Partition coefficients of uranium for some rock-forming minerals

Page 1

Chemical Geology, 15 (1975) 285--294 © Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands

PARTITION COEFFICIENTS OF URANIUM FOR SOME ROCK-FORMING MINERALS

J. DOSTAL and S. CAPEDRI*

Department of Geology, Dalhousie University, Halifax, N.S. (Canada)

(Received January 31, 1975; accepted April 7, 1975)

ABSTRACT

Dostal, J. and Capredi, S., 1975. Partition coefficients of uranium for some rock-forming minerals. Chem. Geol., 15: 285--294.

Partition coefficients of uranium between phenocrysts and their host groundmass have been determined by fission-track mapping. The minerals analyzed include plagioclase, K- feldspar, biotite, olivine, clinopyroxene and orthopyroxene. The data for all these minerals show that U is strongly partitioned into the liquid and only a small fraction of the total whole-rock U content is present in the major rock-forming minerals. In volcanic rocks, the bulk of U is usually contained in glass although in acid volcanic rocks a significant portion may also be present in the U-rich accessory minerals.

INTRODUCTION

A knowledge of the partitioning of trace elements between rock-forming minerals and silicate melts has played an important role in elucidating the genesis and evolution of igneous rocks. Although the partition coefficients for many lithophile elements are fairly well established, only limited empiri- cal data are available for uranium, the most important heat-producing element. In fact, there is relatively little information on U abundances in some common rock-forming minerals. This has led us to determine U in several minerals, main- ly from volcanic rocks.

The purpose of this paper is to report the data on the distribution of U in several igneous rocks and their const i tuent minerals, particularly on the U- partitioning between rock-forming mineral phenocrysts and the groundmass.

SAMPLE DESCRIPTIONS

The samples analyzed in this s tudy are: (1) Acid volcanic rocks from San Vincenzo, Roccastrada, Monte Amiata

and Monte Cimino, Tuscany, Italy and genetically associated plutonic rocks of Monte Capanne, Elba.

*On leave from the Insti tuto di Mineralogia e Petrologia, Universit~t di Modena, Italy.

Page 2

TA

BL

E I

Ch

emic

al c

om

po

siti

on

of

min

eral

s an

d g

lass

es o

f al

kal

i bas

alts

an

d s

tru

ctu

ral

form

ula

e o

f m

iner

als a

t,0

00

W

t. %

O

liv

ine

Cli

no

py

rox

ene

sam

ple

70

b

sam

ple

73

b

sam

ple

70

b

sam

ple

73

c

core

ri

m

core

ri

m

core

ri

m

Pla

gio

clas

e

sam

ple

73

b

core

ri

m

Gla

ss

sam

ple

70

sa

mp

le 7

3

SiO

=

39

.41

3

8.8

9

40

.45

3

9.4

7

52

.11

4

6.8

4

46

.80

4

9.2

2

50

.63

T

iO=

0

.04

0

.04

0

.06

0

.06

0

.26

3

.08

2.

57

0.0

4

0.0

4

AI=

O~

0.0

5

0.0

6

0.0

5

0.07

1

.27

6

.83

5

.59

3

1.6

9

31

.09

C

r20~

0

.03

0

.04

0

.00

0.

01

0.0

0

0.1

8

0.0

7

0.0

1

0.0

1

FeO

d

16

.46

1

8.6

8

13

.93

1

8.4

5

11

.14

6

.75

7

.67

0

.28

0

.26

M

nO

0.2

1

0.31

0

.20

0

.33

0

.40

0

.10

0

.10

0

.01

0

.01

N

iO

0.1

7

0.1

7

n.d.

n.

d.

0.9

0

0.0

1

n.d.

n.

d.

n.d.

M

gO

43

.27

4

1.3

1

44

.90

4

1.1

3

12

.77

1

2.7

6

13

.52

0

.06

0

.05

C

aO

0.2

3

0.3

8

0.0

9

0.2

5

21

.38

2

2.8

1

22

.43

1

5.5

8

14

.50

N

a20

0

.04

0

.03

0

.02

0

.03

0

.41

0

.53

0

.60

2

.62

3

.17

K

20

0

.02

0

.06

0

.02

0

.03

0

.03

0

.03

0

.08

0

.26

0

.33

Su

m

99

.93

9

9.9

7

99

.72

9

9.8

3

99

.77

9

9.9

2

99

.43

9

9.7

7

10

0.0

9

Si

0.9

99

0

.99

7

1.0

12

1

.00

9

1.9

68

1

.75

3

1.7

69

9

.03

8

9.2

36

IV

0

.00

1

0.0

02

0

.00

1

0.0

02

0

.03

2

0.2

47

0

.23

1

6.8

58

6

.68

4

~Al V

I 0

.02

5

0.0

54

0

.01

8

Ti

0.0

01

0

.00

1

0.0

01

0

.00

1

0.0

07

0

.08

7

0.0

73

0

.00

5

0.0

05

C

r 0

.00

1

0.0

01

0

.00

5

0.0

02

0

.00

1

0.0

01

M

g 1

.63

4

1.5

78

1

.67

4

1.5

67

0

.71

9

0.7

12

0

.76

2

0.0

16

0

.01

4

Ni

0.0

04

0

.00

4

Fe

0.3

49

0

.40

0

0.2

91

0

.39

4

0.3

52

0

.21

1

0.2

42

0

.04

3

0.0

40

M

n 0

.00

5

0.0

07

0

.00

4

0.0

07

0

.01

3

0.0

03

0

.00

3

0.0

02

0

.00

1

Ca

0.0

06

0

.01

0

0.0

02

0

.00

7

0.8

65

0

.91

5

0.9

08

3

.06

5

2.8

34

N

a 0

.00

2

0.0

02

0

.00

1

0.0

01

0

.03

0

0.0

39

0

.04

4

0.9

33

1

.12

1

K

0.0

01

0

.00

2

0.0

01

0

.00

1

0.0

01

0

.00

1

0.0

04

0

.06

1

0.0

77

X

0.9

0

0.9

6

0.9

6

4.1

3

4.0

9

Y

2.0

0

2.01

1.

97

1.9

8

1.1

2

1.0

7

1.1

0

Z

1.0

0

1.0

0

1.01

1.

01

2.0

0

2.0

0

2.0

0

15

.90

1

5.9

2

nd

. =

no

t d

eter

min

ed

56

.03

5

4.0

7

1.5

9

1.89

2

1.6

5

20

.15

0

.00

0

.00

4

.33

4

.66

0

.07

0

.07

0

.02

n.

d.

1.0

9

1.3

0

1.9

3

2.2

7

e 7.

11

e 7

.46

86

.71

9

8.9

8

a C

alcu

lati

on

on

th

e b

asis

of

4, 6

an

d 3

2 o

xy

gen

s p

er f

orm

ula

fo

r o

liv

ine,

cli

no

py

rox

ene

and

pla

gio

clas

e, r

esp

ecti

vel

y.

b P

hen

ocr

yst

. c

Gro

un

dm

ass.

d

To

tal

iro

n a

s F

eO.

e N

a=O

an

d K

20

co

uld

no

t be

mea

sure

d d

ue

to r

apid

dec

reas

e o

f X

-ray

in

ten

sity

.

Page 3

287

(2) Alkali olivine basalts from Pozzomaggiore, Sardinia. The acid volcanic rocks from Tuscany are ignimbrites of late Tertiary to Quaternary age. Their composition varies from one volcanic centre to another, quartz-latites, rhyo- lites and trachytes being the dominant rock types, whereas the plutonic rocks are mainly granodiorites. The volcanic rocks contain phenocrysts of sanidine, biotite and plagioclase in a glassy groundmass. At San Vincenzo and Rocca- strada, ignimbrites also occasionally contain crystals of quartz and cordierite while hypersthene occurs in ignimbrites of Monte Amiata and Monte Cimino. Detailed geological, petrographic and chemical information on these rocks was given in a series of articles (Marinelli, 1961; Mazzuoli and Pratesi, 1963; Mittempergher and Tedesco, 1963; Barberi et al., 1967; Mazzuoli, 1967; Dupuy, 1970; Dupuy and Allegre, 1972}. In particular, the chemical composition of analyzed minerals, glasses and rocks was reported by Dupuy (1970).

The alkali basalts from Pozzomaggiore, NW Sardinia, are of Pliocene- Quaternary age (Coulon et al., 1974). These rocks consist of phenocrysts of olivine, clinopyroxene and plagioclase in a groundmass of glass, plagioclase, clinopyroxene, olivine and opaque minerals. The phenocrysts are zoned and generally euhedral in shape. The chemical composition of minerals and glasses analyzed for U is given in Table I together with their structural formulae.

ANALYTICAL METHODS

The uranium determinations were made by the fission-track method as adapted at Dalhousie University (Aumento and Hyndman, 1971; Mitchell and Aumento, 1974). Thin pellets of the pulverized samples were used for the whole-rock determination, while the analyses of the mineral phases and glasses were done on the polished thin-sections. The precision and accuracy of these determinations can be evaluated from the data on the standard rocks BCR-1 and DTS-1 which were analyzed simultaneously with the samples studied. The values obtained are 1.79+0.12 ppm U (BCR-1) and 0.0037+-0.0005 ppm U (DTS-1). The values given by Flanagan (1973) are 1.74 and 0.004 ppm U, respectively. Each value for the reference rocks is an average of three deter- minations, while the error indicated is the standard deviation, which is less than 15%. The precision of some values for mineral phases is, however, poorer, probably up to 20%. The chemical compositions of mineral phases and glasses given in Table I were obtained by electron-microprobe at Dalhousie University.

RELIABILITY OF THE DATA

In addition to a number of problems and assumptions involved in the deter- mination of partition coefficients for trace elements (e.g., Schnetzler and Philpotts, 1968, 1970; Philpotts and Schnetzler, 1970; Albarede and Bottinga, 1972) there are two important factors when considering the reliability of our data. The first is the mode of occurrence of U in the analyzed minerals and the second is whether the assemblages studied at least approach quasi-equi-

Page 4

TA

BL

E I

I

U c

on

cen

trat

ion

s (p

pm

) in

roc

ks,

mat

rice

s an

d m

iner

als

tO

0V

0V

Ro

ck

Mat

rix

B

ioti

te

K-f

eld

spar

P

lagi

ocla

se

Cli

no-

Ort

ho

- O

livi

ne

Co

rdie

rite

p

yro

xen

e p

yro

xen

e

Gra

no

dio

rite

M

. C

apan

ne

4.6

0

Ign

imb

rite

S

an V

icen

zo

7.69

10

.1

Ign

imb

rite

R

occ

astr

ada

8.22

12

.2

Ign

imb

rite

M

. A

mia

ta

7.97

11

.6

Ign

imb

rite

M

. C

imin

o

8.01

10

.6

Alk

ali

basa

lt

sam

ple

70

1.66

3.

04

Alk

ali

basa

lt

sam

ple

73

1.50

2.

72

0.2

74

0

.06

59

0

.07

54

0.2

32

0

.06

50

0

.07

56

0.27

7 0

.06

55

0

.07

87

0.2

40

0

.05

90

0

.07

77

0.2

25

0

.06

48

0.

0741

0.0

24

6

0.13

4

0.11

1

0.0

63

0

0.0

76

0

0.0

08

17

0.0

06

61

Page 5

289

librium systems. With regard to the occurrence of U in major rock-forming minerals, Rogers

and Adams (1969) have pointed out that the trace amounts of U may be present in them as: (1) isomorphous substi tution in the lattice; (2) concen- tration in lattice defects; (3) adsorption along crystal imperfections and grain borders; and (4) inclusions as microcrystais o:~ uranium minerals.

Although small U-rich crystals producing a high agglomeration of tracks as well as a higher density of fission tracks along the fractures and even grain borders have been occasionally observed, particularly in the granodiorite, U was determined only in clear unfractured crystals wi thout any visible inclu- sions and with a homogeneous distribution of tracks. The petrographic obser- vations also indicate that post-crystallization alteration probably did not in- fluence the U content of mineral phases since in all samples studied even the glassy matrix is fresh without any signs of oxidation. The homogeneous distri- bution of tracks together with the regularities of determined parti t ion coeffi- cients suggest that the measured U in mineral phases is, for the most part, present in solid-solution state.

As far as the acid volcanic rocks from Tuscany are concerned the problem of equilibrium between phenocrysts and glassy matrix has been treated in some detail by Dupuy (1970) who concluded that at least quasi-equilibrium was attained for a number of major and trace elements for sanidine, plagio- clase and biotite. This probably also applies to U which in each of these min- eral phases appears to be uniformly distributed throughout a given sample. Although or thopyroxene in these rocks is of xenocrystic origin, it still might yield meaningful partition coefficients provided that equilibrium was achieved. Dupuy (1970) has indeed suggested that even this mineral phase was probably in equilibrium with liquid.

With respect to alkali basalts, some petrographic features such as the ab- sence of reaction rims and corrosion and well-defined crystal boundaries with the matrix are consistent with the equilibrium conditions. On the other hand, as can also be ascertained from Table I, phenocrysts are composit ionally zoned implying that only their outermost zones may be in equilibrium with the matrix. The distribution of tracks in these phenocrysts does not, however, show any obvious zonation, at least not outside the limits of analytical pre- cision. This also appears from a comparison of data on cl inopyroxene in both analyzed samples. The U partition coefficient of the small c l inopyroxene crystals of the matrix of sample 73 is almost the same as that of large pheno- crysts of sample 70 although the major-element composi t ion of the former is rather similar to the outer zone of the latter. Thus, the partition coefficients measured for these zoned phenocrysts might indeed approximate the actual coefficients operative during the evolution of the magma.

RESULTS

U abundances of individual mineral phases and corresponding glassy ma-

Page 6

290

trices are given in Table II while the simple partition coefficients calculated as (U concentrat ion in phenocrys t ) / (U concentrat ion in matrix) (D U) are presented in Table III. Regarding the distribution of U in whole-rock samples, Table IV shows that only a small fraction of U is concentrated in the main rock-forming minerals. In the case of the granodiorite, most of U occurs in accessory minerals (zircon, allanite, apatite), although an appreciable amount of U appears to be present along the fractures and grain borders of the rock- forming minerals. In volcanic rocks from Tuscany, a major port ion of U is held by glass. But once again accessory minerals contain a significant part of the total whole-rock U concentration. It is of interest that the U content of individual mineral phases is about the same in volcanic and plutonic rocks, thus agreeing with a comagmatic origin as invoked by Marinelli (1961) and Dupuy (1970). The bulk of U in alkali basalts is contained in glass; no acces- sory U-rich minerals were observed in these samples.

TABLE III

Partition coefficients of U (D U × 103 ) between minerals and liquid

Biotite K-feldspar Plagioclase Clino- Ortho- Olivine pyroxene pyroxene

Ignimbrite San Vicenzo 23 6.4 7.5

Ignimbrite Roccastrada 23 5.4 6.4

Ignimbrite M. Amiata 21 5.1 6.7

Ignimbrite M. Cimino 21 6.1 7.0 6.0

Alkali basalt sample 70 44 2.7

Alkali basalt sample 73 9.0 41 2.4

A comparison of our data with previously determined U abundances of rock-forming minerals and their partit ion coefficients may be of some interest In general, the U contents of minerals of acid volcanic and plutonic rocks given in Table II are at least of an order of magnitude lower than those com- piled recently by Rogers and Adams (!969) . The concentrat ion of U in clino- pyroxene, olivine and plagioclase of alkali basalts is, however, comparable to the values reported by Nagasawa and Wakita (1968), Henderson et al. (197 i ) and Nishimura (1972) for the same mineral phases f rom basic vol- canic and plutonic rocks.

Page 7

TA

BL

E I

V

Co

ntr

ibu

tio

n o

f in

divi

dual

n~

ajor

ph

ases

to

wh

ole

-ro

ck U

co

nte

nt

Bio

tite

K

-fel

dsp

ar

Pla

gioc

lase

C

M

C

M

C

M

Py

rox

ene

C

M

Oli

vine

C

M

Co

rdie

rite

C

M

Mat

rix

C

M

U

(pp

m)

Gra

no

dio

rite

' M

. C

apan

ne

Ign

imb

rite

3

San

Vic

enzo

Ign

imb

rite

3

Ro

ccas

trad

a

Ign

imb

rite

2 M

. A

mia

ta

Ign

imb

rite

2 M

. C

imin

o

Alk

ali

bas

alt

sam

ple

70

Alk

ali

bas

alt

sam

ple

73

0.2

74

x

0.1

2

0.0

66

x

0.2

3

0.0

75

x

0.4

0

0.2

32

x

0.1

0

0.0

65

x

0.0

9

0.0

76

x

0.2

0

0.27

7 x

0.0

4

0.0

66

x

0.17

0

.07

9

x 0

.10

0.2

40

x

0.0

3

0.0

59

x

0.1

4

0.0

78

x

0.1

6

0.2

25

x

0.0

9

0.0

65

x

0.11

0

.07

4 x

0

.24

0.0

25

x

0.07

0.0

63

x

0.0

5

0.1

34

X 0

.24

0.1

11

x 0

.21

0.0

08

x

0.1

8

0.0

07

x

0.1

7

0.0

76

×

0.0

2

10.1

x

0.5

4

12.2

x

0.5

3

11.6

x

0.6

2

10.6

x

0.5

0

3.0

4 x

0

.51

2.7

2 x

0

.52

0.0

78

(4

.60

).

5.4

98

(7

.69

)

6.4

98

(8

.22

)

7.2

20

~

7.9

7)

5.3

48

(8

.01

)

1.5

84

(1

.66

)

1.4

41

(1

.50

)

C =

U c

on

ten

t (p

pm

); M

= m

od

al p

rop

ort

ion

. (

) D

eter

min

atio

n o

f U

co

nte

nt

in w

ho

le-r

ock

sam

ple

. '

Zir

con,

all

anit

e, a

pat

ite,

sp

hen

e (a

cces

sory

min

eral

s p

rese

nt)

. 2

Zir

con,

all

anit

e, a

pat

ite

(acc

esso

ry m

iner

als

pre

sen

t).

3 Z

irco

n, a

pat

ite

(acc

esso

ry m

iner

als

pre

sen

t).

b~

¢D

Page 8

292

There are only a few published data on partition coefficients for U. For natural clinopyroxene, Nagasawa and Wakita (1968) have given values (D U X 103 ) from volcanic rocks ranging 4--84, while Kleeman et al. (1969) obtained D U X 103 values of 4--10 from lherzolite inclusions of basanites. Seitz and Shimizu (1972) and Seitz (1973) have reported experimental data on clino- pyroxene-liquid and orthopyroxene-l iquid partitioning of U at high pressures and temperatures. Their data indicate that the partitioning of U is not ap- preciably dependent on the composi t ion of the melt or temperature or pres- sure. The D U X 103 of 8.0 for synthetic or thopyroxene given by Seitz and Shimizu (1972) is in good agreement with the hypersthene value in Table III (6.0).

The D U values for all rock-forming minerals in Table III are substantially lower than 1, resulting in a considerable enrichment of U in the residual liquid phase during differentiation. This is consistent with the nearly universal trend for U to increase towards the more differentiated rocks in a con- sanguinous igneous series. The data also indicate that the variation of U, par- ticularly in the more acid rocks might be strongly influenced by the acces- sory U-rich minerals.

PETROLOGICAL IMPLICATIONS

As an example of an application of partition coefficients of U, the ultra- mafic rocks of ophiolite suites and the olivine gabbro inclusions of alkali basalts are considered below. Ultramafic rocks of Tethyan ophiolites, which probably are slices of oceanic lithosphere tectonically emplaced on the con- tinent, correspond predominantly to harzburgites and lherzolites. The U con- tents of cl inopyroxene and olivine in a number of these rocks from the nor- t hem Appennines and Hellenides range from about 4.4 to 6.2 and from 0.3 to 1.3 ppb respectively (our unpublished data). If these minerals were in equi- librium with mafic magma, the U content of such a melt may be calculated using the D U given in Table III at about 100--150 ppb for c!inopyroxene and about 110--550 ppb for olivine. These U concentrations are similar to those found in typical oceanic tholeiites which contain 100--300 ppb (Tatsumoto et al., 1965; Aumento, 1971). Thus, the U data are consistent with models for the origin of ophiolites where ultramafic rocks are either cumulates from or residuum after extraction of the oceanic tholeiite liquid.

For magnetite-bearing olivine gabbro inclusions of alkali basalts from Iki Island, Japan, Kuno and Aoki (1970) invoked their derivation from moderate- ly differentiated alkali basalt magma, while Nishimura (1972) suggested that they represent residuum after the extraction of tholeiitic melt. The uranium content of olivine from these inclusions is about 5 ppb (Nishimura, 1972). Using D U reported here, the U concentrat ion in the hypothet ical liquid in equilibrium with the olivine would be about 1,850--2,090 ppb U. Such values are similar to those observed in some alkali basalts but differ notably from those of tholeiites, thus agreeing with the hypothesis of Kuno and Aoki (1970).

Page 9

293

ACKNOWLEDGEMENT

We thank Dr. C. Dupuy for providing the samples from Tuscany and Dr. C.A.R. Albuquerque for his helpful comments.

REFERENCES

Albarede, F. and Bottinga, Y., 1972. Kinetic disequilibrium in trace element partitioning between phenocrysts and host lava. Geochim. Cosmochim. Acta, 3 6 : 141--156.

Aumento, F., 1971. Uranium content of mid-oceanic basalts. Earth Planet. Sci. Lett., 11: 90--94.

Aumento, F. and Hyndman, R.D., 1971. Uranium content of the oceanic upper mantle. Earth Planet. Sci. Lett., 12: 373--384.

Barberi, F., Innocenti, F. and Mazzuoli, F., 1967. Contributo alla conoscenzo chimico- petrografica e magmatologica delle rocce intrusive, vulcaniche e filoniane del Campigliese (Toscana). Soc. Geol. Ital., Mere. 6: 643--681.

Coulon, C., Demant, A. and Bobier, C., 1974. Contribution du pal~omagnetisme ~ l'~tude des s d r i e s volcanique cdnozoiques et quarternaires de Sardaigne nord-occidentale. Tectono- physics, 22: 59--82.

Dupuy, C., 1970. Contribution ~ l'dtude des fractionnements gdochimiques des alcalins, des alcalino-terreux et du gallium au cours des processus magmatique.~. Thesis, University of Montpellier, Montpellier, 351 pp.

Dupuy, C. and Allegre, C.J., 1972. Fract ionnement K/Rb dans les suites ignimbritiques de Toscane (Italie). Geochim. Cosmochim. Acta, 36: 437--458.

Flanagan, F.J., 1973. 1972 values for international geochemical reference samples: Geo- chim. Cosmochim. Acta, 37: 1189--1200.

Henderson, P., Mackinnon, A. and Gale, N.H., 1971. The distribution of uranium in some basic igneous cumulates and its petrological significance. Geochim. Cosmochim. Acta, 35: 917--925.

Kleeman, J.D., Green, D.H. and Lovering, J.F., 1969. Uranium distribution in ultramafic inclusions from Victorian basalts. Earth planet. Sci. Lett., 5: 449--458.

.Kuno, H. and Aoki, K., 1970. Chemistry of ultramafic nodules and their bearing on the origin of basaltic magmas. Phys. Earth Planet. Inter., 3: 273--301.

Marinelli, G., 1961. Genesi e classificazione delle vulcaniti r~ centi toscane. Atti Soc. Toscana Sci. Nat. Pisa, Mere., 68: 74--116.

Mazzuoli, R., 1967. Le vulcaniti di Roccastrada. Atti Soc. Toscana Sci. Nat. Pisa, Mem., 74: 1--59.

Mazzuoli, R. and Pratesi, M., 1963. Rilevamento e studio chimico-petrografico delle rocce vulcaniche del Monte Amiata. Atti Soc. Toscana Sci. Nat. Pisa, Mem., 70: 355--429.

Mitchell, W.S. and Aumento, F., 1974. A geochemical comparison of minerals of oceanic and continental ultramafic origin. J. Geopbys. Res., 79: 5529--5532.

Mittempergher, M. and Tedesco, C., 1963. Some observations on the ignimbrites, lavas, domes and lava flows of Monte Cimino (Central Italy). Bull. Volcanol., 15: 1--16.

Nagasawa, H. and Wakita, H., 1968. Partition of U and Th between augite and host lavas. Geochim. Cosmochim. Acta, 32: 917--921.

Nishimura, S., 1972. Partition of uranium between peridotite nodules and host basalt. Chem. Geol., 10: 211--221.

Philpotts, J.A. and Schnetzler, C.C., 1970. Phenocryst-matrix partition coefficients for K, Rb, Sr and Ba with applications to anmthosite and basalt genesis. Geochim. Cosmochim. Acta, 34: 307--322.

Rogers, J.J.W. and Adams, J.A.S., 1969. Uranium. In: K.H. Wedepohl (Editor), Handbook of Geochemistry. Springer, Berlin, II/3, Sect. 92.

Page 10

294

Schnetzler, C.C. and Philpotts, J.A., 1968. Partition coefficients of rare-earth elements and barium between igneous matrix material and rock-forming mineral phenocrysts. In.' L.J. Ahrens (Editor), Origin and Distribution of the Elements. Pergamon, London, pp. 929--938.

Schnetzler, C.C. and Philpotts, J.A., 1970. Partition coefficients of rare-earth elements be- tween igneous matrix material and rock-forming mineral phenocrysts II. Geochim. Cosmo- chim. Acta, 34: 331--340.

Seitz, M.G., 1973. Uranium and thorium parti t ioning in diopside-melt and whitlockite-meit systems. Carnegie Inst. Washington Yearb., 72: 581--586.

Seitz, M.G. and Shimizu, N., 1972. Partitioning of uranium in the Di-Ab-An system and a

spinel lherzolite system by fission track mapping. Carnegie Inst. Washington Yearb., 71: 548--553.

Tatsumoto, M., Hedge, C.E. and Engel, A.E.J., 1965. Potassium, rubidium, strontium, thorium, uranium and the ratio of strontium-87 to strontium-86 in oceanic theoliitic basalt. Science, 150: 886--888.