American Mineralogist, Volume 68, pages l12-121, 1983
Variations in the chemicalcompositionsand lattice dimensionsof
(Ba,K,Na)-feldsparsfrom Otjosondu,Namibia and their significance
KnrsnNeuooRTHy VrsweNernAN AND Heowrc-Mnnrn
KrnlnonN
Mine ralogisch-Petrogr aphisches I nst it ut
TechnischeUniversitrit, Braunschweig
Federal Republic of Germany
Abstract
X-ray and chemicalstudiesof twenty samplesof (Ba,K,Na)-feldsparsfrom Otjosondu,
Namibia reveal that they can be divided into the following three groups: (l) Cn(Ba
Al2Si2Os)-poor,Ab-rich ones containing unmixed albite phases (2) Cn-rich varieties
showing occasionalunmixed celsian phases (3) those with Cn contents intermediate
between those of the first two groups. The compositionsof the unmixed phasesin the
secondgroup of specimens,determinedwith the electron microprobe,are Cn5aOr3aAb12
and Cne5Or5.
Almost all containconsiderableamountsof Fe3* (up to 10atomic Vo),which
probablyreplacesAl3+. As a result,their latticedimensionsare slightlylargerthan thoseof
other (Ba,K)-feldsparsreportedpreviously.
In order to estimatethe influenceof the Na+-ion on the lattice constantsof (Ba,K,Na)feldspars,all of the specimensweie ion-exchangedin KCI-melt. The lattice parameters
were correctedfor the presenceof Fe3* and Na+ and then correlatedwith volumesor Cn
contentsin order to determinetheir variationswith respectto chemicalcompositionand
structuralstate.
It was observedthat the cell edgesof (Ba,KFfeldsparsbehavesimilarly to those of the
plagioclases;however, the magnitudesof the changesinvolved are different in the two
groups. This, in turn, can be attributed to differencesin their respective chemical
composition and to certain structural features of feldspars in general.
Introduction
Natural and synthetic feldspars, particularly the
alkali feldspars and plagioclases, have been the
subjectof carefuland intensivestudy during the last
thirty years.Resultsof thesestudieshave revealed
that the physical properties and structures of all
feldsparmixed crystals dependupon the order/disorder of the (Si,Al) atoms and unmixing into different phases. However, both of these phenomena
were found, in turn, to dependupon the composition and structure of the end members involved.
For example, the similarity of the end members of
the alkali feldspars with respect to both these
features leads to simpler ordering schemes and
exsolution textures compared to those of plagioclasefeldspars.The latter exhibit very fine exsolution lamellaeand complicatedsuperstructures.Intergrowths of Ba-poor and Ba-rich feldspars
(Viswanathan,1978)in a few specimensfrom Otjosondu, Namibia, suggestthat the barium feldspars
are probablyso influencedby thesetwo phenomena
0003--(M)083/0
I 02-0I I 2$02.00
t12
that exsolutiontexturescoarserthan thosefound in
the basic plagioclasesappear. Hence, the aim of
this paper is to study the variationsin the chemical
compositionand structuralstateof (Ba,K,Na)-feldspars from Otjosondu in order to determine their
effect on the lattice parametersand, if possible,to
determine the exact composition of the unmixed
phasesin the specimensshowing exsolution. It is
hopedthat such investigationsmay also contribute
to the understandingof the complicatedsubsolidus
relations in the plagioclasefeldspars, which are
chemically similar to the (Ba,K)-feldsparsbecause
of the coupledsubstitutionof Na (or K) + Si by Ca
(or Ba) + Al. Preliminaryresultsof this study were
presentedby Viswanathanand Brandt (1978).
All the investigatedspecimensfrom Otjosondu
occur in the metamorphosed manganesedeposits
which were described by De Villiers (1952) and
Roper (1956).According to Katz (1978)the barium
feldsparsare found mainly in bandedand massive
manganeseores, which also contain yellow garnets
or pyroxene.The feldsparcrystals occur mostly in
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
pocketsof irregular sizes.The crystals are usually
large, some of them reachinga few millimetersor
even centimeters.Ba-feldsparswith compositions
Cnro (Cn refers to BaAl2Si2Os)and Cneawere first
reported by De Villiers (1951). Later Vermaas
(1953)describedtwo specimenswith the compositions Cn56and Cn55.
Gay and Roy (1968)investigatedboth synthetic
and partially ordered natural Ba-feldspars from
different localities. They correlated the lattice parameterswith Cn content and establishedthe following trends:
(l) The volume of Ba-feldsparsincreaseswith
increasing Cn content. The lattice dimension a
showsa substantialincrease.
(2) The angle B decreasesalmost linearly with
increasingCn content.
(3) The c-dimensiondecreaseswith disorderfor a
given composition.
(4) The synthetic high temperature(Ba,K)-feldsparsform a continuous solid solution series.But
for the natural, partially ordered ones they postulated two miscibility gaps-one in the Cn-poor region
(Cno-Cnro)to account for the observed complex
diffraction effects in single crystal photographsand
the other one in the Cn-rich region (Cnog-Cnao)
on
the basis of a compositional gap in the natural
specimens.
The figuresin Gay and Roy's paper show clearly
the above mentionedtrends but the exact positions
of the limiting lines are uncertain because of the
limited numberof availablespecimens,and because
of the considerablescatterof the plotted data. The
latter can be attributedto the following reasons:(l)
Only partial chemical analyseswere availablefor
many specimens.Moreover, the influenceof Na+
and Fe3* on the lattice constants could not be
estimated.(2) Someof the specimenswere zonedso
that an exact correlation betweenthe composition
and the lattice constantsmight not have beenpossible.
None of these difficulties arose in this work
becausethe Otjosondu specimenswere predominantly homogeneous,and completechemicalanalysescould be carriedout. In addition, Na-free specimenscould be producedby ion-exchangemethods.
Two unzoned specimens, one from Jakobsberg,
Sweden, and one from Yugoslavia, were also included in this study. Dr. Gay kindly supplied ten
samplesfrom his collection.The lattice constantsof
thesesampleswere redetermined(Table l), however, the amounts were not enough for complete
chemicalanalyses.
n3
Experimental
Guinier-Jagodzinskipowder patterns of all the
specimenswere made using CuKal-radiation, and
the lattice constants were refined using a slightly
modified version of C. W. Burnham's LcLSe program (Table 1). For monoclinicfeldsparsabout 4050 lines were used for refinement and for the
triclinic ones more than 65 lines were used.Precession photographswere made of many specimens
whose powder patternscontainedadditionalreflections indicatingthe presenceof more than one Bafeldsparphase.
The chemicalcompositionsof individual crystals
were determined with an electron microprobe.
Composition was determined at five different points
on each crystal and an average was taken. The
elementsdeterminedwere Ba, K, Na, Al, Fe, and
Si. No attempt was made to determine trace elements such as Sr, Rb, etc. As a preliminary study
indicated only minor amounts of calcium (<0.02
mol.Vo),it was also ignored during calculation of the
formula; all the iron presentwas assumedto be Fe3*
Oable 2).
In order to eliminate the effect of sodium on the
volume and the lattice constants,all the specimens
were ion-exchangedin KCI-melt at 900"C.Just as in
the plagioclasefeldspars (Viswanathan, l97la), a
finer fraction had to be used in the case of Ba-rich
specimensin order to obtain an almost complete
exchangeof sodium by potassium.The lattice parametersof the ion-exchangedspecimensare given
in Table 3.
All attempts to exchangethe potassiumin Bafeldspars by sodium in NaCl-melt proved futile.
Hence K-free (Na,Ba)-feldsparscould not be produced by this method.
Results
In this section we consider the characteristic
chemicalfeaturesof the (Ba,K,Na)-feldsparsfrom
Otjosondu.The compositionsof the 20 specimens
(Table 2) are plotted in a ternary Cn-Ab-Or diagram (Fig. l). Thesespecimenscan be divided into
three types:
( l ) C n - r i c h s p e c i m e n sw i t h c o m p o s i t i o n s
These show small amounts
Cna6-55Or33-arAbrr-rr.
of probablyunmixed,almostpure celsianas seenin
powder patterns.It is interestGuinier-Jagodzinski
ing to note that one of them (SpecimenNo. 16,
Table I and 2) is triclinic.
(2) Cn-poor, Ab-rich ternary feldsparswith the
These contain
composition Cne-21Or55-TrAbrz-rs.
varying amountsof unmixed albite.
114
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
Table l. Lattice constantsof natural (Ba,K,Na)-feldspars
S p e ci m e n
No.
I
2
4
5
6
7
I
9
l0
ll
12
13
l4
15
16
17
l8
l9
20
21
22
23
24
25
zo
27
28
29
30
31
3Z
o
8 . 5 4 0( 3 ) I 2 . 9 9 8 ( 3 ) I 4 . 3 9 6( 3 ) 9 0 . 0 0
115.95Q) 9 0 . o o
1437.0(9)*
8 . 5 0 9 ( 2 ) r 3 . 0 0 8 ( 2 ) r 4 . 4 0 0 ( 3 )9 0 . 0 0
1 1 5 . 9 4 ( l )9 0 . 0 0
1433.3(6)
8 . 5 1 4 ( 2 ) r 3 . 0 1o ( 3 ) 1 4 . 4 0 4 ( 4 )9 0 . 0 0
11 5. 9 3 ( 2 ) 9 o . o o
1434.8(9)
8 . 5 5 2 ( 2 ) 13.022(3) r l r . 4 rr ( 4 ) 9 0 . o o
1 1 5 . 8 6 ( 2 )9 0 .o o
I 4 4 4 . 2( 9 )
8 . 5 18 ( 2 ) 13.025(3) | 4 . 4 12 ( 4 ) 9 0 . 0 0
ll5.9l(2) 90.oo
r437.8(9)
8.515(1) 13.023Q)1\.\12(2) 90.00
1r5.90(r)90.00
1437.8(5)
L 5 6 \ ( 2 \ 1 3 . 0 1 9 ( 2 ) r 4 . 4 0 7( 4 ) 9 o. o o
1 1 5 . 8 7 ( l )9 0 . 0 0
r445.3(r)
8 . S z \ ( z ) 1 3 . 0 2 3 Q ) 1 4 . 4 1 4 ( 4 )9 0 . o o
1 1 5 . 8 9 ( 1 )9 0 . 0 0
1439.5(8)
8 . 5 4 2 ( 1 ) 1 3 . 0 3 6 ( 2 )14 . 4 ro ( 3 ) 9 0 .o o
1 1 5 . 8(11 ) 9 0 . o o
1444.5(6)
8 . 5 \ 6 ( z ) 1 3 . 0 4(13 ) 1 4 . 4 2(14 ) 9 0 . o o
115.82(2) 9 0 . o o
1446.9(9)
8 . 5 4 7 Q ) 1 3 . 0 4 5 ( 3 )1 \ . 4 2 6 ( 4 ) 9 0 . o o
1 1 5. 8 2 ( 2 ) 9 0 . 0 0
14 4 8 . 0( 9 )
8 . 5 \ 7 Q ) 1 3 . 0 4 5 ( 2 )r 4 . 4 1 (13 ) 9 0 . 0 0
1 1 5 . 7 3 Q ) 9 0 .o o
1 \ \ 7. 5 ( 7 )
8 . 5 5 5 Q ) l ' 3 . 0 5 1 ( 2 )1 4 . 4 r 9 ( 4 )9 0 . o o
1 1 5 . 7 8 ( 2 )9 0 . o o
1449.7(8)
8.575Q\ 13.055Q\ t\.4258) 90.00
t15.55(l) 90.00
1455.7(6)
8.576(2) r3.054(2) 1\.\22(3) 90.00
11 5 . 6 2 ( 1 ) 9 0 . 0 0
1 \ 5 5 . 7( 7 )
8.575Q) 13.053(2\ 1\.\22(3) 90.02(2) r 1 5 . 5 6 ( l ) 8 9 . 9 3 ( 1 ) 1 4 5 5 . 0 ( 7 )
8 . 5 8 2 ( 2 ) 1 3 . 0 5 8 ( 2 )r 4 . 4 2 0 ( 3 ) 9 0 . 0 0
| 1 5. 6 7( 2 ) 9 0 .o o
1\56.\(7)
8 . 5 7 5 Q ) 1 3 . 0 5 6 ( z \ 1 4 . 4 3 0 ( 3 )9 o . o o
r 1 5 . 5 (21 ) 9 0 . 0 0
1456.7(8)
8.570(z) 13.052(2) 1\ .\23(3)- 90.00
1 1 5 . 5 5 ( 1 )9 0. o o
1 t + 5. t2+( 7 )
(
4
)
8 . 5 7 8 ( 2 \ 1 3. 0 5 70 ) 1 4 . 4 3 r
115.64(2) 9 0 . o o
r457.1(8)
90.00
8 . 5 4 4 ( 2 ) 1 3 . 0 3 0 ( 2 )1 4 . 3 9(r4 ) 9 0 . 0 0
1t 5 . 6 8 ( 2 ) 9 0 . 0 0
I443.8(8)
8 . 6 3 4 ()1 1 3 . 0 4 9 ( 2 )1 \ . \ 1 2 ( z ) 9 0 . 0 0
r15.r4(1)90.00
1 4 7 0 .(15 )
8.555(2) 13.ooz(z) 14.397(\) 90.00(2) 115.92(2) 8 9 . 9 0 ( 2 ) 1 4 4 0 . 2 ( 9 )
8 . 5 3 5 Q ) 1 2 . 9 9 5 Q ) r 4 . 4 0 3 ( 3 )9 0 . 0 0
1 1 5 . 9 I r ( l )8 9 . 9 5 *
1\36.6(7)
8 . 5 7 6 ( 2 ) 1 3 . 0 2 2 ( z ) r 4 . 4 0 5 ( 3 )9 0 . 0 4 ( 2 ) 1 1 5 . 8(11 ) 8 9 . 8 3 ( r ) 1 4 4 8 . 2 ( 8 )
8 . 5 9 3 ( 1 ) r 3 . 0 3 8 ( 2 ) r 4 . 4 0 9 ( 3 )9 0 . 0 0
1 1 5 . 7 0 ( 1 )9 0 . ( ) o
1454.5(6)
8 . 5 \ 7 Q ) 1 3 . 0 2 9 Q ) 1 4 . 3 9 5 ( 3 )9 0 . 0 0
11 5. 7 6 ( 1 ) 9 0 . 0 0
1 \ \ 3 . 70 )
8 . 5 6 2 ( 2 ) 1 3 . 0 3 7 ( 3 ) r 4 . 4 0 3 ( 4 )9 o . o o
115.73Q) 90.00
14 4 8 . 3( 9 )
8 . 5 5 5 Q ) 1 3 . 0 3 7 Q ) 1 4 . 3 9 5 ( 3 )9 0 . 0 0
1 1 5 . 5 8 ( 1 )9 0 . 0 0
1447.0(6)
8 . 6 3 (11 ) t 3 . 0 4 8 ( r ) 4 . 4 0 9 ( 2 ) 9 0 . o o
ll5.l5(l) 90.oo
r469.0(4)
8 . 5 4 0 ( 1 ) 1 3 . 0 4 3 ( t ) 1 4 . 4 0 4 ( 2 )9 o . o o
r5.16(1) 90.00
1469.2(5)
8 . 5 3 6 ( 1 ) 1 3 . 0 4 7 ( 1 )r 4 . 4 0 8 ( 2 ) 9 o . o o
1469.5(Ir)
t 1 5 . 1 5 ( 1 )9 0 . 0 0
N o t e s o n s o u r c e s ; I t o 2 0 f r o m O t j o s o n d u ; 2 ' l f r o m Y u g o s l a v i a ; 2 2 f r o n J a k o b s b e r g , S w e d e n ;2 3 t o 3 2
s p e c i m e n so f G a y a n d R o y ( 1 9 6 8 ) , t h e l a t t i c e c o n s t a n t s o f w h i c h w e r e d e t e r m i n e d b y t h e a u t h o r s .
T h e . n u m b e r gs i v e n b y G a y a n d R o y a r e : ( 2 3 ) g 1 t 1 5 6 , ( z \ ) g 2 1 0 0 9 , Q 5 ) g 1 1 1 3 0 , ( 2 6 ) F . S \ , Q 7 ) 8 . A . 8 \ 7 6 5 ,
( 2 E ) B . r.' f 3 r 9 9 0 , ( 2 9 )1 9 5 8 5 7 , ( 3 0 ) s r R / 5 / 1 5 , ( . 3 1I)0 0 9 7 8 , ( 3 2 ) B . H 1
. 958,284.
*Estimated standard errors are ln brackets and refer to the last declmal place.
**o
which tend
a n d y a r e e s t l m a t e d b a s e do n t h e w i d t h o f t h e l l n e s ( e , 9 . 1 3 2 , \ 2 \ , 5 S o , T e z , 2 8 0 ) ,
to split in triclinic varleties.
(3) Specimenswith intermediate compositions,
which show neither unmixed albite nor celsian.
Thus the first two types appear to limit the
compositional variations of the Otjosondu
(Ba,K,Na)-feldsparswith respectto the celsianand
albite end members.
Another interesting feature is that all these specimenscontain considerableamountsof Fe3+ (0.050.13, Table 2). It increaseswith celsian contenr
reaching a maximum value (0.17) in an unmixed
celsian phase. Thus, the incorporation of Fe3+ in
these specimens appears to depend upon the
amount of aluminium present in the structure. On
the other hand, the albite content decreaseswith Cn
content. Though this trend is expectedbecauseof
the tendency of albite to unmix from (Ba,K,Na)-
feldspars, it is noteworthy that even Ba-rich feldspars(type 1) containas much as 10-12molVoAb in
solid solution. Naturally-occuffing Ca-rich plagioclasesdo not contain more than 4-5VoOr molecule
(Corlett and Eberhard, 1967).
The ion-exchangeexperiments also revealed that
(Ba,K)-feldsparsbehave differently from (Ca,Na)feldspars.It is known that Na+ in plagioclasecan be
exchangedwith K* and that the latter in the exchangedplagioclasescan be reexchangedwith Na+
(Viswanathan,1972).In the (Ba,K,Na)-feldsparsit
was possibleto replacethe small Na'-ion with the
larger K+-ion, but an exchangeof K+-ion by Na+ion could not be achieved either in the natural
ternary Ba-feldspars or in the K-exchanged Bafeldspars. All attempts to exchange K* by Na*
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS l15
Table 2. Chemical formulas of (Ba,K,Na)-feldspars,
determinedby microprobe analysis and calculatedon the basis
ofeight oxygen atoms
Fe'
Ba
Sp No.
0.09r
0.139
0.143
0.17t
0 .1 8 9
o t9o
0.202
0.205
9
t0
0.302
0.545
0.6!5
0 515
0.532
0.569
0.590
0.711
0.581
0.4811
0.489
0 3\9
0.215
0.234
0.213
0.213
0-220
0.122
0.20r1
0.210
0.200
0.027
0.056
0.059
0.087
0.051
0.050
0.010
0.088
0.096
0.088
0.998
r.078
1 10t
1.143
1 - 1t 2
r r3o
1 203
1. 1 2 7
r.099
1. 1 9 2
0.207
0.20r
0.165
0.!35
o .I 0 5
0.091
0.061
0.088
0.110
0.081
0.r33
0.129
0.127
0.120
0.12i
0.137
n.d.
0.109
0-123
0.| 29
0.135
o.t3o
00
l 183
1 231
r.300
1. 3 3 9
1.3j6
1. 3 5 2
t.!ol
1. 3 7 6
r .389
1 389
l . 4 1r
n.d.
n.d.
r4
15
0.308
0.338
0.393
0.469
0.47r
17
r8
19
20
o.lr97
0.5r5
0.54r
0.545
0.5It6
0.5r3
0.q41
0.448
0.405
o.\27
0.388
0.353
0.356
0.345
0.335
21
22
0.397
o.95
0.455
0.05
ll
12
n.d....not
AI
delemined.
See Table
1 for
Table 3. Lattice parametersof ion-exchanged(Ba,K)-feldspars
Speciren
No
si
2.963
2.867
2.8trr
z 781
2.822
2-820
2 781
Z-790
2 781
2.717
2.711
2.701
2.610
2.5\6
2 539
2.517
2.481r
2 480
2 \68
2 \13
2.592
n.d.
1 3 . 0 1 3 Q ) 1 4 . 4 i l( 4 ) 11 5 . 9 6 ( 2 ) 1 \ j t . z ( g ) x
1 3 . 0 2 6 ( 2 ) 1 4 . 4 1 9 ( 3 ) i l 5 . 9 r ( 2 ) r 4 5 q . 4( 7 )
1 3 . 0 2 2 ( 2 ) 1 4 . 1 2(r3 ) i l 5 . 9 r( r ) 1 1 r 5 4 . 2 ( 6 )
r 3 . 0 3 4 ( 2 )1 4 . t 2 5 ( 4 ) r 1 5 . 8 5 ( 2 ) 14 5 7. 5 ( 8 )
1 3 . 0 2 6 ( 3 )1 4 . l r l 6 ( l ) 1 1 5 . 8 6 ( 2 ) r 4 t 5 . o ( 8 )
8 . 6 0 7 Q i 1 1 . 0 2 9 ( 3 1 I r { . 4 r8 ( 4 ) 1 1 5 . 8 5 ( 2 ) r r r 5 5 . 0( 9 )
8 . 5 u 0 ( 2 ) 1 3 . 0 r 7 ( 2 ) r ! . q 0 5 ( 4 ) 1 1 5 . 8 3 ( 2 )1 4 5 r . 3 ( 8 )
8 . 6 0 8 ( r ) r 3 . 0 3 3 ( r ) 1 4 . 4 1 9 ( 2 )i l 5 . 8 4 ( r ) 1 4 5 5 . 9 ( 9 )
8 . 6 0 1( 2 ) r 3 . 0 3 9 ( 3 ) r 4 . A 1 r r ( 3 ) 1 1 5 . 7 5 ( 2 ) r 4 5 6 . r( 8 )
8.612(2\ 13.0\7(3) r4.424(rr) l't5.7\(21 1 q 5 o . o ( 9 )
8 . 6 0 7 ( 2 ) r 3 . 0 4 3 ( 3 ) r 4 . 4 2 0 ( 4 ) 1 1 5 . 7 3 Q ' I 4 5 8 . 3( 9 )
8 . 6 1 3 ( 2 ) 1 3 . 0 4 7 ( 3 )l 4 . r r 2 l ( 4 ) 1 1 5 . 6 8 ( 2 )1 4 5 0 . 7 ( 9 )
8 . 6 1 1 ( l ) l 3 . o ! 8 ( 2 ) 1 4 . 4 1 7 ( 2 )i l 5 . 6 7 ( 1 ) r 4 5 9 . 9 ( 5 )
8 . 6 r 9 ( r ) r 3 . 0 6 r ( r ) 1 4 . q 2 3 ( 2 ) i l 5 . 6 0 ( l ) 146rr.3 (4)
8 . 6 1 7 ( l ) 1 3 . 0 5 9 ( 2 )1 4 . 4 2 5 ( 2 )1 r 5 . 5 9 ( r ) I 4 6 r rI. ( 5 )
*
8.6r5(2) 13.060(2)14.\26(3\ 115.59(r) 1463.7(6)
8 . 6 r 9 ( 2 ) r 3 . 0 6 5 ( 2 ) r 4 . 4 3 0 ( 3 ) 1 1 5 . 5 9 ( 2 ) I 4 6 5 . 4( 7 )
(
6
)
8 . 5 1 6 ( 1 ) 1 3 . 0 6(12 ) 1 4 . 4 2 6 ( 3 )1 1 5 . 5 0 ( 1 )14 6 4 .I
8 . 5 t 6 ( 2 ) 1 3 . 0 6 1 ( 2 )1 \ . \ 2 7 { . 3 \ 1 r 5 . 6 0 ( r ) 14 6 4 .r ( 7 )
8 . 6 t r r ( 2 ) r 3 . 0 5 7 ( 2 ) r4.424(4) 115.59Q) 1 4 6 3 . 2 ( 8 )
8.607(2)
8.607(2)
8.509(1)
8.6r4(2)
8.609(2)
8
9
10
1l
tt
r4
t<
l5
17
rg
r9
20
8 . 6 0 0 ( 2 ) r 3 . 0 3 8 ( 2 ) 1 4 . 4 0 3 ( 3 )i l 5 . 7 3 ( t ) 1 4 5 4 . 9 ( 7 )
22w
sources.
always lead to the destructionof Ba-feldsparsand
formation of sodalite. This appearsto indicate that
the K-equivalentsof partially-orderedplagioclases
can exist metastably, but the Na-equivalentsof
(Ba,K)-feldsparscannot.
One of the main aims of this investigationwas to
determine the exact composition of the unmixed
phases, especially the Ba-poor phase. The end
membercompositionsgiven by Viswanathan(1978)
were derived from the lattice constants. which are
affectedby the presenceof minor amountsof Na+
and Fe3+as shown later. On the basisof this study
of Cn-rich feldspars, the following observations
were made:
(l) Powder patternsof all but one Cn-rich specimen show the presenceof minor amounts of the
unmixed celsian phase (about 5Voas estimatedby
the intensities of the powder lines). Precession
photographsof single crystals chosen from these
specimensrarely showed the celsian phase. It is
therefore concluded that the unmixing is coarse,
and the unmixed celsian is not homogeneously
distributed throughout the specimens.In all the
precessionphotographsin which both phaseswere
seen,the orientation relationshipwas the sameas
that observedby Viswanathan(1978),namely, the
a*-axes of the unmixed phasesare split, whereas
the b* and c* coincide.
The one specimenwhich showedlarger amounts
of celsian in powder patterns and which showed
more frequently both unmixed phasesin the precession patternshad an approximatebulk composition
(Viswanathan,1978).
of Cn71Or27Abz
+Estimated standard errors are in brackets and refer to the last
decimal place.
#Calculated rcnoclinic ignoring the small splitting of a few lines.
*tNot
exchanged.
(2) It was not possible to observe the unmixed
phasesmicroscopically.Attempts to locate the unmixed celsian with the electron microprobe in the
specimenswith minor amounts of celsian proved
futile. Hence only the compositionof the Ba-poor
phase could be determined on these specimens,
(averageof
which was found to be Cn5aOr3aAb12
five values).
(3) Only in the specimens containing larger
amounts of celsian could two types of unmixing
(Fig. 2 and 3) be observed.In Figure 2 a photomicrograph made with BaKo radiation shows three
regions: (a) A darker region containing only a Cnpoor phasewith a compositionof about 50-52VoCn,
36-35VoOr. l4-l3Vo Ab (Fe3* almost nil). (b) A
Fig. l. Variations in the chemical composition of the
(Ba,K,Na)-feldspars from Otjosondu, Namibia. O: Cn-rich
specimens; O: Cn-poor specimens; O: Specimens with
intermediatecompositions.
l16
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba.K.Na)-FELDSPARS
synthesizedor all the Ab-componentwhich is invariably present in the natural (Ba,K)-feldspars
must be ion-exchanged.Becauseordered (Ba,K)feldspars have not been synthesizedso far, the
second method was used to produce ordered or
partially ordered specimens.As the ion-exchanged
feldsparsfrom Otjosondu representonly a limited
compositionalrange, a few specimensof Gay and
Roy (1968)havealso beenused.Becausethe chemical analysesgiven by them were not always complete, it was necessaryto find an indirect methodto
estimatethe amountof sodiumin the specimensand
to determinewhich are sodium-free.
A diagram (Fig. a) using a plot of a againstcell
volume was chosen for this purpose, as both are
considerablyaffectedby the exchangeof Na* by
K+. A[ Na-free (Ba,K)-feldsparsshouldfall on the
curvejoining pure sanidine(or microcline),synthetic (Ba,K)-feldsparsof Gay and Roy (1968)and pure
celsian.Those containingNa should lie below the
Fig. 2. X-ray (Ba.iKc)area scanshowingthe coarseunmixing in
a barium feldspar from Otjosondu (see text). Scale bar = 17 p.
curve. The ion-exchangedOtjosondufeldsparsplotted very slightly above this curve, thereby showing
bright transitional region very near the sharp slightly larger valuesfor a. This could be attributed
boundary containing dark lamellae of apparently to the presenceof larger Fe3* ion in the structure,
Cn-poor phase, whose composition could not be which is assumedto be in tetrahedralcoordination.
determined exactly becausethe lamellae were narAs the data of Otjosondu feldspars cannot be
rower than the electron beam. However, the aver- used further without correcting for the effects of
age Cn content increased abruptly at the sharp Fe3+two independentmethodswere consideredto
boundary. (c) As the distancefrom the boundary estimatethem. The first one was to determinethe
increases,there exists a region (not shown in the effects in synthetic feldspars containing Fe3*. A
figure) where the dark lamellae completely disappear and the Cn content reachesa maximum value
of about 95% (Ab almost nil, Fe3* : 0.17).This
coarserunmixing can be compared with a finer one
(Fig. 3), in which Cn-rich and Cn-poor (dark) regions are traversed by a system of fine lamellae,
which terminate abruptly in the lower Ba-rich region.
In view of the identical orientation relationsbetween the Cn-poor and Cn-rich phase in all the
specimens,and considering the similar compositions of the specimensshowing small amounts of
unmixed celsian, it is concluded that the compositions of the unmixed phases must be about
Cn52-5aOr35-roAbrr-rz
and Cne5-lsoOrs-o.
Having consideredthe variationsin the chemical
compositions,let us now consider their effects on
lattice parameters. In order to determine the
changes in the various lattice parameters with
increasing Cn content in the binary system
BaAl2Si2Os-KAlSi3O8,
either (Ba,K)-feldsparswith
Fig. 3. X-ray (BaKa) area showing the finer unmixing (see
different (BalK)-ratios and structural statesmust be text). Scale bar : l7 p
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
comparisonof the volume of KAlSi3Oswith that of
KFeSi3Os(type HS or LM; Smith, 1974,p. 219)
revealsthat the replacementof Al by Fe3* resultsin
an increasein volume by about 50A3 (considering
the 144-cell). Again, the volume of BaAlFeSi2Og
(V: 1515.143
a ,: 8 . 7 8 A ,
b:13.169A,
":
14.5734,F : 115.04;Pentinghaus,1981,personal
communication)is larger than that of BaAl2Si2Og
(Smith, 1974,p. 219) thereby suggestingthat the
replacementof one Al by one Fe3* in the formula of
celsian increasesthe volume by 45A3. If a linear
influenceis assumed,a replacementof 0.I Al by 0.1
Fe3* shouldcausean expansionin volumeby about
+.543 in all (Ba,K)-feldspars.
The second approach is based on the reasonable
assumptionthat the cell volumes of ordered and
disordered(Ba,K)-feldsparsmust be nearly equal
because the volumes of high sanidine and low
microcline are nearly equal. Hence the discrepancies in the volumes of the ion-exchangednatural
specimenscan be determinedby comparingthem
with those of synthetic specimenswith same Cn
contents(Fig. 5). By plotting the differences(AV)
againstthe observedFe3* contenta slightly smaller
value of 3.5A3for the replacementof 0.1 Al by 0.1
Fe3* was obtained for the Otjosondu specimens.
Using this observedvalue the volumes of the samples in this study were corrected.
*
(43)
Fig. 4. The a cell dimension of (Ba,K,Na)-feldspars vs.
volume. x : (Ba,K)-equivalentsof (Ba,K,Na)-feldsparsproduced
by ion-exchangeof specimensfrom Figure 1; O: Ion-exchanged
specimenscorrected for their Fe3* content; O: Natural feldspars
of Gay and Roy (l%8); I: Synthetic feldspars of Gay and Roy
(1968); O: Natural (Ba,K,Na)-feldspars from Otjosondu,
Namibia; Cn: celsian, Sa: sanidine,Mi: microcline. The straight
lines join ion-exchangedspecimensand thus show the efect of
ion-exchange.
ll7
vl
(43)|
1470
1460
1450
1140
50+
mol o/oCn
Fig. 5. Variation of unit cell volume of a synthetic (Ba,K)feldspar with Cn content (data from Gay and Roy, 1968).
The next problem was to determinethe effectsof
Fe3* on different lattice dimensions.A comparison
of the data of the synthetic specimensshows that
the increaseof 4.5A3in volume is accompaniedby
an increaseof about 0.008A in a, 0.011A in b,
0.0204 in c and a negligible change in B. As a
correspondingreplacementcausesa smallerexpansion (3.543)in Otjosonduspecimensa proportional
increase in the different lattice parameters (La :
0.006,Ab : 0.008,Ac : 0.016)was assumedto be
due to the incorporationof Fe3*.
After the correction, the Otjosondu specimens
plotted (Fig. a) slightly under the curve for Na-free
varieties in accordance with the fact that they are
partially ordered.As there was little An-moleculein
any specimens,it could be safely ignored in all
theseconsiderations.
Figure 4 shows the following features:
(1) An exchange of Na by K produces larger
changes in a and cell volume than the coupled
substitution of (Ba+Al) by (K+Si). Hence this
diagram can be used to estimate the Ab content of
(Ba,K,Na)-feldspars.
(2) In Ba-poor (Ba,K)-feldspars a changes only
slightly with increasingCn content, but more rapidly in Ba-rich varieties.
(3) Fe3*-free(Ba,K)-feldsparsfall on or near the
curve joining celsian and sanidine.As microcline
plots near sanidine, the effect of ordering of (Al,Si)
atoms can be considered negligible, especially in
Cn-rich varieties.
(4) If specimensplot above the curve joining Sa
and Cn, they are likely to contain considerable
amounts (>0.05 atomicVo)of Fe3* in tetrahedral
coordination.Thus, Ab content and Fe3* content
show opposite efrects.
ll8
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba.K.Nd-FELDSPARS
Gay and Roy (1968) suggested that a diagram
correlating p with Cn content is most suitable to
determinethe Cn in (Ba,K)-feldspars.This is illustrated in Figure 6 where B is plotted against cell
volume. It is noteworthy that the exchangeof K by
Na considerablydecreasesthe volume, but only
slightly increasesB. As microcline is also plotted, it
can be inferred that the ordering of Si/Al atoms has
a negligible effect on B. Hence pvalues of natural
barium feldspars containing small amounts of Abcomponent and possessingdifferent degreesof order can still be used to determine the approximate
Cn content. A suitable diagram correlating p with
celsian content can be drawn using Figures 5 and 6.
Figure 6 clearly shows the expansion in cell volumes caused by the additional Fe3+, which increaseswith increasing Cn content in the K-exchangedOtjosondufeldspars.
The variation of the b-dimensionwith Cn content
of almost Na-free (Ba,K)-feldspars is shown in
Figure 7. As the synthetic sanidine of Gay and Roy
(1968)possessesa smaller b-value than the high
sanidine, it must be assumedthat the former is
partially ordered. The K-rich ion-exchangedOtjosondu specimensshow (after correction for Fe3*content) even smaller values, thereby indicating
that they are more ordered. It can be seenthat there
is only a small change in the b-dimension of the
disordered(Ba,K)-feldsparseries,high sanidineto
celsian.
In order to study the variation of b in partially
ordered natural (Ba,K)-feldspars,a few specimens
of Gay and Roy (1968)(Cn > 35 molVo)were also
included in Figure 7. Though a few of them may
contain small amounts of Ab in solid solution. ion
BI
(') |
116.0
115.5
115.0
V (A)
Fig. 6. Lattice angle p of (Ba,K,Na)-feldsparsplotted against
volume. x: Data point of unmixed celsian (Viswanathan, 1978).
See Figure 4 for other symbols. The straight lines join ionexchangedspecimens.
b1
( A )|
13.06
66od
13.04
13.02
I
.
*
Y"'a
;t
r
a.-
t
t
t
t'
o
,'556
-d
13.
12 . 9 8
mol
o/oCn
Fig. 7. Variation of D of (Ba,K)-feldspars with Cn content
(molVo).See Figure 4 for explanation of symbols.
exchangeexperimentshave shown that 10molVoltb
decreases b only insignificantly in such Cn-rich
specimens(about 0.0064). The partially ordered
natural (Ba,K)-feldspars show a rapid increasein b
for K-rich compositions until about Cn_15
and then
almost no change occurs. Viswanathan (1971b,
1972)and Ribbe (1975)made a similar observation
in low plagioclases and arrived at the conclusion
that the rapid increase in b-parameter of natural
plagioclases from Ans to approximately An36-35
must be attributedto the disorderingof Al,Si atoms,
which accompanies increasing calcium content.
The similarity in the nature of variation of b is also
reflected in the magnitudes.The difference in the &
value of albite (low) and anorthite (Smith, 1974,p.
219) is 0.088A, whereas that between microcline
(low) and celsian is 0.0804. A discussionof the
behaviorof b is presentedlater.
The best parameterto determine the differencein
the structuralstatesof Fe3*-free(Ba,K)-feldsparsis
c (Fig. 8) becauseit is more sensitiveto changesin
the structural states of specimenswith Cn content
greater than 30-40 molVo. The cell volume was
chosen as the abscissainstead of Cn content becauseit is possibleto considerall (Ba,K)-feldspars,
with or without sodium. Thus the effect of ionic
exchangeon c and cell volume can be simultaneously shown. In Figure 8 the synthetic feldspars studied by Gay and Roy (1968), all their natural feldspars for which complete analyses are available,
and all the natural and ion-exchanged feldspars
from this study are plotted after correcting for the
presenceof Fe3*.
The following observations can be made regarding this figure:
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
l19
structure determination. It should be possible to
estimate the approximate AVSi-distributions in the
14.tr4
tetrahedral positions, T1 and T2, if structure data of
only a few more barium feldspars were available,
and if curves joining (Ba,K)-feldspars having similar structural states(as defined by Viswanathan and
tco
Brandt, 1980)are drawn in this figure.For example,
the Cn-rich specimenlies below the line (not shown
I
in Fig. 8) joining the structurally similar Spencer
a^
>q
I
"C" and the Cn-poor specimen(Viswanathanand
Brandt,
1980), thereby suggestingthat it should
_}
have
less
than0.494Al in T1 and more than 0.21I Al
V (43)
in T2-positions,which is in agreementwith the
Fig. 8. c-dimensionof (Ba,K,Na)-feldspars)plotted against
observeddata.
volume. A: (Ba,K,Na)-feldspars for which structure data are
(5) Many ternary (Ba,K,Na)-feldsparsfrom Otjoavailable; tO: Natural (Ba,K,Na)-feldspars from Otjosondu,
corrected for their Fe3+-content.SeeFigure 4 for explanationof
sondu lie outside the field Mi-Sa-Cn. This is due to
other symbols; B, C: refers to sanidines,Spencer "B" and
the effect of the sodium in solid solution. It is
Spencer "C" (Colville and Ribbe, 1968); The dashed line
noteworthy that ion-exchangecausesall the speciconnects unmixed barium feldspar phases(corrected for Fe3*joining the plots of
content, see text). The continuous lines join ion-exchanged mens to move towards the line
phases
the
Ba-poor
and
Ba-rich
of the unmixed
specrmens.
specimens(Viswanathan, 1978).This line can be
assumed to represent the approximate boundary
(l) All the disorderedsyntheticspecimensof Gay line for Fe-free,Na-free(Ba,K)-feldsparswith maxand Roy have the smallestc-values.For a specimen imum possibleorder, and thus it differs in position
with a given cell volume the higher the order of from that published earlier (Viswanathan and
(Si,Al)-atomsthe larger is the c value. In making Brandt, 1980).Hence it can be concludedthat many
this conclusionthe small changesin volume caused (Ba,K,Na)-feldsparsfrom Otjosonduare relatively
by orderingare ignored.
ordered.This should not be surprisingbecausethe
(2) The c-dimension of a disordered synthetic very presenceof a few unmixed specimensindi(Ba,K)-feldsparincreasesrapidly at Ba-poor com- cates that the P-T conditions in Otjosondu were
positionsand only slightly at the Ba-rich composi- favorable for ordering. Another explanation is that
tions, a trend which is oppositeto that of a.
the considerable amount of the Ab component
(3) The fact that c of celsianis less than that of present in these specimensmust have accelerated
microclineclearly indicatesthat the increasingdis- the ordering and probably also the unmixing in
order of Si/Al-atoms (considering the 7A-cell) them. The observation that albite or Ab-rich feldwhich accompaniesincreasing Cn content has a sparsorder faster than Or-rich ones seemsto supdiminishingeffect on c as in plagioclases(Viswan- port this view. Probably both factors are important.
athan, 1972).However, the magnitude of reduction
Discussionand conclusions
is smallerthan that in plagioclases.A discussionof
the behaviourof c is presentedlater.
It is well known that the lattice parameters of
(4)That the c-parametercan be usedto determine feldspars are influenced differently by the size of
the Al/Si-distributions on (Ba,K,Na)-feldsparsis the largerinterstitialcations,the sizeofthe tetrahedemonstrated using the data of two barium feld- dral cations and finally by AVSi order-disorder. In
sparswhose structureshave been refined (Viswan- the case of monoclinic (Ba,K)-feldsparsthe lattice
athanand Brandt, 1980;Viswanathanand Kielhorn, angles, a and y, are constrainedto be 90.0" and
1983).It is seenin Figure 8 that one of them, the Cn- hence, only B varies with Cn content. It is not
poor variety, has a larger c-dimension,even after therefore possible to compare the variations in the
coirecting for the presenceof Fe3*. The Cn-rich lattice angles of triclinic plagioclaseswith those of
specimen, on the other hand, has a smaller c- monoclinic (Ba,K)-feldspars and to correlate them
parameter both before and after the exchange of with compositionalchanges.Hence further discusNa. This suggeststhat it possessesa more disor- sion will be confined only to unit cell volumes and
deredAl/Si-distribution,which is confirmedbv the lattice dimensionsand will include an interpretation
rirt
120 VISWANATHAN AND KIELHORN: COMPOSITIONS ANt
only of the trends and not the exact magnitudesof
the changesin different lattice parameters.
First, we examine the volumes of the high-temperature disordered plagioclases and (Ba,K)-feldspars. Becausethe mean Ca-O bond in anorthite is
0.16A shorter than the mean Na-O bond in high
albite, and the meanAl-O bond is 0.134 largerthan
the mean Si-O bond, the coupled substitution of Na
+ Si by Ca + Al should be compensatingin its
effect on volume. Thus the fact that the volumes of
all high-temperature plagioclasesare nearly equal
can be understood. Similarly, as the mean Ba-O
bond is only 0.054 shorter than the mean K-O
bond, a larger difference in the volumes of sanidine
and celsianshouldbe observed.In the caseof low
plagioclasesthe volume is also affected considerably by the changes in the (Al,SiFO framework
accompanyingAl/Si order, whereasin the (Ba,K)feldspars the larger interstitial cations appear to
permit only very small changesin the framework.
Though the a-dimensionshows only small changes
in Ba-poor (Ba,K)-feldsparsits behavioris, in general, similar to that of volume in both groups of
minerals.
The other two lattice dimensions.b and c. are
considerably influenced by the distribution of the
Si,Al atoms, as in alkali feldspars (Wright and
Stewart, 1968, Stewart and Ribbe, 1969).The bdimensionsof the high-temperatureplagioclasesare
nearly equal, whereasthose of the high-temperature
(Ba,K)-feldsparsincrease slightly (about 0.0154)
towards celsian. Though this trend is the same as
that exhibited by the cell volume, the magnitudeof
the change in b is much smaller. Two different
causesare probably responsiblefor this effect. First
is the influenceof the coupled substitution,which
hasa smallresultanteffecton b of (Ba,K)-feldspars.
The second, probably a more important reason,
appearsto be the inability of a feldspar with a given
(Al,SiFO framework to expand in b direction beyond a certain value. This can be inferred if one
comparesthe b-dimensionsof high sanidine and
high RbAlSi3Os(13.0294and 13.034A,respectively; Smith, 1974p. 219) or those of low microcline
and low RbAlSi3Os029964 and 12.964A,respectively; Smith, 1974p. 219).lt is observedthat after
b reaches a value of about 12.964. it is more
influencedby a changein the framework causedby
ordering (or disordering) of Al/Si atoms than by
replacementof existing interstitial cations by even
larger ones. Perhapsthis also explainsthe smaller
changes noted in b (as compared to c) where
tetrahedralAl3+ is replacedby Fe3+.
LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
The rapid increase in b in natural (Ba,K)-feldspars (and plagioclases)must be attributed to the
increasing Al/Si disorder which accompanies increasing Cn content (An content). Hence the
change in b can be considered as a measureof the
accompanyingorder/disorder in Ba-poor feldspars.
In both low plagioclasesand (Ba,K)-feldspars,the
b-dimension increases rapidly until about 3V40Vo
(An or Cn, respectively) and then it remains constant. In both cases, a doubling of the cell must
occur in the An-rich (or Cn-rich) members.Hence
the distinct two-trend variation of b cannot be
considered accidental. The explanation that b increases rapidly until 30-40% An in plagioclases
because disordering takes place predominantly
within the 7A-cell appears plausible and can be
applied to (Ba,K)-feldspars.The inability of b to
increase further in Cn-rich members can also be
attributed to another type of disorder that takes
place in them. Viswanathan(l97lb, p. 41) suggested that this ordering involved the enrichment of Al
in a tetrahedral position in one 7 A-cell and a
reduction of Al in the corresponding tetrahedral
position in the adjacent cell-a process, which
should ultimately produce an anorthite-like cell.
This process need not necessarily take place between two adjacent cells. It may take place over
regions containing a few cells as indicated by the
superstructuresin more calcic plagioclases.
The c-dimension of disordered plagioclase feldspars decreaseswith increasing anorthite content,
whereas that of disordered (Ba,K)-feldspars increaseswith increasingCn content. This behavior
clearly revealsthat the differencesin the meanbond
lengths of the interstitial cations and those in the
mean bond lengths of the tetrahedral cations have
larger effects on c than on D. This should not be
surprisingbecausesubstitutionof K by Rb in Rbfeldspars(c : 7.250A in low RbAlSi3O8,7.1824 in
high RbAlSirOs;Smith, 1974,p.219) causesc to
expand much more than b, and becausethe variation in c in feldspars containing Fe3* is more
sensitive to the size of the tetrahedral cation.
The reduction of c of high plagioclasesmust be
attributed predominantly to influence of the smaller
Ca-O bond whereas the increasein c of disordered
(Ba,K)-feldsparsmust be due to the dominating
influence of the larger Al-O bond. In examining the
variation of c in natural, ordered or partially orderedfeldspars,the influenceof AllSi distributions
should also be considered. All three factors affect
the c-parameter of low plagioclasesin such a way
that only an approximate linear relationship with
VISWANATHAN AND KIELHORN: COMPOSITIONS AND LATTICE DIMENSIONS OF (Ba,K,Na)-FELDSPARS
anorthitecontent is observed(Fig. 2, Bambaueret
al., 1967).The available data (Fig. 8) are inadequate
to make similar statements regarding (Ba,K)-feldspars; we may, however, state that a two-trend
variation, as shown by b, is not observed.
It is now recognizedthat Al/Si disorder is introduced in the feldspar structure either by heating a
low albite (or low microcline) or by the coupled
substitutionof Na * Si by Ca + Al (or of K + Si by
Ba + Al), and that such a disorder leads to an
increasein b and a reduction in c. The differencein
behavior of the b-c relationship may lie in the fact
that alkali-feldspars do not show any doubling of
the unit cell. As order/disorder in alkali-feldspars
takesplace only within the 7A unit cell, the second
type of disorder never becomes important, and
hence both lattice constants are subjected to the
sametype of disordering.Thus the influenceof the
second type of disorder appears to be different on
the two lattice constants b and c of plagioclases.
The influenceon b is such that there is no further
changein it. But c, which is more sensitiveto the
size of tetrahedral cation, shows a net resultant
effect. Thus c appearsto be more useful for determining the structural state of the whole plagioclaseor hyalophane-series,whereas b appearsto be a
more sensitiveindicator in the range of 0 to 3V40Vo
(An or Cn, respectively).
Before concluding,it is relevant to comment on
the triclinic (Ba,K,Na)-feldspars. Gay's triclinic
specimenshave a Cn content less than 30 molVo,
whereasthosefrom Otjosondupossessas much as
49 molVo.This rather surprising observation suggests that the composition at which the change of
symmetryoccursin naturalbarium feldsparsshould
be at least50 molVoCn. If it can be assumedthat the
degreeof triclinity shown by a (Ba,K,Na)-feldspar
is a function of its chemicalcompositionand structural state, and that the triclinity values of the
barium feldsparswith maximum possibleorder (as
definedby Viswanathanand Brandt, 1980)increase
with decreasingCn content, then we concludethat
the structural statesof two specimens,Nos 16 and
25, Table 1, are nearest to the maximum order
possiblefor their respectivecompositionsas compared to others.
Acknowledgments
We are grateful to Professor E. Woermann, Mineralogical
Institute, Aachen, for lending the specimensfrom Otjosondu.
Thanks are also due to Dr. P. Gay, Cambridge, England, for
lendingten samplesofbarium feldsparsfrom his collection. Prof.
E. Eberhard, Hannover, Drs. H. Kroll and H. Pentinghaus,
Mtinster, Prof. M. Okrusch, Braunschweig, and Prof. J. V.
Smith, Chicago, critically read the manuscript.
l2l
References
Bambauer,H. U., Eberhard, E., and Viswanathan,K. (1967)
The lattice constantsand related parametersof "Plagioclases
(low)". SchweizerischeMineralogische und Petrographische
Mitteilungen, 47, 351-164.
Colville, A. A. and Ribbe, P. H. (1968)The crystal structure of
an adularia and a refinement of the structure of orthoclase.
American Mineralogist, 53, 25-37.
Corlett, M. and Eberhard, E. (1967)Das Material fiir chemische
und physikalische Untersuchungenan Plagioklasen.Schweizerische Mineralogische und PetrographischeMitteilungen,
47,303-316.
De Villiers, J. E. (1951)The manganeseores of Otjosondu,
South-WestAfrica (Namibia). TransactionsGeologicalSociety of South Africa, 54, 89-98.
Gay, P. and Roy, N. N. (1968)The mineralogy of the potassiumbarium feldspars, III, Subsolidusrelationships. Mineralogical
Magazine,36,914-912.
Katz, K. (1978) Metamorphe Silikat-reiche Manganerze der
Lagerstette Otjosondu (Siidwest-Afrika). Diplomarbeit, Uni
versitat Braunschweig,W. Germany.
Ribbe, P. H. (1975)Optical properties and lattice parametersof
plagioclasefeldspars. In P. H. Ribbe, Ed., Feldspar Mineralogy, p. R53-R72. Mineralogical Society of America, United
Statesof America.
Roper, H. (1956)The manganesedepositsat Otjosondu, SW
Africa (Namibia). Symposium on manganese,XX., International Geological Congress,II, Mexico.
Stewart, D. B. and Ribbe, P. H. (1969)Structural explanationfor
variations in cell parameters of alkali feldspar with AUSi
ordering. American Journal of Science, 267-A, 4'14 462.
Smith, J. V. (1974) Feldspar Minerals, Volume 1., Crystal
structure and physical properties. Springer-Verlag, Berlin
Heidelberg New York.
Vermaas,F. H. S. (1953)A new occurence of Ba-feldsparat
Otjosondu, SW-Africa (Namibia) and a X-ray method for
determiningthe composition of hyalophane.American Mineralogist,3E,845.
Viswanathan, K. (l97la) A new X-ray method to determine the
anorthite content and structural state of plagioclases.Contributions to Mineralogy and Petrology, 30,332-335.
Viswanathan, K. (l97lb) Kationenaustauschan Plagioklasen.
Habilitationsschrift, University of Hannover, West Germany.
Viswanathan, K. (1972) Kationenaustausch an Plagioklasen.
Contributions to Mineralogy and Petrology, 37,277-290.
Viswanathan, K. (1978) Intergrowth of Ba-rich and Ba-poor
phasesin barium feldsparsfrom Otjosondu, South-WestAfrica (Namibia). Mineralogical Magazine, 42, 294.
Viswanathan,K. and Brandt, K. (197E)Geordneteund entmischte Barium-Feldsptiteaus Otjosondu/Siidwestafrika(Namibia). Fortschritte der Mineralogie, 56, Beiheft I, 140.
Viswanathan,K. and Brandt, K. (1980)The crystal structure of a
ternary (Ba,K,Na)-feldspar and its significance. American
Mineralogist.65, 47247 6.
Viswanathan, K. and Kielhorn, H.-M. (1983)Al,Si distribution
in a ternary (Ba,K,Na)-feldspar as determined by crystal
structure refinement. American Mineralogist, 68, 122-124.
Wright, T. L. and Stewart,D. B. (196E)X-ray and optical study
of alkali feldspars:I. Determination of composition and structural statefrom refined unit-cell parametersand 2V. American
Mineralogist,53, 38-87.
Manuscript received, January 4, 1982;
acceptedfor publication,August 23, 1982.