Australites, Henbury impact glass and subgreywacke: a comparison of the abundances of 51 elements

Australites, Henbury impact glass and subgreywacke: a comparison of the abundances of 51 elements

4Jeochimica &Cosmochimica Acta,[email protected]&3.v0l. 3o,pp.1121tQ1133. Australites, Henbury impact glass and mbgreywacke: a comparison of the abundancesof 51 ele...

1MB Sizes 0 Downloads 11 Views

4Jeochimica &Cosmochimica Acta,[email protected]&3.v0l. 3o,pp.1121tQ1133.

Australites, Henbury impact glass and mbgreywacke: a comparison of the abundancesof 51 elements S. R. TAYLOR Department of Geophysics and Geochemistry, Institute of Advanced Studies Australian National University, Canberra,Australia (Received 21 March 1966; in revkedfom

9 May 1966)

Ah&&-Data are presented for 51 elements in two subgreywaokesand a meteorite impact glass from Henbury, Australia, and for five australitescoveringthe range in compositionfor this group of tektites. There are no sign&ant differencesin composition between the impact glass and the parental subgreywackes, and no effective alteration in concentrationhas taken place during melting for the elements studied. The australites are closely similar in composition to the Henbury glass and sedimentary rocks, except that Ca and Sr are enriched, and the chalcophile elements Cu, Pb, Sn, Tl, In and Bi are depleted in the tektites. From analogy with the impact glass-sediment relations, these variations are considered to be present in the tektite parent material, which otherwiseis equivalent in compositionto terrestrialsubgreywacke.

INTRODUCTION THE chemical composition of tektites can provide useful evidence about their place of origin, providing that the processes of formation of tektite glass have not seriously altered the composition of the parent material. Direct evidence of the effects of melting have been difficult to obtain, and contradictory results have been obtained from fusion experiments on rock types (FRIEDMAN et al., 1960 ; LOVERING, 1960). TAYLOR (1962) and TAYLOR and SACHS (1964) concluded from the overall composition and from the high concentration of elements such as cesium in tektites that selective volatilization had not modified the composition of the parent material to the extent that it could not be identified. They concluded that this composition was close to that of terrestrial argillaceous sandstones. Many workers (LOVERING, 1960; CHAO, 1963; SCRNETZLERand PINSON, 1963 ; O’KEEFE, 1963) have proposed that tektites are derived from material of granitic composition. This view is perhaps compatible with a lunar origin for tektites, since the presence of igneous rocks on the moon seems inherently more probable than the occurrence of sedimentary material. Selective volatilization is appealed to to explain the wide differences in composition between granitic rocks and tektites. Several uncertainties beset this approach : (1) Lunar rocks may differ substantially from their terrestrial analogues. (2) Extrapolation to possible parent materials is very difficult if selective volatilization has operated to the degree required to convert terrestrial-type granites to tektites. (3) The selective removal of elements will obscure major element relationships and will make comparison with terrestrial rocks by standard petrochemical procedures difficult. Recent work in the related field of the chemistry of impact glasses has yielded results of much relevance to studies of tektite parent material. The occurrence of 1121

1122

S. It.

TAYLOR

glass formed during meteorite impact at Henbury led TAYLORand KOLBE (1964a, 1965) to investigate the chemistry of the glass and its parent material for 22 elements. The evidence from Henbury is relevant for the following reasons: (1) Meteorite impact occurred at a transition from subgreywacko to quartzite. (2) The glass formed has a composition close to that of the subgreywaeke. (3) No appreciable change in composition for the 22 elements studied occurred during glass formation. (4) Volatile elements such as the alkalis have been retained in the glass. (5) The chalcophile elements have been retained in the glass. (6) The parental subgreywaeke, and the glass, are very similar in composition to australites. The production of the impact glass at Henbury is on a much smaller scale than the event which produced the australite group of tektites. Nevertheless it is possible that the conditions for glass formation are similar whether the events are large or small. No data exist for the age of the Henbury glass. KOLBE(in preparation) suggests an age of 3000 yr, based on an assessment of the geology and of the ages for the arrival of the iron meteorite, based on the decay of cosmogenic nuclides. This age appears to be much younger than the ages assigned to australites based on K-Ar dating of 0.5 to 1 x lo6 yr (Z~HRI~~E~, 1962; MCDOUOALL, personal communication). BAKER(1959) has given an age of 5000 yr for the time of arrival of australites on the earth. The geochemical study of the Henbury glass noted above was carried out on data for 22 elements. The adaptation of a spark source mass spectrograph to the analysis of geological samples (TAYLOR, 1965a, b) enabled the concentrations of another 30 elements to be determined. These include such critical elements as the rare earths, Tl, Bi and Hf. The principal purpose of this paper is to present the new analytical data and to study the relative compositions of the australites, Henbury impact glass, and parental sedimentary rocks. Nearly all the elements present as cations in silicates are included in the present study. SAMPLES Sample descriptions and localities are given in Table 1. The five tektites were selected to cover the variation in australite composition. Their silica concentrations are 69.7, 70.4, 73.1, 74.8 and 77.4, covering nearly the observed range of this constituent in australites. The two subgreywaekes from Henbury are the sediments closest in composition to the impact glass (TAYLORand KOLBE, 1965). Sample (a) is generally closer in composition to the analysed impact glass. A large number of additional glass and sedimentary rock samples from Henbury are being analysed. It should be noted that the meteoritic iron, cobalt and nickel contribution has been subtracted from the composition of the impact glass, as described by TAYLORand KOLBE(1965). ANALYTICALMEZ’HODS AXD DATA New data for the following elements are given: Tl, Pb, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, I’m, Y, Th, U, Hf, Sn, Nb, MO, W, Bi, In and Sb. These elements were determined

by spark source mass spectrography

using the method

1123

Austraiites, Henbury impact glass and subgreywacke

described by TAYLOR(1965a, bf. The data for the remaining elements have been published previously. The tektite data are from TAYLORand KOSSE (1964a, 1965). The major elements were determined by Dr. H. B. Wiik, using wet chemical methods, and the remaining elements by emission spectrography, using the methods described by AHRENS and TAYLOR(1961), TAYLOR(1960, 1962), TAYLORand KOLBE (1965), TAYLOR and SACHS (1964) and TAYLOR and SOLOMON(1962), and calibrated using the standard values reported by TAYLOR and KOLBE (1964b). Table 1. Sample and locality data (a)

~~tr~~t~

Museum

Sample no.

XLO.

25

11947D* 1004*9t 266t 11945*

28 32 35 39

s.

f3ung

37%7’ 30”49’ 24”35’ 25’55’ 30”47’

Boat core Core core

(b) Impad glass Sample no. HB-1

L&t.

Type

E. Long.

Locality

142”31’ 121”32’ 133”15’ 134”55’ 121”29’

Caramut, Vio. Hannans Lake, W.A. Henbury, N.T. Charlotte Waters, N.T. Boulder, W.A.

Locality Collected from area10 m diameter, North rim of main crater, Henbury, N.T. S. Lat. 24”35’, 133”09’

(c) Subgreywacke

Sample no.

Locality

A (5)

Block, north wall of main crater, Henbury, N.T. S. Lat. 24’35’, E. Long. 133”09’ East wall of main crater, Henbury, N.T.

J3 (3) * Western Australian Museum. t South Australian Museum.

The gallium data in the subgre~ackes and impact glasses were obtained using Ga 2943.637 as the analytical line. This normally usable line is subject to interberenee from Ni 2943.914 at the high concentrations due to meteoritic contamination observed in the impact glass. The lines are nearly resolved by the spectrograph used but the toe of the Ni line contributes about 30 per cent of the total intensity. The Ga data for HB-1 have been corrected for this. The other Ga lines at 2874 and 4172 A suffer from iron interference and Ga 4033 from Mn interference. All the samples have been analysed under closely similar conditions and the data may be compared directly, without the analytical uncertainties common to data from different laboratories. Detailed information on analytical precision is given in most of the papers quoted above, and summarised in TAYLORand KOLJIE(1964b) and TAYLOR and SACHS (1964) for the spectrographic data and in TAYLOR (1965b) for the mass spectrographic data. The precision (reproducibility) of the trace element

1124

8. R. TAYLOR

determinations expressed as relative deviation generally between five and ten per cent.

(per cent standard

deviation)

lies

PRESENTATION OF DATA The problem of presenting, in usable form, abundance data for over 50 chemical elements is difficult. A common procedure in tabulating geochemical data has been to arrange them in order of ionic radii since geochemists have recognised size as the prime factor controlling the distribution of the elements in mineral lattices. Alternative tabulations have been arranged by groups according to the periodic classification of the elements, by atomic number, by elemental abundance or in alphabetical order. All suffer from various defects. Principal objections are that elements of diverse character are brought together and that the tables become unwieldy when the number of elements exceeds fifteen or twenty. In this paper a tabulation of the chemical elements based on geochemical association in common rocks and minerals is used (see also TAYLOR, 1965b). Elements are arranged according to a combination of the important geochemical factors of size (ionic radius), valency (charge) and bond type (ionic-covalent). Within each division the arrangement is in order of decreasing ionic radius. Elements may occur in more than one table, reflecting their dual geochemical behaviour, but the amount of duplication is small. The advantages of the tabulation are: (1) Elements whose geochemical behaviour is similar and whose ratios a,re significant in geological processes are brought together. (2) The useful order of ionic radius is preserved within each group. (3) The system is flexible, and can be adapted readily for additional groups (e.g. the platinum elements). (4) The individual tables are of convenient size. The principal objective is to make a workable system to handle abundance data. It is not desirable to introduce a new terminology, but it is useful to have some identification for each division, based on commonly used terms. (a) The large cations (potassium type) These comprise the elements which occur in 8-12 fold co-ordination with oxygen in silicates, and form dominantly ionic bonds. Potassium is the dominant major cation. Elements such as lead and thallium, which form strongly covalent bonds, are included because of their apparent ability to enter potassium lattice sites. The common elements sodium and calcium, are conveniently included here. (b) The rare earth elements It is convenient to list these adjacent to table (a) since their ionic radii overlap with that of the common cation, calcium. (c) The large highly charged cations (zirconium type) These comprise the large cations with valencies of 4,5 and 6. These elements have di&ulty in entering common rock-forming minerals because of valency difficulties. They also occur as complex ions.

1126

Australites, Henbury impect glass and subgreywacke Table 2. Analytical data

The All date are given in p&e per million (wt) except where indicated ea wt. per ceneno date ev&ble. meteoritic iron. cob& end nickel contribution hes been eubtraoted from the impnot glaes. (See TAYLOB and KOLBJE, 1965.) The sedimentary rock date are expressed on e water free beeis. Ionic radii are from AEXEXO (1952). Analyete: Major elements: H. B. Wiik. Other elements: 8. R. Taylor, M. Saabs, A. Capp and A. Haleey

Anelyeis no. Museumno.

Radius

Large~0th

(a)

26 18

28 11947-D

Tektites 32 1004.9

35 266

38 1194s

Subgreywaokes B A 3 5

tipect Bh 1 HB-1

(K type)

03’

1.67

Rb+

1.47

4.0

3.7

103

96

2.8 78

2.6 92

2.0 79

Tl+

1.47






B3+

1.35

610

590

600

580

600

2.1

3.0 110

93 0.6

0.6 620

430

4.2 127 0.4 640

%K+

1.33

2.13

2.04

1.80

1.84

1.79

2.44

2.30

Eu’f

1.25

I.3

1.3

1.1

1.4

1.2

1.3

1.9

Pb’+

I.20

4.0

1.4

4.0

7.8

5.3

Sf+

1.18

c/&s’+

1.02

2.24

2.13

3.34

2.59

1.94

0.60

0.34

0.53

%Na+ K/Rb

0.97 -

1.11

1.10

0.87

0.85

0.87

0.74

0.61

0.69

195

150

235

185

180

2.65 1.2

11

15

15

90

69

85

207

213

230

200

227

222

247

209

K/Cs

-

5330

5510

6430

7080

8950

8130

10950

6310

Rb/Cs

-

26

26

28

35

40

37

Be/Rb

-

5.9

6.1

7.7

6.3

7.6

5.6

4.6

5.0

Be/&

-

3.1

3.9

2*6

3.1

3.3

6.9

6.2

7.5

Rb/Sr

-

0.53

0.64

0.33

0.50

0.44

0.82

1.34

1.49

44

30

(b) Rare earth LB*+

1.14

54

60

52

52

37

95

36

56

[email protected]+

1.07

88

84

88

78

60

140

SO

76

Pl+

1.06

10.2

11.5

Nda+

1.04

29

33

8.5 22

8.0 24

9.0 34

12.0 32

8.0 27

8.8 29

Sma+

1.00

5.0

4.6

4.1

5.6

4.4

5.3

5.9

Eu*+

0.98

1.3

1.3

1.1

1.4

1.2

1.3

1.9

1.2

Gd’+

0.97

5.2

5.5

3.9

4.8

4.2

5.2

6.0

5.1

Tb’+

0.93

0.8

1.0

0.8

0.9

0.8

1.1

1.5

0.9

Dy”+

0.92

4.5

4.3

3.8

4.4

4.2

3.9

4.9

4.6

Ho*+

0.91

1.0

1.2

0.9

1.2

1.0

1.3

1.7

1.1

Es+

0.89

1.9

2.2

1.9

3.0

1.9

2.8

3.9

2.1 0.4

T&+

0.87

0.3

0.4

0.3

0.4

0.35

0.5

0.8

Ybw

0.86

1.8

1.9

1.6

2.2

1.9

2.6

3.3

Ys+

0.92

(c)

34

Large,highly-charged,c&fm.3 (Zr

25

37

28

25

30

36

4.6

2.0 36

type)

Th’+

1.02

14.5

13.4

9.0

11.5

12.1

14.8

9.5

U’f

0.97

3.1

1.9

2-o

2.8

2.1

3.8

4.1

C#+

0.94

88

84

88

78

60

140

260

204

360

250

295

390

SO -

9.6 2.4 76

zr4+

0.79

HP+

0.78

3.7

f3l#+

0.71

0.73

Nb6+

0.69

15

13

19

13

16

13

11

15

Ti’+

0.68

5400

4700

3700

3900

3600

4930

4330

5280

MO’+

0.62

0.4

-

0.4

1.0

0.6

0.9

0.8

I.4

we+

0.62

0.6

-

0.6

0.8

0.6

1.8

0.7

1.3

Th/U

-

Zr/Hf

-

4.7 70

4.0 -

7.1 51

400

3.7

4.7

5.0

8.5

6.6

3.7

0.67

1.7

1.1

4.0

3.9

3.6

4.5 97

4.1 53

5.8 59

3.9 46

2.3 -

4.0 108

1126

8. R. TAYLOR

Table 2 (cont.)

Anelysis no. Museum no. Radius (d) Fervmcqwaian %Fe total

25 18

28 11947-D

Tektites 32 1004.9

35 206

38 11945

Subgreywackes A R 5 3

Impwt glass 1 HB-1 ---

elements

-

4.03 760

3.99 680

3.28

3.33 870

760

3.24 410

Mn’+

0.80

%Fe*+

0.75

3.36

3.41

2.80

2.80

2.74

cu=+

0.72

5.6

56

6.3

5.8

57

2.88 840

3.33 290

0.80

l-20

18

34

(2.88) 1110 39

cw+

0.72

15

16

15

15

15

11

--

(11)

IG+

0.69

32

36

31

27

28

26

Li+

o&?

44

46

37

35

33

35

28 45

(26) 40

%Mgs+

0.66

X/Co

-

I*18

1.44

1.19

1.10

0.97

1.22

2.1

2.2

2.1

1.8

1.9

2.4

0.81

12

14

11

10

I2

Va+

0.74

93

81

71

65

62

65

62

63

Ti*+

0.68

5400

4700

3700

3900

3600

4930

Fes+

0.64

Crs’

0.63

89

GG+

0.57

10

%AIS+

0.51

G/V

-

il’@s

-

0.58 76

8.2

0.48

0.53 65

61

8.8

1.44 -

W’

0.67

8.2

1.29 -

0.50 63

2.08 56

4330 2.13 58

5280 83

6.0

8.6

8.9

8.3

9.0

9.4

8.5

7.03

7.35

5.75

5.86

5.21

5.74

5.81

6.51

0.96

0.94

0.92

0.94

1.0

0.86

0.94

7030

12250

6690

6580

6280

6380

6180

1.3 7650

(e) S9nuiEe&en-s (Si #we) %A1=

0.51

%W+

0.42

PA+

0.35

7‘03 32.5

7.35 32.9

5.86

5.75 34.1

35.0

5.21 35.9

520

170

310

ad.

220






5.74 36.1 100

5.81 36.2 100

&51 36.0 150

(f) ChukophGe elements Tl+

1.47

Pb=+

1.20

4.0

BiS+

O-96

Ins+

0.81

SbS+

0.76

0.19

CUB+

0.72

5.6

snp+

0.71

0.73

n.d.
1.4 n.d.
7.8

4.0 n.d.
5.3

0.6 ii

0.6 15

o-4 15

n.d.

n.d.

0.59

0.65

0.37

<0*1

-co*1

0.15

0.10

0.11

0.64

0.28

0.2

0.38

0.4

6.3

5.8

&7

0.87

1.7

1.1

18

34

4-o

3.9

@22 39 3.6

(g) iwajor elements o/oSiO,

-

89.67

70.39

73.08

74.82

77.39

77.26

77.53

75.09

O-‘W,

-

13.29

13.89

10.87

11.07

IO*10

10.85

1iaJ

12.32

%FG’,

-

0.96

0.83

O-76

0.69

0.72

2.97

3.05

-

%FeO

-

4.32

4.39

3.60

3.60

3.52

1.03

1.54

(371)

%MgO

-

1.96

2.38

1.98

1.82

1.63

2.01

2.13

2.39

%CaO

-

3.14

2.98

4.67

3.63

2.39

0.83

0.47

0.74 o-92

%Na,O

-

1.50

1.48

1.17

1.15

1.15

1.00

0.82

%W’

-

2.57

2.46

2.17

2.22

2.17

2.95

2.76

3.19

%TiO,

-

0.90

0.78

0.62

0.65

0.60

0.82

0.72

0.88

%P,O,

-

-

0.05

0.12

0.04

0.07

98.42

99.62

99.00

SD.66

99.72

5.18

5.14

4.28

4.22

4.17

0.02 9974

0.02

0.03

100.04

99.27

4.28

(3.71)

o/oTotalFe as Fe0

3-71

Australites,Henburyimpactglassand subgreywacke (d) The ferromagnesian

1127

elements

Here are listed the group of trivalent and divalent elements which occur in 6-fold co-ordination with oxygen in silicates. It is convenient to list separately the trivalent and divalent elements, and also to include titanium and lithium. Copper and zinc, although chalcophile elements, are also listed. (e) The small cations (silicon type)

Here belong the small cations occurring in 3- and 4-fold co-ordination with oxygen in silicates. (f) The chalcophile elements These form dominantly covalent bonds and are the typical elements entering sulphide phases. The ionic radii in Angstrom units are from AE~RENS (1952), and AHRENS (personal communication). The method of graphical presentation of the data is to plot, on a logarithmic scale, the tektite abundances against the sediment and impact glass abundances. A 45” line is drawn so that elements which have the same concentration in both tektites and sediments will fall on this line. Increasing distance from this line represents increasing disparity in composition. The error expressed as relative deviation of any point due to analytical uncertainty is about f 10 per cent. GEOCHEMICAL COMPARISONS The overall composition of tbe subgreywacke and impact glass is compared with that of the austrahtes in Fig. 1. The concentration range extends over a factor of lo*, from 0.01 ppm to 100 per cent. The average data for both groups are plotted. This graph gives an overall picture of the element relationships, which are shown in detail in Figs. 3-6. The most striking feature is the close agreement for most elements between the australites and the sedimentary rocks. The average concentrations agree within less than 10 per cent for most elements. This close to the analytical uncertainty. Eight elements, indicated on Fig. 1, (Ca, Sr and the chalcophile elements Cu, Pb, Sn, Tl, In and Bi) show significant differences in composition. Comparison with granitic average composition

Although it is not a primary purpose of this paper to compare the composition of other rock types with that of australites, it is instructive to present the abundance data for granitii: rocks on a similar diagram. These are the other commonly suggested source material. Figure 2 shows the comparison between the australite averages and average granitic rocks, derived by averaging the three estimates of their composition from TUREKIANand WEDEPOHL(1961) for high Ca-granites, VINO~RADOV (1964) and TAYLOR(1964). Si, Al, Mn, Li, Y and Sb are nearly equivalent in composition. The other elements show wide differences, equivalent to those between granite and subgreywacke. The enrichment in the ferromagnesian group and depletion in alkalis and chalcophile elements are to be noted. There is no ready correlation with volatility, for Th, U and Y are not enriched in the tektites. It is accordingly difficult to derive tektite composition from granitic rocks by selective volatilization, as discussed 2

6. R. TAYLOR

1128 100

10 %

Ca1.0

I,

SrP P co

100

10 -CU

wm

-Pb -Sn

-TI

-01

-01

0.1

1.0 Ppm

10

100 Sediment

1000

1.0

ya

lo

Fig. 1. Comparison of the averageelement abundances in au&al&s compared with the averqea for Henbury subgreywackeand impact glass. Note the coincidence in composition over wide concentrationranges, except for Ca, Sr and the chalcophileelements.

by TAYLORand SAC- (1964). Losses of the order of 10-20 per cent are not precluded by the present data, but losses of this magnitude do not affect the conclusions of this paper. It should be noted that averages are used, but that any bias is toward granites with SiO, values of about 76-72 per cent. Use of data from granites with higher silica values would accentuate the difference8 observed. DETAILEDCOMPARISON OF TEKTITEAND SEDIMENTARY COMPOSITIONS The detailed comparisona between these types are prmnted in Egs. 3-6. The spread in composition is reprmented by vertical lines for the tektites, and horizontai lines for the aubgreywackes and impact glasses. The open circles represent the values for the impact glasses.

1129

Australites, Henbury impact glass and subgreywacke 100

10

Fe Ca 41-

% 1

-No

ti1000 -Mn

2 z * f

100

-Rb

-Y

wm

-Th

10

-cu -

Pb

rC; -Sn

14

Sb0

/.

:

J

1

O-1

1-o

, 10

Pm

rgi, _-TIIn

100

,

1000

,

1

10

(

loo

%

Granite

Fig. 2. Comparisonof the overagecompositionof australiteswith element averages for granitic rocks (see text). The scale is the same as in Fig. 1. Note the wide dispersionin concentration.

Nfljor

elements

Figure 3 shows the relative concentrations

of Si, Al, Fe, Ca, Mg, K, Na and Ti.

Most of these elements fall close to the line of equal composition.

Potassium

is

slightly lower, and sodium slightly higher in the tektites, but the only real discrepancy is in the concentration of calcium, which is higher in the australites. It is a variable constituent in tektites generally (TAYLOR, 1962). The Ca content of bediasites, which averages O-65 per cent CaO, is comparable with that of the subgreywackes. The variations in calcium are accordingly attributed Ca as a major cation, in the parent material.

to variations in a phase containing

1130

S. K. TAYLOR

50

10

5 Q .% 5

Ca

co

r y 5 P

1

-5

/

I

-5

I

1 per cent

L

I

I

5

10

50

Sediment

Pig. 3. Comparison of the abundanoes of the major elements in australites (vertical lines), subgreywackea (horizontal lines) and impact glass (0). Note the enrichment of calcium in the tektites and the close correspondence for the other elements.

Trace elements

It is convenient to show the relationships for these elements on two separate graphs covering different concentration ranges. Figure 4 shows those elements in the range 6-1000 ppm and Fig. 5 in the range 0.01-10 ppm. In Fig. 4, in order of decreasing concentration, Ba, Zr, Rb, C’e, V, Cr, La, Li, Ni, Nd, Y, Nb, Co, Th, Pr and SC show a close correspondence in concentration. The average value for Mn is close, but this element shows wide variations in concentration which makes it an unsatisfactory index element. This is probably due to the varying oxidation states and wide mobility of Mn, particularly in sedimentary processes. The anomalous element is Sr, which resembles Ca in being enriched in the australites. Bediasites contain from 60-140 ppm Sr (CHAO 1963), illustrating the sympathetic variation with Ca content. The similarity

in gallium

abundances, low in both tektites and sedimentary

Austrctlites,Henbury impact glsss and subgreywaoke

1131

St

+ V

/

LO

10

wm

50

100

500

1000

Sadimen,

Fig. 4. ~orn~&risonof the a;bu.ndancesof the trace elements in the range 61000 ppm in austr&ites (v&&al lines), sub~ywac&es (horizontal Iines) and impwt glass (0). Note the enrichment of strontium and the depletion of copper in the tel-tites and the close correspondencein composition for the remaining elements.

material, is noteworthy. The overall correspondence in ~orn~tio~ shown in Fig. 4 is striking. The depletion in Cu in tektites shown in Fig. 4 is discussed in the next section. Figure 5 shows the rarer trace elements. In order of decreasing concentration, Sm, Gd, Hf, Dy, Cs, Er, Yb, Eu, Ho, Yb, Sb and Tm show a close sim~arity in tektites and the Henbury sedimentary rocks and impact glasses, Uranium, but not thorium, is lower in the tektites, as is shown by the ThjU ratios in Table 2, (c). These range from &l-7*1 for the tektites compared with subgreywacke and impact glass ratios of about 4. TILTON(1958) ascribed the high Th/U ratios to ~iobi~ is present at the same selective removal of U, possibly during w~the~g. level in both. The Hf content appears slightly lower in tektites, leading to higher ZrjHf ratios. Both MO and W are slightly lower in the tektites.

S.

1132

R . TAYLOR

Figure 5 shows ths striking depletion compared with the sedimentary material. cu 3.5 Pb 2.5 Sn 4

of Cu, Pb, Sn, Tl, In and Bi in tektites The depletion factors a,re: Tl > 6 In > 20 Bi > 20

These elements are mainly “chalcophile,” entering sulphide phases in common geological materials. Even their reported presence in many silicate minerals may be due to minute sulphide inclusions and their geochemical distribution is imperfectly understood for this reason (TAYLOR,1965). The rare earths

HASEINand GEHL (1963) have publish~ values for the rare earth elements in an australite, and the present data on five fully analysed samples are in general agreement with their data. The abundance data are plotted in Fig. 6, together with the values for the Henbnry impact glass. All data are normalised to Yb = 1.00, and divided by the Yb-normalised chondritic abundances from HA~XIN et al. (1965). In Fig. 6, the horizontal line represents the chondritic abundances. The enrichment of the impact glass and the australites in the lighter rare earths is clear. The identity of the two rare-earth patterns is clear and striking. In addition to the relative patterns shown in Fig. 6, the absolute abundances in Table Z(b) and shown in Figs. 4 and 5 are very close. COMPOSITIONAL DIFFERENCES The Ca-Sr anomaly Among the major elements, Ca alone does not display a significant negative correlation with SiO, (TAYLORand SACHS,1964). Although it is occasionally stated that the inverse relation between SiO, and other constituents is only of statistical si~~nee, the lack of secant correlations with certain elements (Ca, Sr, Zr, M.n, Cu etc.) indicates a genetic, as well as a mathematical control. The mobility of minerals such as calcite during weathering and sedimentation are responsible for the wide variations in the calcium content of sedimentary rocks. The lack of correlation between Ca and SiO, in australites, unique among the major elements, indicates that a major portion of Ca in the parent material of tektites may be in a separate non-silicate phase. This conclusion is reinforced by the similar behaviour of strontium, alone among the trace elements in this behaviour. According to preference, the geoche~st may choose from the common simple calcium minerals such as fluorite, calcite, anhydrite, gypsum or pexhaps oldhamite. If the analogy with sedimentary material is pursued, then c&it-e, is a logical choice. Ok&am&e seems excluded because of the low’ sulphur content and the lack of other chalcophile elements. Depletion

of chaleophib

eZement8

One of the principal observations of this paper is that the chalcophile elements, as a group, are depleted in a~trali~s. TALON (1953) observed that lead was low in tektites, GREBNLBND and Lovxnr~u f X963) noted the low concentration of !Zn

Allstra~~, Henbury impactglass md subgceywmke

1133

Sediment

Fig. .5. Comparison of the abundances of the rarer elements in au&r&&s (vertical lines), subgreywackes (horizontal lines) and impact glass (0). Note the depletion of the chalcophile elements (Cu, Pb, Sn, Tl, In, and Bi) in tektites ad the close correspondence in composition for the remaining elements.

tektites compared with shales and granites, and TAYLOR (1962) and TAYLOR and noted the low abundance of copper in australites. The deficiency of the chalcophile elements may be a primary feature of the parent material at the time of melting, or may be due to loss during melting. The evidence in this paper favours the first alternative. The Henbury glass has retained these elements during melting. The alkali elements were also retained by the Henbury glass, and also apparently by the austrahtes. There ia certainly no need to invoke any loss to account for the in

SACHS (1964)

0.

14

5

10

20

I CePr

I

I Nd

I

Sm

I I

Eu Cd

I lb

I

I

Dy Ho

I Er

I I

Tm Yb

Fig. 6. The abundances of the rare earth elements in australites and Henbury impact glass. The data are normalised to Yb = 1.00 and divided by the Yb-normalised chondritic abundances (HASKIN et al., 1965), which are aocordingly represented by the line at 1.00. Note the coincidence of the two abundance patterns. The absolute abundances also show a close correspondence.

La

t

Au&al&es, Henbury impact glass and subgreywacke

1136

observed element relationships. It seems safer to appeal to these observations than to calculations based on considerable extrapolation and uncertainty regarding temperatures during formation of tektite glass. Sulphides are rapidly oxidised and removed from rocks exposed to weathering at the surface of the earth. In arid regions, such as Henbury, oxidation will occur, but the sulphates are not readily removed (GOLDSCHMI~T, 1954). It is thus possible to account for the differences in the abundances of these elements by conventional geological processes operating at the surface of the earth. Alternatively, the depletion in these elements may be a primary feature of the parent material of tektites. It may be noted that the ordinary chondrites are likewise deficient in the chalcophile elements compared with the carbonaceous chondrites, and assumed primitive solar system abundances. SUMMARY Conclusions based on the abundances of one or two elements must be treated with caution. Analytical error, non-homogeneous distribution and sample contamination When numbers of diverse elements are may all affect individual comparisons. considered, these possibilities become remote. Similarities based on the abundances of 50 elements must be considered well established. The data given in Table 2 and plotted in Figs. 1 and 3-6 show the following features : (1) The Henbury subgreywacke and the impact glass compositions are similar for all elements, and no effective change in composition has been effected by the melting process. (2) The composition of the australites is very close to that of the terrestrial sediment and impact glass. The average compositions agree within about 10 per cent over a concentration range of 10 million for all elements except Ca, Sr, and the chalcophile elements Cu, Pb, Sn, Tl, In and Bi. This is a remarkable coincidence in composition. (3) The evidence for the impact-glass-parental sediment relationships suggests that the differences in composition between tektites and subgreywacke are primary, and not caused by the melting process. These differences are reasonably explained as due to variation in the abundance of a calcium mineral and to the low concentration of sulphides in the parent material of tektites. (4) The virtual identity in composition of australites and terrestrial sedimentary rocks of subgreywacke type is established. Acknowledgements-Theauthor is grateful to Miss A. C. CAPP who carriedout the mass spectrographic determinations,and to Dr. P. KOLBE for assistancein collecting samples.

REFERENCES L. H. (1952) The use of ionization potentials, Part I. Ionic radii of the elements. Geochim. Cosmochim. Actu 2, 155-169. AERENS L. H. and TAYLOR S. R. (1961) Spectrochemical Analysis (2nd Edition) Addison-Wesley. BAKERG. (1959) Tektites. Mem. Nat. Mwr. Victoria 2% CHAO E. C. T. (1963) The petrographic and chemical characteristicsof tektites. In Tektites, (Ed. J. O’KEEFE) Chapter 3, pp. 51-94. University of Chicago Press. AHRENS

S. R. TAYLOR

1136

FRIEDMANI., THORPEA. and SENFTLEF. E. (1960) Comparisonof the chemicalcompositionand magnetic propertiesof tektites, and glasses formed by fusion of terrestrialrocks. Nature 187, 1089-1092. GOLDSCHMIDT V. M. (1954) Geochemistry(Ed. A. MUIR). Oxford University Press, 730 pp. GREENLANDL. and LOVERIN~J. F. (1963) The evolution of tektites: elemental volatilization in tektites. Geochim. Coemochim. Acta 2’7, 249-259. HASKIN L. A., FREY F. A., SC~MIYL-IY R. A. and SMITE R. H. (1965) Meteoritic, solar and terrestrialrare-earth distributions. General Atomic Report GA-6800. -KIN L. and GEHL MARY A. (1963) The rare-earth distribution in tektites. Sciem~ [email protected], 1056-1058. KOLBEP. (1966) Australian Meteorite Craters (in preparation). LOVERING J. F. (1960) High temperaturefusion of possibleparent materials for tektites. Nature 186, 1028-1030. O’KEEFE J. A. (1963) The origin of tektites. In Tektites (Ed. J. O’KEEFE), Chap. 8, pp. 167-188. University of Chicago Press. SCHNETZLER C. C. and PINSONW. H. (1963) The chemical composition of tektites. In Tektites (Ed. J. O’KEEFE) Chap. 4, pp. 95-129. University of Chicago Press. TAYLORS. R. (1960) Abundance and distribution of alkali elements in australites. Geochim. Cosmochim.

Acta 20, 85-100.

TAYLORS. R. (1962) The chemical composition of australites. Geochim. Coamochim. Acta %_I, 685722. TAYLORS. R. (1964) Abundance of elements in the continental crust: a new table. Geochim. Coamochim. Acta 28, 1273-1285. TAYLORS. R. (1965a) Geochemicsl applicationof spark sourcemass spectrography. Natwe ao5, 34-36.

TAYLOR S. R. (1966b) Geochemical analysis by spark source mass spectrography. Geuchim. Coemochim. Acta 29, 1243-1262. TAYLORS. R. (196%) Similarity in composition between Henbury impact glass and australites. Geochim. Gosmochim. Acta 29, 599-601. TAYLORS. R. (1965d) The application of trace element data to problems in petrology. Phys. Chem. Earth. VI, Chap. 2, 133-213. TAYLORS. R. and KOLBEP. (1964e) Henbuty impact glass: Parent material and behaviour of volatile elements during melting. Nature %08, 390-391. TAYLORS. R. and KOLBE P. (1964b) Geochemical standards. Geochim.Cownochim. Acta 28, 447-454.

TAYLOR S. R. and KOLBE P. (1965) The geochemistry of Henbury impact glass. Geochim. Cowrwchim.

Acta 29, 741-754.

TAYLORS. R. and SACHSM. (1964) Geochemicalevidence for the origin of australites. Geochim. Coswwchim.

Acta 28, 235-264.

TAYLOR S. R. and SOLOMON M. (1964) The geochemistryof Darwin Glass. Geochim.Cosnnochim. Acta 28, 471494. TILTON G. R. (1958) Isotopic composition of lead from tektites. Geochim. Cosmochim. Acta 14, 323-330.

TUREKIAN K. K. and WEDEPORLK. H. (1961) Distribution of the elements in some major units of the Earth’s crust. Bull. Geol. Sot. Am. 72, 175-192. VINOGEADOV A. P. (1962) Average contents of chemical elements in the principal types of igneous rocks of the Earth’s crust. Geokhimiga 1962 (7) 641-664. Z;/HRINGER J. (1963) Isotopes in tektites. In Tektites (Ed. J. O’KEEFE) Chap. 5, pp. 130-149. University of Chicago Press.