertainty of the
delta^34S values is between +1.5 and +5 per mil, depending
mainly on S content. The yellow glasses are heterogenous in
terms of both S and delta^34S values. Their S contents range
from 0.3 to 1.25 wt% SO3. The spectral position of the S Ka
line in the electron microprobe shows that S speciation is
~94.5% sulfate in the high-Ca glass (sulfate/sulfide =
17.2), in agreement with high O fugacity inferred from
Mossbauer spectra [7]. Melting experiments with gypsum and
andesite have reproduced the high-Ca glass trend, both for
major oxides and S content [2]. A S geothermometer,
calibrated by melting experiments, indicates that the
crystal-free glasses were quenched in the range 1257-1360
degrees C.

The delta^34S values range between +2.9 and +14.5 per mil,
confirming earlier results [2]. A strong isotopic
heterogeneity is observed at all scales, between +2.5 +- 5.1
per mil and +10.9 +- 1.9 per mil within a single glass
particle, and between +2.9 +- 6.8 per mil and +12.3 +- 3 per
mil between different particles. This degree of isotopic
heterogeneity is not known in any terrestrial volcanic
glasses [8] and was not observed in melting experiments [2],
which produced glasses with homogenous delta^34S values,
within analytical uncertainty (+2.8 per mil). The variations
of delta^34S are not clearly correlated to other parameters,
however, all the samples show a general tendency of increase
of delta^34S values with increase of S and Ca content (Fig.
1). The only previous yellow glass measured for delta^34S
[2] follows the same trend.

The delta^34S values of the yellow glasses can be used to
constrain the origin of S. Clearly sedimentary rocks are the
only ones which have high enough S contents to account for
the high S content of the yellow glass. The hypothesis of an
anhydrite-rich evaporitic source was tested by modeling, on
basis of mass fraction of S degassing calculated from known
quenching temperatures of the glasses, isotopic
fractionation of S, and the observed delta^34S of the yellow
glasses. Assuming mixing between anhydrite (41% CaO and 59%
SO3) and a silicate melt from a crustal component with
roughly andesitic composition (5% CaO, 0.3 wt% SO3), the
fraction of anhydrite required to account for the CaO in the
yellow glass was calculated from mass balance, ranging from
0.3 to 0.6. This would have produced a silicate melt plus
anhydrite mixture with ~20 to 36 wt% SO3. Initial delta^34S
value of this mixture, before S degassing, is assumed to be
that of the anhydrite. Loss of S from the melt during
degassing/volatilization lowered their S content to 0.3 to
1.25 wt% SO3, causing isotopic fractionation. We have
calculated the degree of fractionation on basis of quench
temperature (from the S geothermometer), and from dependence
of the isotopic fractionation on temperature at a dissolved
sulfate/sulfide ratio of 17.2 in the glasses [9]. The
results shown in Fig. 2 are for a delta^34S of +12 per mil
(the minimum value for Cretaceous evaporite [10]). It is
clear from Fig. 2 that some of the glasses have delta^34S
values in agreement with the hypothesis of isotopic
fractionation during anhydrite volatilization, while others
have too low delta^34S values. This might be related to the
different trends observed in Fig. 1. It is clear that a
high-S source such as anhydrite is required, but these
initial results indicate that in addition to anhydrite,
other high-Ca sediments, such as carbonates, may have
participated in the formation of the yellow glasses,
contributing to the large delta^34S range. As discussed
previously [4] the geology of the Chicxulub region of the
Yucatan peninsula in Mexico is consistent with the above
scenario, with interbedded Cretaceous evaporites and
carbonates. Further S and O isotope measurements are in
progress in order to quantify the role of these two
components in the genesis of the yellow high-Ca impact
glasses.

References: [1] Sigurdsson H. et al. (1991) Nature, 349,
482-487. [2] Sigurdsson H. et al. (1991) Nature, 353, 839-
842. [3] Blum J. D. and Chamberlain C. P. (1992) Science,
257, 1104-1107. [4] Sigurdsson H. et al. (1992) EPSL, 109,
543-560. [5] Chaussidon M. and Lorand J.-P. (1990) GCA, 54,
2835-2846. [6] Deloule E. et al (1992) Chem. Geol., 101,
187-192. [7] Oskarsson N. et al. (1991) LPS XXII, 1009. [8]
Taylor B. E. (1986) Reviews in Mineralogy, 16. [9] Zheng Y.-
F. (1990) Terra Nova, 2, 74-78. [10] Claypool G. E. et al
(1980) Chem. Geol., 28, 199-260.

Fig. 1. Mean CaO content vs. delta^34S (per mil CDT) in
high-Ca and high-S yellow impact glass particles from the KT
boundary in Haiti. Open circle shows results from an earlier
study [2].

Fig. 2. Isotopic fractionation model of delta^34S evolution
in high-Ca yellow impact glasses. Calculated delta^34S is
based on fractionation factors from [9], quench temperature
determined from S geothermometry and measured
sulfate/sulfide ratio in the glass. Open circle shows
results from previous study [2].


D'Hondt S.   Sigurdsson H.   Hanson A.   Carey S.   Pilson M.
   Sulfate Volatilization, Surface-water Acidification, and Extinction at the
   KT Boundary

It appears that the severity of environmental effects
related to the Cretaceous-Tertiary (KT) boundary impact may
have been largely due to the unusual geochemistry of an
evaporite-rich impacted terrane. KT tektite composition,
experimental results and comparison to Yucatan stratigraphy
all indicate the presence of gypsum or anhydrite at the
tektite source (Chicxulub crater). Thick (~1 km) Late
Cretaceous evaporite sequences occur in the region of the
Chicxulub KT impact structure [1]. Ca-rich KT tektites
contain high SO3 concentrations (0.83% [2]; 0.53% [3]; range
= 0.20 to 1.0%). They also exhibit delta^34S values typical
of evaporites (13.2 [2,4]). Experimental results have
duplicated the high-Ca tektite composition by high
temperature melting of evaporite + andesite [2,5].

Several studies have estimated the amount of SO2 volatized
from target evaporites by the KT impact. An estimate of 1.3
* 10^16 was based on the proportion of unaltered high-Ca
glass in Haiti and the global thickness of the KT boundary
clay. This assumed that 1000 km^3 of impact glass was
created and 2% was high-S glass derived from evaporite
source [2,5]. This represents a minimum estimate, since it
is limited to SO2 released by tektite formation and does not
include SO2 released from solid rock by shock, or SO2
released by initial volatization of the target. More
comprehensive estimates have been derived from
reconstruction of Chicxulub geology and assumed shock
pressures required for sulfate release. These rely on
variable estimates of transient crater diameter (80 to 146
km), stratigraphy of Chicxulub region (0.5 to 1.5 km
anhydrite), and shock pressure of sulfate release (20 to 40
Gpa). Based on such criteria, Sigurdsson et al. [5]
estimated that 2.4 * 10^18 to 8.4 * 10^18 g SO2 was
released, Brett [6] estimated that 4 * 10^17 g SO2 was
released, and Pope et al. [7] estimated that 5.4 * 10^17 to
1.6 * 10^18 g SO2 was released.

Combined with the great optical depth loading previously
estimated to result from the KT impact "dust" cloud, the
resultant stratospheric sulfur aerosols may have contibuted
to a rapid decline in global surface temperatures to near-
freezing in about one week [5]. Time-dependent conversion of
stratospheric SO2 to H2SO4 would have prolonged this cooling
for several years [5]. These stratospheric sulfate aerosols
may also have caused global "blackout," preventing
photosynthesis for months and disrupting it for years [7].
These relatively long-lived effects of the estimated SO2
release (global cooling and global darkness) appear likely
to have been major causes of the global KT extinctions.

The estimated mass of sulfuric acid generated by this event
is equal to or greater than the most tightly constrained
estimates of nitric acid formed by the same event (HNO3
yields corresponding to 10^15 moles NO(sub)x--perhaps higher
if the impact was oblique [8]). Depending on the
stratospheric conversion rate of SO2 to H2SO4 and the rate
of transport to surface waters, these combined sulfuric and
nitric acids may have led to varying levels of acidification
in surface marine and freshwater systems. Instantaneous and
uniform global rainout of maximum KT sulfuric acid estimates
(10^17 moles H2SO4) would have temporarily destroyed the
buffering capacity of surface marine waters (the upper 100 m
of the watercolumn) and driven the pH of nearsurface marine
and fresh waters below 3. Assuming complete atmospheric
conversion of SO2, most estimates of KT SO2 release would
create ~1016 moles of atmospheric H2SO4. Instantaneous and
uniform global rainout of this mass would have driven marine
pH below 6 in the upper 25 m of the watercolumn. Its mixing
throughout a 100 m watercolumn would have depressed marine
pH below 7 and unbuffered freshwater values below 3. Similar
rainout of 10^15 moles of H2SO4 would have depressed surface
marine pH from 8.2 to 7 only in the upper 10 m of the
watercolumn, but would have depressed the pH of unbuffered
fresh waters to about 4.5 throughout a 100 m watercolumn.
Similar rainout of 10^15 moles of HNO3 would have had a
still smaller effect on watermass acidification. In all
cases, seasurface pH would have subsequently increased by
varying amounts as the pCO2 of the atmosphere and surface
oceans re-equilibrated (on timescales of a few months to a
few years). On slightly longer timescales, seasurface pH
would closely approach pre-impact values as surface and deep
waters exchanged.

Possible environmental effects of KT surface ocean and
freshwater acidification include (i) increased availability
of free metal ions in surface waters, (ii) loss of CO32- and
HCO3- from surface marine waters, and (iii) selective
extinction in surface water environments [9,10]. Nontheless,
acidification of the marine and freshwater photic zone may
not have been a primary cause of global KT extinctions. Even
assuming maximum estimates of KT SO2 volatization, our
maximum estimates of global KT watermass acidification are
probably unrealistic since the time-dependent conversion of
SO2 to H2SO4 may preclude instantaneous global injection of
the total H2SO4 mass into surface waters. Furthermore,
analysis of vertebrate fossil records indicates that
freshwater organisms of the western interior United States
exhibited much higher KT survivorship than terrestrial
organisms [11], suggesting that extinction in that region
primarily resulted from factors other than freshwater
acidification. Finally, extant carbonate-secreting marine
phytoplankton and planktic foraminifera have respectively
survived pH values of less than 7.6 [12] and less than 7.2
[13] in culture experiments. This probably precludes rapid
global and uniform rainout of 10^15 moles of H2SO4 and/or
HNO3 from consideration as a primary cause of global
extinction of carbonate-secreting marine plankton. However,
it allows the possibility of regional or depth-related
extinctions resulting from acidification of marine or
freshwater photic environments if rapid nonuniform acid rain
occurred. Furthermore, photic zone acidification remains a
possible cause of global carbonate-secreting plankton
extinction in the case of instantaneous and uniform global
rainout of more than 10^16 moles of H2SO4 and/or HNO3.

References: [1] Lopez Ramos E. (1979) Geologia de Mexico,
Tomo III, 2nd ed. Mexico. [2] Sigurdsson H. et al. (1991)
Nature, 353, 839-842. [3] Koeberl C. and Sigurdsson H.
(1992) GCA, 56, 2113-2129. [4] Chaussidon et al. (1994) this
volume. [5] Sigurdsson H. et al. (1992) EPSL, 109, 543-559.
[6] Brett R. (1992) GCA, 56, 3603. [7] Pope K. O. et al.
(1993) LPSC XXIV, 1165-1166. [8] Zahnle K. (1990) GSA Spec.
Paper, 247, 271-288. [9] Lewis J. et al. (1982) GSA Spec.
Paper, 190, 215-221. [10] Prinn R. and Fegley B. Jr. (1987)
EPSL, 83, 1-15. [11] Sheehan P. and Fastovsky D. (1992)
Geology, 20, 6, 1992. [12] Griffis K. and Chapman D. (1990)
Lethaia, 23, 379-383.[13] Spero, 1993, personal
communication.


Rocchia R.   Robin E.   Froget L.   Gayraud J.
   Ni-rich Spinels (Meteoric Spinels) as Indicators of the KT Event Timing

Introduction: Highly oxidized Ni-rich spinels are minerals
formed when a piece of extraterrestrial matter (rich in Ni)
is melted and oxidized in the Earth atmosphere (high O
fugacity). Such minerals have no counterpart in terrestrial
rocks but are systematically found in the fusion crust of
meteorites [1,2], in cosmic spherules [2,3,4], in the
ablation products found in a Jurassic horizon [5] and in
particles associated with oceanic impact debris from late
Pliocene sediments [6]. They are specific markers of
meteoritic material which interacted at high temperature
with the Earth atmosphere [2,7]. Moreover they are easily
synthesized in the laboratory by simple fusion of meteoritic
material [8]. Therefore, they can be labelled meteoric
spinels since their formation is associated with the
occurrence of meteors.

Relevance to the KT event: (1) KT boundary spinels. In
addition to high concentrations of PGEs (Platinum Group
Elements) KT boundary sediments also contain Ni-rich spinels
[9,-11]. These minerals, found in nearly all KT sections,
cannot be distinguished, by their morphology and
composition, from meteoric spinels derived from pure
meteoritic material. The appearance of these minerals in KT
boundary sediments, in exact coincidence with the planktonic
crisis, is a major argument supporting a cosmic origin for
the KT boundary biological crisis [2].

(2) A short catastrophic event: In all marine KT sites where
the Ir anomaly has been observed, the Ir enrichments are not
confined to the boundary clay layer: overabundances are also
observed on both sides of the boundary. Such extended
distributions are sometimes considered as the indication of
a long duration event. This observation is used to support
very different scenarios: the occurrence of a shower of
comets or a period of intense volcanic activity. However,
the extent of the Ir distribution is not necessarily due to
a long event and can be explained in another way. It might
also be due to the chemical properties of elemental Ir which
can stay for long in the ocean reservoir [12] and may have
diffused in the sediments after deposition. KT meteoric
spinels can be used to remove the ambiguity. These minerals
have a quite simple behaviour: they are locked in rather big
spheroids, a few tens or hundreds microns in size, which
cannot stay for long in sea water and which, once settled,
cannot move inside the lithified sediment. Therefore, the
stratigraphic distribution of meteoric spinels can provide a
less questionable information about the duration of the KT
event than Ir and other chemical markers.

In all KT sites analyzed for Ir and spinels, both markers
have their maxima at the same stratigraphical level. This is
a consequence of the fact that both have the same cosmic
origin. However, as expected, Ir and spinels stratigraphical
distributions are quite different. Ir exhibits broad
distributions but spinel distributions are always very
narrow. Data from the sites of El Kef, Caravaca and Hole
761C, leg 122, speak by themselves: anomalously high Ir
concentrations are found over about 1 meter when the
distribution of spinels never exceeds a few centimeters in
thickness (Fig. 1a, b and c). Two sections deserve a
particuliar attention.

El Kef: This section, the stratotype of the KT boundary in
the mesogean domain, is considered as the most complete one
for paleontological studies. Sediments of the KT transition,
nearly totally deprived of coarse detrital component, appear
to have been deposited under extremely quiet conditions. The
section offers the advantages of a rather low bioturbation
and a high sedimentation rate permitting the study of the KT
sequence with the highest resolution. Actually, this is the
section where we have observed the most different Ir and Ni-
rich distributions. The Ir anomaly is extremely dissymetric
and characterized by a rather narrow (2 cm wide) peak
followed by a low concentration tail extending in the basal
Danian over about 2 m. Overabundances are also oberved in
the last tens of centimeters of the upper Maastrichtian. The
integrated Ir flux, ~70 ng/g, is comparable with the world
average value. The maximum concentration of 18 ng/g is found
in a 1-2 mm thick brown-reddish layer in exact coincidence
with the mass extinction level (drop of carbonate). This
layer contains 5-10% of the total Ir flux. Ni-rich spinels
have a quite different distribution: 95 % of them are found
in this millimetric layer [13]. Taking into account the
sedimentation rates evaluated in the upper Cretaceous and
early Danian, this pulse-like distribution correspond to a
deposition time which cannot exceed 100 years. The infall of
meteoric spinels at El Kef is the result of a quasi-instant
event.

Gubbio Sections (Botaccione, Gubbio): This historical site
has been the subject of deep investigations. The finding of
shocked quartz and Ir secondary maxima in levels distinct
from the boundary layer has pushed Crocket et al. [14] to
propose that multiple events, probably explosive volcanic
eruptions, occurred over a period of more than 3.10^5 years.
These results initiated additional studies and a blind
sampling investigation by different laboratories. Available
data [15] show that secondary Ir maxima are real. However,
it is not demonstrated that these maxima result from
multiple Ir events or from post depositional diffusion and
dissolution. Meteoric spinels can provide conclusive
informations about that point. Our preliminary data show
that, in the Bottacione section, these minerals are not
present in secondary Ir levels: they are found only, like in
other sections, in the boundary clay. The stratigraphical
distribution of meteoric spinels, even if the cosmic origin
of these minerals is ignored, does not support the volcanic
hypothesis of Crocket et al. [14]. Long duration mechanisms
are excludes: all data points to a single cosmic event
occurring right at the paleontological boundary.

Conclusion: In all KT sections analyzed so far, the
stratigraphic distribution of meteoric spinels is extremelly
narrow. This indicates that the meteoritic material,
responsible for the formation of spinels and also for the Ir
enrichments, was accreted during a single and extremelly
brief event, or series of events, which occurred in exact
coincidence with the planctonic crisis. This result is
inconsistent with scenarios based on continuous or
repetitive long duration events. In particular it excludes
the occurrence of a comet shower over several 10^5 years
produced by the perturbation of the Oort cloud at the time
of the KT transition.

References: [1] Brownlee D. E. et al. (1975) JGR, 80, 4917-
4924. [2] Robin E. et al. (1992) EPSL, 108, 181-190. [3]
Robin E. (1988) Thesis, University of Paris XI, 131. [4]
Koeberl C. et al. (1989) GCA, 53, 937-944. [5] Jihanno C. et
al. (1988) LPSC 18, 623-630. [6] Margolis S. et al. (1991)
Science, 251, 1594-1597. [7] Robin E. et al. (1994) this
volume. [8] Gayraud J. et al. (1994) this volume. [9]
Montanari A. et al. (1983) Geology, 11, 668-671. [10] Smit
J. et al. (1984) Nature, 310, 403-405. [11] Kyte F. et al.
Geology, 14, 485-487. [12] Goldberg, E. D. et al. (1986)
Applied Geochem., 1, 227-232. [13] Robin, E. et al., (1991)
EPSL, 107, 715-721. [14] Crocket et al. (1988) Geology. [15]
Rocchia et al., (1990) EPSL, 99, 206-219.


Robin E.   Froget J.   Gayraud J.   Rocchia R.   Turpin L.
   Characteristics and Origin of Spinel-bearing Spheroids at the KT Boundary

Introduction: Spinel-bearing spheroids have been reported at
several KT boundary sites throughout the world [1]. Although
the origin of these spheroids is not clearly established,
they are interpreted as condensed products deriving from the
cloud of vaporized projectile and target material produced
by a large asteroid impact [1-5]. The occurrence of numerous
meteoric spinels in these spheroids shows that at least some
of them could derive from meteoritic debris heated and
oxidized in the atmosphere [6]. Here we report on the
mineralogical and chemical features of the spinel-bearing
particles at the KT boundary and we discuss on their
possible origin(s).

Samples and Procedure: From the 18 KT sites we have
investigated so far [7], only DSDP site 577 still contain
well-preserved spinel-bearing spheroids. At the other sites,
the spheroids have experienced intensive alteration and in
most cases only spinels have survived. However, few
spheroids from Beloc (Haoti), Petriccio (Italy), El Kef
(Tunisia), and Caravaca (Spain) have also been recovered and
studied. At site 577 about 100 particles were analyzed for
Ir, Ni, Co, Cr, Sc, and rare-Earth elements with a high-
purity Ge gamma-ray detector [8]. At other sites, particles
were analyzed only for Ir. Debye-Scherrer and x-ray
diffraction analyses have been performed on a set of
particles in order to characterize mineralogical phases.
Individual particles were then mounted in epoxy and polished
for detailed scanning electron microscopy and electron
microprobe analysis.

Results: DSDP site 577 (North Pacific Ocean). Two distinct
populations of well-preserved spinel-bearing particles have
been found right at the boundary [8]: spheroids with
dendritic spinel textures uniformly dispersed in the whole
particle and irregularly shaped fragments with tiny spinels
essentially confined to the rim. Chemical analyses of
spinels contained in the particles are consistent with those
previously reported [4]: they have high Ni (> 1%) contents,
low Ti contents (<1%) and a high Fe oxidation state (Fe3+/Fe
(sub)total ~ 96 atom%, see reference [7]). In addition to
spinel, augite crystals are observed and pure augite
spherules, some of them containing spinels, are found.
Although most of the minerals originally present in the
particles are now replaced by smectite, their morphology and
composition have been preserved. In 86 analyzed spheroids,
Ir contents range from less than 2 ng/g up to 610 ng/g,
consistent with the range of Ir contents observed in cosmic
ablation spheres [8]. In fragments (11 particles analyzed as
yet), Ir contents range from 95 ng/g to 1200 ng/g, with an
average value of 530 ng/g, very close to the Ir abundance of
meteorites. We also found systematic Cr, Co, and Ni
enrichments in spheroids and fragments as well, with
abundances sometimes approaching those in meteorites, and
low REE contents, with a relatively flat pattern (Fig. 1),
which excludes a mixing with more than 10% of terrestrial
material. The small Ce anomaly may be attributed to
postdepositional contamination by the surrounding REE-rich
sediment.

Beloc (Haiti): Only spheroids are found. Spinel are embedded
in a smectitic matrix enriched in iron oxide (goethite). The
spinel composition differs from the one measured at site 577
[7]. No augite crystals are observed. Ir analyses of 10
spheroids reveal concentrations ranging from less than 2
ng/g up to 100 ng/g.

Petriccio (Italy) and El Kef (Tunisia): At both sections,
many flatttened spheroids are found but few of them contain
spinels. At Petriccio, the particles are altered and filled
with glauconite, K-feldspaths and iron hydroxide. At El Kef
they are mainly composed of goethite and smectite. Rare
kaolinite and K-feldspaths are present. No augite crystals
are observed. Ir analysis on a set of spheroids from
Petriccio gives an average content of 40 ng/g, slightly
lower than previously measured [2]. Spinel composition at
these sites still differs from the one observed at site 577
and at Beloc and also differs between them [7].

Caravaca (Spain): K-feldspaths spheroids are abundant [2]
but none of them contain spinels and their Ir content is
lower than in the surrounding matrix. Numerous lenses almost
entirely made of spinels are present and their Ir content is
comparable to the one in the surrounding matrix (~70 ng/g).

Discussion and Conclusions: From the study of the spinel-
bearing particles at site 577 we can draw out the following
conclusions: (1) morphological, mineralogical and chemical
evidences show that spinel-bearing particles at the KT
boundary derive from nearly pure meteoritic debris. If KT
spheroids are formed in the cloud of vaporized projectile
and target material, then the target contribution is
negligible. (2) The occurrence of partially melted
meteoritic debris is inconsistent with a condensation
process.

This rule out some proposed mechanisms for the origin of
spinel-bearing spheroids at the KT boundary. Formation
during a cratering event on land is unlikely as no
meteoritic material survived in such an event. In addition
the contribution of the target is always high in impact on
land and the amount of meteoritic material in impact-melt
rocks is in general <1% relative to C1-chondrites and still
lower in the long range ejecta. At last but not least,
impact products, i.e., tektites and microtektites are
extremely reduced objects free of any crystalline inclusion.
Condensation from the bolide itself is also unlikely as
spinel crystallization requires high O fugacities [9] and
that there is no available O in meteorites.

Considering the formation conditions of spinels [9] we can
envisage two possible origin for KT spheroids: (1) ablation
and oxidation of meteoritic debris in the atmosphere [6].
This hypothesis implies that several objects resulting
either from an oblique impact or from a fragmented comet are
involved [7]. (2) Oxidation of meteoritic debris in the
water vapor cloud produced by an impact in the ocean. In
such an event pure meteoritic debris can be recovered and
highly oxidized spinel can be formed [10].

References: [1] Smit J. and Romein A. J. T. (1985) EPSL, 74,
155-170. [2] Montanari A. et al. (1983) Geology, 11, 668-
671. [3] Smit J. and Kyte F. T. (1984) Nature, 310, 403-405.
[4] Kyte F. T. and Smit J. [1986] Geology, 14, 485-487. [5]
Bohor B. F. et al. (1986) EPSL, 81, 57-66. [6] Robin E. et
al. (1992) EPSL, 108, 181-190. [7] Robin E. et al. this
volume. [8] Robin E. et al. (1993) Nature, 363, 615-617. [9]
Gayraud J. et al. this volume. [10] Margolis S. V. et al.
(1991) Science, 251, 1594-1597.

Table 1, which appears here in the hard copy, shows
elemental contents determined by INAA for 97 spinel-bearing
debris (86 sperules and 11 fragments) from site 577.

Figure 1, which appears here in the hard copy, shows REE
contents relative to chondrites.


Kyte F. T.   Bostwick J. A.   Zhou L.
   The KT Boundary on the Pacific Plate

We will review our results to date on the geochemistry and
impact mineralogy of the Cretaceous Tertiary (KT) boundary
on the Pacific Plate. Over the past several years we have
been relatively successful at locating the KT boundary at
Pacific sites. In addition to the two KT boundaries in
calcareous sequences at DSDP Sites 465 and 577, the boundary
is now confirmed in three pelagic clay cores--LL44-GPC3,
DSDP 576, and DSDP 596--and in the relatively barren
boundary sequence at ODP Site 803. At each locality the KT
boundary is characterized by an Ir anomaly, spherules, and
shocked quartz. We have extracted and analyzed
magnesioferrite spinel from each site except DSDP 465.

These six sites encompass a region of ~20 million km^2 (4%
of the Earth's surface), but they must be used as
representative of a region at least five times larger. The
pre-Cenozoic portion of the Pacific Plate was situated
generally southwest of the Yucatan Peninsula at 65 Ma and
these sites were ~6500 (GPC3) to ~10,000 km (DSDP 596)
distant from the Chicxulub impact structure. Of the six
sites, all except DSDP 465 are bioturbated and sediments
from this site were severely disturbed by drilling. Thus it
is nearly impossible to perform detailed analyses of fine-
scale stratigraphic units in this portion of the world.

In DSDP Hole 465A we were fortunate to recover a clast
containing 1.3 cm of stratigraphy including the KT boundary.
In this sample, a distinct boundary clay ~2-3 mm thick was
recovered. Detailed geochemical analyses were performed on
four layers from this clast in conjunction with leaching
experiments and isotopic analyses [1]. Siderophile element
abundances are quite variable across the clast (e.g., Ir/Au
varies by a factor of 65). However, as at other sites [2],
integration of the samples results in siderophile element
abundances approaching chondritic values with element/Ir
ratios for Ni, Pd, Os, Pt, and Au within a factor of 2 of
chondritic ratios. These data are consistent with a
chondritic source for the siderophiles, but require
fractionation by impact or terrestrial processes, such as
diagenetic mobilization.

The KT boundary at each site is characterized by a large Ir
anomaly. Ir fluences range from 61 ng cm^-2 in DSDP Hole
577B [3] to 320 ng cm^-2 in DSDP Hole 596 [4]. At each of
the three pelagic clay sites (GPC3, 576, and 596) Ir
concentrations have been measured across the entire
Cenozoic. These Ir profiles are sufficient to rule out the
existence of global Ir anomalies comparable to the KT signal
at any time since 65 Ma. These data confirm earlier results
[5] and cast serious doubt on hypotheses of periodic comet
showers as a cause of mass extinctions.

These sediments are particularly amenable to trace mineral
analyses. Sediment samples (typically 0.5 g) are
disaggregated and dispersed, the fine-fraction pipetted off,
and biogenic and hydrogenous components dissolved by
sequential leaching. This procedure should quantitatively
retain trace mineral fractions >30 micrometers in size.
Residues typically consist principally of quartz and
magnesioferrite spinel. Quartz is defined as shocked if
planar deformation features (PDF) are observed during SEM
analyses. Approximately 40% of the quartz grains do not have
PDF, but these are also apparently derived from the KT
impact event. Large grains of unshocked and shocked quartz
from DSDP 596 [4] and DSDP 576 [6] are restricted to the Ir-
enriched interval. Terrigenous eolian components apparently
do not contribute to the >30-micrometer fraction at these
deep-sea sites.

The largest shocked quartz grain recovered was from GPC3
with a maximum diameter of 230 micrometers. We assume that
larger grains have been deposited on the Pacific Plate, but
our ability to define a maximum size is inhibited by small
sediment sample sizes. Size distributions of shocked quartz
at three North Pacific sites are similar, with mean maximum
diameters of 58, 61, and 76 micrometers (for the >30-
micrometers fraction) at DSDP 576, GPC3 and DSDP 465,
respectively. This is considerably less than the mean
maximum diameters (typically ~200 micrometers) reported for
North American sites [7], but consistent with the ~70
micrometers reported for DSDP 596 [4]. We estimate a total
fluence of 1200 grains cm^-2 of shocked quartz >30
micrometers at DSDP 576, comparable to the 1800 grains cm^-2
at South Pacific DSDP site 596. Our data are not yet
sufficient to quantify the fluence of grains at other sites,
but we expect similar results.

Electron microprobe analyses of magnesioferrite spinel
confirm previous results that indicated regional variations
in KT boundary spinel compositions [8]. Our extensive data
(349-point analyses) from all Pacific sites except DSDP 465
show that spinel on the Pacific Plate is compositionally
distinct from non-Pacific sites (Furlo, 127 points;
Caravaca, 19p; DSDP 524, 7p; ODP 761, 13p). The most
striking differences are higher ratios of Fe2O3/FeO and
Al2O3/Fe2O3 at the Pacific sites with the most extreme
values observed at northwestern DSDP sites 576 and 577.
These data indicate a clear compositional asymmetry of KT
impact debris relative to the proposed impact site in the
Yucatan. Models to explain these variations could include
asymmetric ejection from Chicxulub resulting from processes
such as low-angle impact or heterogeneous target materials,
or heterogeneous accretion of multiple projectiles [9,10].
The compositions of Pacific spinel indicate crystallization
at high O fugacities and high temperatures. We suspect that
these compositions, in part, reflect chemical fractionation
during vapor condensation.

Mineral and chemical data from the KT boundary on the
Pacific Plate must be accounted for by models that are used
to explain KT boundary phenomena. Key constraints imposed by
data from this and earlier studies include (1) significant
amounts of shocked debris and spherules thousands of
kilometers from any continental source, (2) chemical
asymmetry of high-temperature mineral phases, and (3)
absence of global Ir anomalies during the Cenozoic. The
Mesozoic portion of the Pacific Plate is now only a small
remnant of a vast region of the Earth s surface, most of
which was lost by subduction. However, the KT boundary
appears to be very well preserved over most of the plate,
commonly buried by only 20 to 60 m of clay, and readily
accessible by ocean-drilling or long-piston coring.

Supported by NSF EAR-9118701.

References: [1] DePaolo D. J. and Kyte F. T. (1984) LPS XV,
220-221. [2] Kyte F. T. et al. (1985) EPSL, 73, 183-195. [3]
Michel H. V. et al. (1985) Init. Rep. DSDP, 86, 533-538. [4]
Zhou L. et al. (1991) Geology, 19, 694-697. [5] Kyte F. T.
and Wasson J. T. (1986) Science, 232, 1225-1229. [6]
Bostwick J. A. and Kyte F. T. (1993) LPS XXIV, 157-158. [7]
Izett G. A. (1990) GSA Spec. Paper, 249, 100. [8] Kyte F. T.
and Smit J. (1986) Geology, 14, 485-487. [9] Kyte F. T. et
al. (1980) Nature, 288, 651-656. [10] Robin E. et al. (1993)
Nature, 363, 615-617.


Holsapple K. A.
   An Estimation of the Measures of the Chicxulub Cratering Event

1. Introduction: The impact of a large asteroid or comet
into the Earth sets in motion a very complex event with
severe consequences. The energy from the impact of a 10 km
body at 20 km/sec velocity is 10,000 times greater than the
sum of the energy from the simultaneous detonation of all
nuclear devices in the world [1]. That energy is deposited
into what is, compared to the range of the effects, a single
spot in the lithosphere. The consequences of such an impact
are well outside any known experience, and require extreme
extrapolations of limited data to estimate.

The most important link to the estimation of the other
aspects of the crater formation is the remaining final
crater. Here we review the existing methods of estimating
the relations between the conditions of the impact: the
impactor composition, size and velocity; and the resulting
effects: crater size, ejecta deposition, dynamic flow
fields, melt volumes, and so on.

2. The Database: The quantification of cratering phenomena
are based on results of four types. Those types and some
comments on their importance and limitations follow: (1)
Experiments in the laboratory. Since the 1960s experimenters
have formed centimetersized craters in a variety of
materials at velocities up to about 6 km/sec. In the 1980s,
experiments were also performed at high gravity on a
centrifuge [2]. Primarily from these data, two distinct
physical regimes for cratering have been identifi