A Possible Tektite Strewn Field in the Argentinian Pampa

Abstract

Impact glass associated with 11 elongate depressions in the Pampean Plain of Argentina, north of the city of Rı́o Cuarto,
was suggested to be proximal ejecta related to a highly oblique impact event. We have identified about 400 additional elongate
features in the area that indicate an aeolian, rather than an impact, origin. We have also dated fragments of glass found
at the Rı́o Cuarto depressions; the age is similar to that of glass recovered 800 kilometers to the southeast. This material
may be tektite glass from an impact event around 0.48 million years ago, representing a new tektite strewn field.

Full text

the BSB-S
2
/THPMA-MMA material system
can be used to pattern complex 3D structures
and suggests that microfluidic structures and
even microoptical structures, such as grat-
ings, waveguides, and photonic lattices, can
be fabricated readily. The low irradiation
power used to pattern the structure also dem-
onstrates the high two-photon sensitivity of
this microfabrication material system.
References and Notes
1. S. Maruo, O. Nakamura, S. Kawata, Opt. Lett. 22, 132
(1997).
2. B. H. Cumpston et al.,Nature 398, 51 (1999).
3. S. Kawata, H.-B. Sun, T. Tanaka, K. Takada, Nature
412, 697 (2001).
4. S. Maruo, K. Ikuta, Proc. Soc. Photo-Opt. Instrum.
Eng. 3937, 106 (2000).
5. H. E. Pudavar, M. P. Joshi, P. N. Prasad, B. A. Reinhardt,
Appl. Phys. Lett. 74, 1338 (1999).
6. W. Denk, Proc. Natl. Acad. Sci. U.S.A. 91, 6629
(1994).
7. R. M. Williams, D. W. Piston, W. W. Webb, Fed. Am.
Soc. Exp. Biol. J. 8, 804 (1994).
8. K. D. Belfield et al.,J. Phys. Org. Chem. 13, 837 (2000).
9. M. Albota et al.,Science 281, 1653 (1998).
10. M. Rumi et al.,J. Am. Chem. Soc. 122, 9500 (2000).
11. C. G. Willson, H. Ito, J. M. J. Fre´chet, T. G. Tessier, F. M.
Houlihan, J. Electrochem. Soc. 133, 181 (1986).
12. J. M. J. Fre´chet et al.,J. Imag. Sci. 30, 59 (1986).
13. J. M. J. Fre´chet et al.,inFunctional Polymers,D.E.
Bergbreiter, C. R. Martin, Eds. (Plenum, New York,
1989), pp. 193–200.
14. R. W. Angelo et al., U.S. Patent 5,047,568 (1991).
15. R. W. Angelo et al., U.S. Patent 5,102,772 (1992).
16. F. D. Saeva, Adv. Electr. Trans. Chem. 4, 1 (1994).
17. F. D. Saeva, D. T. Breslin, P. A. Martic, J. Am. Chem.
Soc. 111, 1328 (1989).
18. W. Zhou, S. M. Kuebler, D. Carrig, J. W. Perry, S. R.
Marder, J. Am. Chem. Soc. 124, 1897 (2002).
19. J. March, Advanced Organic Chemistry ( Wiley, New
York, ed. 3, 1985).
20. Details of the molecular synthesis and polymeriza-
tion studies are available on Science Online at www.
sciencemag.org/cgi/content/full/296/5570/1106/
DC1.
21. Solutions of BSB-S
2
in acetonitrile (4.0 ⫻10
⫺4
M)
were irradiated at 400 nm with either a Xenon lamp
or a frequency-doubled mode-locked Ti:sapphire la-
ser. At this concentration, more than 99% of the light
was absorbed. The photogenerated acid was titrated
by the addition of excess rhodamine B base in ace-
tonitrile (6.0 ⫻10
⫺5
M after addition) and quantified
spectrophotometrically from the absorbance of pro-
tonated rhodamine B base (27). The acid yield in-
creased linearly with both the excitation intensity
and time, consistent with acid generation being ini-
tiated by one-photon excitation.
22. M. Shirai, M. Tsunooka, Prog. Polym. Sci. 21, 1 (1996).
23. Irradiating 4 ml of BSB-S
2
at 8 ⫻10
⫺7
M in 80% (by
volume) cyclohexene oxide in dichloromethane at
419 nm yielded polymer that precipitated after the
addition of methanol.
24. Y. Boiko, J. M. Costa, M. Wang, S. Esener, Opt. Express
8, 571 (2001).
25. Freestanding columns of cross-linked polymer were
generated when a 10 mM solution of BSB-S
2
in 20%
Epon SU-8 (Shell)/80% 4-vinyl-cyclohexene dioxide by
weight was irradiated at 745 nm for5satathreshold
pulse energy of 940 ␮J, using focused 5-ns pulses (at a
10-Hz repetition rate, 500-mm focal length, and 5-mm-
diameter spot size at the focusing lens). Under similar
conditions, BSB-S
2
also initiates the polymerization of
20% SU-8/80% cyclohexene oxide, neat Araldite
CY179MA, and solid films of SU-8.
26. Freestanding microstructures were fabricated (2,28)
by irradiating solid thin films of 1 weight % (wt %)
BSB-S
2
in SU-8 using tightly focused [numerical ap-
erture (NA) ⫽1.4] 80-fs pulses (at an 82-MHz rep-
etition rate, 745 nm) at average powers as low as 2.4
mW and then dissolving the unexposed resin in an
organic solvent.
27. G. Pohlers, J. C. Scaiano, R. Sinta, Chem. Mater. 9,
3222 (1997).
28. S. M. Kuebler et al.,J. Photopolym. Sci. Technol. 14,
657 (2001).
29. C. Xu, W. W. Webb, J. Opt. Soc. Am. B 13, 481 (1996).
30. Support from the Air Force Office of Scientific Re-
search (grants F49620-99-1-0019 and F49620-97-1-
0014), NSF (grants CHE-0107105 and DMR-
9975961), the Office of Naval Research (grants
N00014-95-1319 and N000141-01-1-0633), the
NSF-funded Cornell Nanobiotechnology Center, and
the Defense University Research Instrumentation
Program (grants N00014-99-1-0541 and F49620-00-
1-0161) is gratefully acknowledged.
30 November 2001; accepted 19 March 2002
A Possible Tektite Strewn Field
in the Argentinian Pampa
P. A. Bland,
1
* C. R. de Souza Filho,
2
A. J. T. Jull,
3
S. P. Kelley,
4
R. M. Hough,
5
N. A. Artemieva,
6
E. Pierazzo,
7
J. Coniglio,
8
L. Pinotti,
8
V. Evers,
9
A. T. Kearsley
10
Impact glass associated with 11 elongate depressions in the Pampean Plain of
Argentina, north of the city of Rı´o Cuarto, was suggested to be proximal ejecta
related to a highly oblique impact event. We have identified about 400 addi-
tional elongate features in the area that indicate an aeolian, rather than an
impact, origin. We have also dated fragments of glass found at the Rı´o Cuarto
depressions; the age is similar to that of glass recovered 800 kilometers to the
southeast. This material may be tektite glass from an impact event around 0.48
million years ago, representing a new tektite strewn field.
Comets and asteroids that impact planets create
circular craters at impact angles between ⬃15°
and 90° (measured from the surface); when the
angle ␪is less than ⬃15°, craters become elon-
gate in shape. On Earth, the only confirmed
low-angle impact structures are the series of
elongate craters at Rı´o Cuarto, Argentina, esti-
mated to be ⬍0.01 million years (Ma) (1)to
⬍0.005 Ma (2,3) in age. Rı´o Cuarto is also the
largest object known to have impacted Earth in
the last 0.1 Ma, an impact that may have been
witnessed by early inhabitants of the Pampean
Plain (1). The largest feature (64°10⬘W,
32°45⬘S) has dimensions of 4.5 km by 1.1 km
and is considered to correspond to the first
impact; an impactor initially 150 to 300 m in
diameter then fragmented and ricocheted to the
south to produce 10 additional elongate depres-
sions (1). Impact formation of the depressions
was questioned (4), but samples of meteorite,
and detailed analysis of glass fragments found at
the site, supported the impact hypothesis (1,5).
Rı´o Cuarto glass is typically vesicular,
with abundant loess inclusions, and lower
surfaces that appear to be sand casts (1,5).
Vesicle-poor splash forms (elongate drops)
are also common (1,5). Although most sam-
ples do not show evidence of much meteoritic
contamination, some fragments exhibit high
Cr [⬎1000 parts per million ( ppm)] and Ni (1
to 2 weight %) concentrations, and one has
metallic Fe and Fe-Ni spherules (5). Sidero-
phile-element abundance and rare earth ele-
ment pattern suggest a chondritic impactor
(6). In addition, very low water content
(characteristic of tektites and other impact
glasses) is typical; vesicular glass contains
⬃0.1 weight % water, whereas splash-form
glass contains 0.06 weight % (5). The pres-
ence of lechatelierite (5), homogeneous oxide
distribution, and high silica content are addi-
tional characteristics in common with tektites
(1). Overall, the glass has a composition sim-
ilar to that of the loessoid sediments that
cover the Pampa (1,7). Because loess only
occurs to a depth of ⬍50 m over a metamor-
phic basement, it was suggested that the glass
originated in a low-angle impact that did not
excavate deeply (5).
We have conducted an extensive remote-
sensing study of the Rı´o Cuarto site and the
surrounding Pampean Plain, using CORONA
and EOS Terra-ASTER multispectral, high-
resolution satellite imagery. This survey re-
veals several hundred elongate depressions
with high length-to-width ratios (Figs. 1, S1,
and 2); 403 are ⬎200 m in their long axes,
201 ⬎1 km, and 6 ⬎5 km. Long-axis orien-
tations vary throughout the region: north-
1
Planetary and Space Sciences Research Institute, The
Open University, Milton Keynes MK7 6AA, UK.
2
Insti-
tuto de Geocieˆncias, Universidade Estadual de Campi-
nas, Campinas, Brazil.
3
NSF–Arizona Accelerator Mass
Spectrometry Laboratory, University of Arizona, 1118
East Fourth Street, Tucson, AZ 85721, USA.
4
Depart-
ment of Earth Sciences, The Open University, Milton
Keynes MK7 6AA, UK.
5
Earth and Planetary Sciences,
Western Australian Museum, Francis Street, Perth,
Western Australia 6000.
6
Institute for Dynamics of
Geospheres, Russian Academy of Sciences, Leninsky
Prospect 38/6, Moscow, Russia 117939.
7
Planetary
Science Institute, Tucson, AZ 85705, USA.
8
Departa-
mento de Geologı´a, Universidad Nacional de Rı´o
Cuarto, 5800 Rı´o Cuarto, Co´rdoba, Argentina.
9
Insti-
tute of Educational Technology, The Open University,
Milton Keynes MK7 6AA, UK.
10
Geology (Biological
and Molecular Sciences), Oxford Brookes University,
Headington, Oxford OX3 OBP, UK.
*To whom correspondence should be addressed. E-
mail: p.a.bland@open.ac.uk
REPORTS
www.sciencemag.org SCIENCE VOL 296 10 MAY 2002 1109

northeast–south-southwest in the north,
northeast-southwest in the south, and north-
northwest–south-southeast in the west. Field
visits to 52 of these features [including sev-
eral of those previously identified (1)] con-
firm that they are morphologically similar to
the original 11 (1): Most depressions have
rims 3 to 10 m above the surrounding plain,
some of which are degraded, and bases 3 to
10 m below the plain. Several elongate fea-
tures, including those described by Schultz
and Lianza (1), have a “lag” deposit of local
country rocks preserved on the crater floor;
these samples, and a hardground layer that is
occasionally exposed, appear to be unaffect-
ed by impact. In addition, we recovered two
meteorites from craters D and E (craters that
also yielded abundant glass; the largest sam-
ple was 17 cm by 9 cm by 6 cm). Glass was
found in two of the new structures, and at a
road section 400 km to the south, near the
town of Santa Rosa.
Assuming that all the depressions have a
similar origin, to account for the hundreds of
additional similar features we see would re-
quire a larger initial object (⬎0.5 km in di-
ameter) that perhaps fragmented in the atmo-
sphere before impact and ricochet. Estimates
of impact rates for bodies in this size range
vary widely, from 0.038 Ma (8) to 0.38 Ma
(9). Such highly elongate features require an
impact angle ␪⬍5°, which occurs only
0.75% of the time (10,11); a 0.5-km-diame-
ter object should impact Earth at ⬍5° about
every 5 to 50 Ma. A 0.5- to 0.7-km projectile
also appears to be the minimum-diameter
stony body that can penetrate Earth’s atmo-
sphere at ␪⬍5° (12). Impact rates for bodies
of the size postulated by Schultz and Lianza
(1) are more frequent by an order of magni-
tude, but it appears unlikely that they could
penetrate Earth’s atmosphere at low angle.
There is an aeolian landform that matches
a number of the morphological characteristics
of the Rı´o Cuarto craters. Upper Holocene
parabolic dunes have been described in the
Pampean Plain (13). These dunes typically
form in semiarid environments around a
“blowout,” or deflation, where the soil sur-
face has been broken and partially removed
(14 ). A dune forms at the downwind edge of
the blowout deflation and migrates down-
wind, the crest extending around the elongat-
ed blowout. Deflation of the interior contin-
ues, forming an elongate feature with a raised
rim and a base that is lower than the sur-
rounding plain (14 ). Before the discovery of
impact-related materials (1,3), the Rı´o
Cuarto depressions and those further south
were considered to be aeolian in origin (15–
17 ). The variation in long-axis orientations
over the region is difficult to reconcile with
the breakup and ricochet of a single impactor,
but it supports an aeolian formation mecha-
nism, with long axes consistent with prevail-
ing wind directions recorded by meteorology
stations. In the Rı´o Cuarto region, where
depressions are oriented north-northeast–
south-southwest, observations over 19 years
find that 38% of winds blow from the north
and 53% from the northeast (18). To the
southwest, in the region of San Luis where
depressions are oriented northeast-southwest,
28% of winds are from the north, 32% from
the northeast, and 35% from the east (19).
Using accelerator mass spectrometry
(AMS), we analyzed cosmogenic
14
C content
in the organic carbon fraction of soil recovered
from craters J, D, and E to determine the expo-
sure age of these surfaces (20). Our results
suggest a maximum age of 0.004 Ma for these
features, consistent with previous estimates
(1–3).
Schultz and Lianza (1) also described two
meteorites found in crater D, both fusion crust-
ed ordinary chondrites, and suggested that they
were fragments of the Rı´o Cuarto impactor. We
have identified and analyzed two new samples,
a weathered ordinary chondrite and a basaltic
achondrite, recovered from craters D and E,
respectively. AMS analysis reveals a
14
C ter-
restrial age of 0.036 ⫾0.004 Ma for the chon-
Fig. 2. Schematic map of the area, showing elongate features (black), associated lakes (gray), local
drainage, and the area originally described by Schultz and Lianza (1) (shaded). General orientations
of elongate features are indicated by arrows. Additional glass localities (see text) are indicated in
red on the context map.
Fig. 1. EOS Terra-ASTER satellite images show-
ing (inset) a portion of the crater field origi-
nally defined by Schultz and Lianza (1), and
elongate features partially infilled with lakes 60
km to the east.
REPORTS
10 MAY 2002 VOL 296 SCIENCE www.sciencemag.org1110

drite, and ⬎0.052 Ma for the achondrite. Vari-
ation in terrestrial age and weathering, the re-
covery of both chondrites and achondrites, and
meteorite terrestrial ages older than surface ages
for the depressions suggest a meteorite accumu-
lation rather than fragments of a single impac-
tor. Meteorite accumulations are commonly
found in blowout deflations (21) where samples
with a range of compositions and terrestrial
ages are revealed after soil is removed.
The geomorphologic and meteoritic evi-
dence suggests that dune formation occurred
⬃0.004 Ma ago, removing meters of loessoid
sediment and exposing rocks previously con-
tained in the sediment column: samples of
country rock, meteorites, and any other ma-
terial deposited onto the surface before defla-
tion. However, it is clear that the glass found
at Rı´o Cuarto is derived from an impact; it
may therefore be distal, rather than proximal,
ejecta. A typical impact (i.e., not highly
oblique) producing a 10-km crater might de-
posit loess-derived glass hundreds of kilome-
ters from the impact site but may itself have
been rapidly infilled.
We have attempted to determine the age of
this event by Ar-Ar dating of Rı´o Cuarto glass.
Analysis of crushed vesicular glass yielded
dominantly atmospheric argon. This is typical of
highly vesicular impactites (22). A sample of
inclusion-free Rı´o Cuarto glass yielded much
higher radiogenic contents (as much as 75%).
Individual fragments of this glass were analyzed
by total fusion, the results giving an isochron
indicating an age of 0.57 ⫾0.1 Ma (2␴error).
This is similar to the results obtained with glass
from Southern Buenos Aires Province [with an
age of 0.46 ⫾0.03 Ma (23)], found in loess
sediments on the coast near the city of Neco-
chea. We suggest that the glass fragments found
at Rı´o Cuarto and Necochea were formed by a
common impact that occurred at ⬃0.48 Ma.
We recently recovered glass from a new
Pampean locality near the city of Santa Rosa.
This material is highly vesicular and has a
bulk chemistry (Table 1) intermediate be-
tween that of clear and light-brown vesicle-
abundant impactite from the Rı´o Cuarto
depressions (1).
The fact that glass is derived from a thin
loess cover does not constrain the size of the
buried crater. Tektites originate from the up-
permost layer of the target surface (24,25);
however, they may result from relatively
large impacts that excavate much deeper cra-
ters. Recent modeling studies rule out low-
angle impacts as possible causes of tektite
strewn fields (26 ). We performed a hydro-
code simulation to model tektite production
from the impact of a 500-m-diameter projec-
tile at ␪⫽30°, with an impact velocity of 18
km/s, into a Pampean-like target (a porous
50-m quartzite surface layer over a granitic
basement). The final crater produced in this
simulation would be ⬃5 km in diameter (27 ).
The results (fig. S2) show that a substantial
portion of distal glass ejecta originates from
the porous upper layer: 90% of the melt from
this layer is ejected early and at high velocity,
with little or no contamination from projectile
and basement material.
Finally, we can place some lower limit on
the spatial distribution of glass from this impact
in the Pampean plain. Schultz et al. (3,28)
recovered vesicular glass with low water con-
tent (0.003 weight %) 500 km downrange of the
Rı´o Cuarto depressions and noted that this ma-
terial and the Rı´o Cuarto glass were of very
similar composition. With the Necochea glass
(800 km southeast of Rı´o Cuarto) and the Rı´o
Cuarto samples displaying similar ages, and the
glass we recovered at Santa Rosa (400 km
south) and the Rı´o Cuarto samples showing
similar composition, this suggests that geneti-
cally related glass samples have a wide geo-
graphical distribution. Montanari and Koeberl
(29) identify seven defining characteristics of
tektites (30) and recommend that the term “tek-
tite” should be reserved for glasses that fulfill
most of these criteria. With the recognition that
Pampean glass occurs as distal ejecta in a geo-
graphically extended strewn field, this material
appears to fulfill all other defining criteria for
tektites (30), so it seems plausible that they are
representatives of a widespread tektite strewn
field in Argentina with an age of ⬃0.48 Ma.
Further work will determine the extent of this
strewn-field, as well as the cogenesis of glass at
different locations and of different morpholo-
gies, and seek the location of the source crater.
The widespread distribution of glass suggests a
source crater ⬎5 km in diameter. Given that
loessoid sedimentation was ongoing throughout
the last 0.5 Ma, we suggest that a recent, pos-
sibly well-preserved complex crater remains to
be discovered beneath the Pampean Plain of
Argentina.
References and Notes
1. P. H. Schultz, R. E. Lianza, Nature 355, 234 (1992).
2. J. A. Grant, P. H. Schultz, Lunar Planet. Sci. XXIII
(abstract), 439 (1992).
3. P. H. Schultz et al.,Lunar Planet. Sci. XXIII (abstract),
1237 (1992).
4.
A. L. Bloom, GSA Abstr. Prog. 24 (abstract), 136 (1992).
5. P. H. Schultz, C. Koeberl, T. Bunch, J. Grant, W. Collins,
Geology 22, 889 (1994).
6. C. Koeberl, P. H. Schultz, Lunar Planet. Sci. XXIII
(abstract), 706 (1992)
7. A. A. Aldahan, C. Koeberl, G. Possnert, P. H. Schultz,
GFF 119, 67 (1997).
8. D. Morrison, C. R. Chapman, P. Slovic, in Hazards Due
to Comets and Asteroids, T. Gehrels, Ed. (Univ. of
Arizona Press, Tucson, 1994), pp 59–91.
9. D. W. Hughes, in Impacts and the Early Earth,I.
Gilmour, C. Koeberl, Eds. (Springer-Verlag, Berlin,
2000), pp. 327–341.
10. G. K. Gilbert, Bull. Philos. Cosmol. Wash. (D.C.)12,
241 (1893).
11. E. M. Shoemaker, in Physics and Astronomy of the
Moon, Z. Kopal, Ed. (Academic Press, San Diego, CA,
1962), pp. 283–359.
12. H. J. Melosh, Impact Cratering: A Geological Process
(Oxford Univ. Press, New York, 1989), p. 245.
13. M. H. Iriondo, Quat. Int. 57, 93 (1999).
14. R. U. Cooke, A. Warren, A. S. Goudie, Desert Geomor-
phology (UCL Press, London, 1993), p. 526.
15. J. L. Allione, Trabajo Final de Licenciatura (Univ. Nac.
De Rio Cuarto, unpublished, 1987).
16. M. P. Cantu, S. B. Degiovanni, Noveno Congreso Geo-
logico Argentino IV, 76 (1984).
17. M. T. Blarasin, M. L. Sanchez, Decimo Congreso Geo-
logico Argentino III, 297 (1987).
18. R. A. Seiler, R. Fabricius, V. H. Rotondo, M. Vinocur,
Agroclimatologı´a de Rı´o Cuarto—1974/93 (Publica-
cio´n de la Ca´tedra de Agrometeorologı´a, Facultad de
Agronomı´a y Veterinaria, Universidad Nacional
de Rı´o Cuarto), vol. 1, p. 68.
19. Source: Meteorological Station of the National Insti-
tute of Agricultural Technology (INTA, Instituto Na-
cional de Tecnologı´a Agraria) at San Luis city
(33°39⬘50⬙S, 65°24⬘37⬙W).
20. Samples of soil and paleosol layers were recovered
from craters D, E, and J. Although an attempt was
made to sample only stable surfaces, soil samples
from craters D and E gave recent ages. Samples of soil
from an exposed paleosol in the floor of crater J gave
the oldest age: 0.004 Ma.
21. M. E. Zolensky, H. M. Rendell, I. Wilson, G. L. Wells,
Meteoritics 27, 460 (1992).
22. P. W. Haines, R. J. F. Jenkins, S. P. Kelley, Geology 29,
899 (2001).
23. P. H. Schultz, M. Zarate, W. E. Hames, Meteorit.
Planet. Sci. 35 (abstract), A143 (2000).
24. D. K. Pal, C. Tuniz, R. K. Moniot, T. H. Kruse, G. F.
Herzog, Science 218, 787 (1982).
25. G. M. Raisbeck, F. Yiou, S. Z. Zhou, C. Koeberl, Chem.
Geol. 70, 120.
26. N. A. Artemieva, Lunar Planet. Sci. XXXII (abstract),
1216 (2001).
27. R. M. Schmidt, K. R. Housen, Int. J. Impact Eng. 5, 543
(1987).
28. P. H. Schultz, T. Bunch, C. Koeberl, W. Collins, Lunar
Planet. Sci. XXIV (abstract), 1259 (1993).
29. A. Montanari, C. Koeberl, Impact Stratigraphy: The
Italian Record (Springer-Verlag, Berlin, 2000), p. 364.
30. Defining criteria for tektites are as follows (29): (i)
glassy; (ii) homogeneous rock melts; (iii) contain le-
chatelierite; (iv) geographically extended strewn-
field; (v) distal ejecta, not around source crater or
within impact lithologies; (vi) poor in water, and
minor extraterrestrial component; (vii) formed from
the uppermost layer of the target.
31. We thank H. Jay Melosh for valuable discussion, M. J.
Abrams for assistance with ASTER data, D. Ducart and
D. Schonwandt for help in the field, and G. Graham
for assistance with glass analyses. P.A.B. thanks the
Royal Society for financial support, and C.R.d.S.F.
acknowledges CNPq-Brazil for research grant
301227/94-2.
Supporting Online Material
(www.sciencemag.org/cgi/content/full/296/5570/1109/
DC1)
figs. S1 and S2
26 November 2001; accepted 16 April 2002
Table 1. Major elements from energy-dispersive
x-ray microanalysis of Santa Rosa glass (average
of six analyses), compared with vesicle-abundant
Rı´o Cuarto impactite (1).
Vesicle-abundant glass
Santa Rosa
Rı´o Cuarto (1)
Clear Light brown
SiO
2
61.2 58.4 58.3
Al
2
O
3
23.6 25.1 15.8
MgO 1.6 0.3 1.7
Na
2
O 3.6 5.4 4.3
K
2
O 2.1 3.2 2.8
CaO 5.0 6.5 4.2
TiO
2
0.3 0.0 2.0
FeO 2.7 1.4 10.3
REPORTS
www.sciencemag.org SCIENCE VOL 296 10 MAY 2002 1111