Canadian Mineralogist
Y ol.' ?-6,pp. 67-79 (1988)
PETROGENESIS
AND CONDITIONSOF CRYSTALLIZATIONOF SPANISH LAMPROITIC
ROCKS
GIANPIERO VENTURELLI aNI EMMA SALVIOLI MARIANI
Istituto di PetrogroJio,Universitddi Parma,43100Parma,Italy
STEPHEN F. FOLEY
Max Plancklwtitut filr Chemie,Abt. Kosmochemie,
65M Mainz, WestGermony
SILVIO CAPEDRI
Istituto di Mineralogiae Petrologia,Universitddi Modena,41100Modena,Itoly
ANTONY J. CRAWFORD
Departmentof Geology,Universityof Tasmania,Hobart, Tasmania,Australio
ABSTRACT
The lamproitic rocks of southeastSpain (8.6-6 Ma) occur
mainly as dykes and plugs in the provinces of Murcia,
Almeria and Albacete. Phlogopite, olivine, apatite, spinel,
sanidine and glass are ubiquitous. Clinopyroxene,
ortiopyroxene, K-rich amphibole,biotite, ilmenite, pseudobrookite, leucite, analcimeand carbonatesmay be additional phases. The magmas crystallized at temperatures
below ca. l200oC. Oxygen fugacity during crystallization
varied from ca. WM at Calasparra to values intermediate betweenMn1_"O- Mn3O4(MnH in the text) and NNO
at Jumilla and Aljorra. The /(HF)f(H2O)
and
ratios were very low (ca. 10.s- l0-4). The
l(Hcl)/f(tt2o)
distribution of high-valency(>3) elementswas largely
influenced by the heterogeneousdistribution of accessory
phasesin the sourceand by the melt structure. In particular, the Ol(Si+Ti+Al+Fd+)
ratio and P2O5are well
correlated, suggestingthat the melts were saturatedin apatite. The magmasweregeneratedfrom an enriched mantle
sourcecontaining phlogopite, ofthopyroxene, olivine and
accesory phases(e.g., apatite).The source,however,before
metasomatism, probably underwent loss of a magmatic
componentas suggestedby the low contentsof Ca, Na,
Al and Sc.
Keywords: lamproites, southeastSpain, mineral chemistry,
petrology.
SovuarnE
Les rocheslamproitiquesdu sud-estde I'Espagne(provinces de Murcia, Almeria et Albacete) se pr€sentent surtout sousforme de dykeset de petitesmassifs. Phlogopite,
olivine, apatite, spinelle, sanidine et verre sont tres repandus. On rencontre aussi clinopyroxbne, orthopyrox€ne,
amphibole potassique,,biotite, ikn6nite, pseudobrookite
mago€sienne,leucite, analcimeet descarbonates.Les magmas ont cristallis6 e une temp€rature infdrieure i environ
1200'C. La'fugacitd d'oxygbne6tait plut6t variable pendant la cristalisalion, prCsdu tampon WM i Calasparra
et interm&iaire entre les tampons MnprO - Mn3Oa(MnH
dans le texte) et NNO d Jumilla et Aljorra, Les rapports
f(IJn/fGl2O) etf(HCl)/f(H2O) dtaient trOsbas (environ
l0-5 - l0-4). La distribution des6l6mentsir valence6lev6e
(>3) a 6tdinfluenc€epar la distribution h€tdrogdnedesphasesaccessoires
i la sourceet par la structure du magma.
En particulier, le rapport O/(Si + Ti + Al + Fd' ) et P2Oj
montretrt une bonnecorr€lation, qui fait penserquelesmagmas dtaient saturesen apatite. Ces magmas ont pris naissancedans une rdgion du manteau enrichie par m6tasomatisme et pofieusede phlogopite,orthopyroxene,olivine et
phasesaccessoires,
dont I'apatite. Avant le m€tasohatisme,
par contre, la r€gion aurait perdu une composantemagmatique, commel'indiqueirt les faibles teneursen Ca, Na,
AI et Sc.
(Traduit par la R6daction)
Mots-cUs: lamproites,Espagne,chimismedes mindraux,
p€trologie.
INrnonucrldu
The ultrapotassic rocks of lamproitic character
fGroup I n Foley et al. (l 987)] of southeastem Spain
are distributed over a wide area (provinces of Murcia,
Almeria and Albacete; Fig. 1) mainly in the form
of plugs and dykes. They range from 8.6 to 5.5(?)
Ma in age (Nobel et al. l9Sl,Bellon et ol. 1983) and
belong to the Na-poor Neogene magmatic province
of southeastern Spain. Igaeous suits of this province
(Lopez Ruiz & Rodriguez Badiola 1980, Puga 1980)
consist mainly of calcalkaline and shoshonitic volcanic rocks; lamproites are volumetrically scarce.
The Spanish lamproites show the following features: (i) magmatic plagioclase is typically absent;
(ii) silica normally exceeds 5390, except for some
Jumillasamplesi MB, Ni, Cr, Zr,Th, LREE, U, Pb,
Ba, Sr and Rb are high (Rb may be relatively low
at Jumila); (iii) the 87Sr/89Srvalue is unusually high
67
68
THE CANADIAN MINERALOGIST
O JUMILLA
}
l o c a nc a n t x
.O
LASMINAS
SPAIN
a
O C A L A S P AR R A
O.TFORTUNA
Almoria
urcia
OTzENETA
tr
M U L AO \
A fr i c a
BARQ
Lorca
tr
A L J O R R AI
Cartagena
Frc. 1. Locations of the outsops of lamproitic rocks in southeasternSpain.
(0.717504.7?183),whereasthe 143Nd/144Nd
value is
low (0.51I 19-0.51126);lead-isotoperatios arein the
range 206pb/2o4pb18.66-18.81, 207pb/204pb
15.67-15.74 and 208Pb/2uPb39.0-30.2.
Detailedpetrographicand geochemicalinvestigations on the Spanishlamproitic rocks werereported
in Fuster et al. (1967),Borley (1967),Fernandez&
HernandezPacheco(1972),Lopez Ruiz & Rodriguez
Badiola (1980),Venturelli et al. (1984),Nixon e/ a/.
(1984)and Nelsonet ol. (198A. Our main aim in this
paperis to contribute to the knowledgeof the crystallization history aqd genesisof theseunusualrocks.
Mgl(Mg+Fe2*)l ranging from 79-82, mainly
occursin severaloutcrops (e.9., Aljorra, Cancarix,
Fortuna) as microlitesof the groundmass,and very
rarely as reaction coronas mantling olivine. Biotite
is an additional phaseatZenetaand Vera. Sanidine
is the most abundant groundmassphasein the glasspoor or glass-freerocks. The glassof the groundmass is very rich in Si (64-7190 SiO2 on a dry
basis).K-rich amphibole(3.04.5t/o K2O, 5.0-6.390
CaO), very high in Ti (5.2-7.94/o
TiO) and very low
in Al (0.0-1.790Al2O) crystallizedlater than sanidine in the groundmass.The amphiboleis yellow to
pinkish and someis mantled by greenrims which are
PstnocRApHy ANDMTNERALCHEMISTRy
Ca-freeand Fe-rich (19-210/oFeO,o,u1).
Carbonates
of presumedmagmatic origin (Fuster et al. 1967,
The petrographic featuresof the Spanishrocks Pellicer 1973)occur at Aljorra, Jumilla and Vera.
with lamproitic affinity are highly variable. All the Sparsecrystalsof leucite occur in somesamplesfrom
rocks are porphyritic and somemay be very rich in Jumilla and Vera @uster et al. 1967).Other phases
glass @ortuna, Vera, Zeneta). Many rocks from
are apatite,Cr-spinel,Fe-Ti-spineland, sometimes,
Jumilla are holocrystallineand coarsegrained. Oli- analcime, ilmenite and pseudobrookite.
vine and phlogopite are presentin variableproporSmall ultramafic fragments and xenocrysts of
tions and are the most abundantmafic phases.Cli- mafic phasesmay be present.They include kinked
nopyroxene may be abundant in some outcrops olivine (Mg# 90-94), orthopyroxene (Mg# 90) and
(Aljorra and Jumilla). The Spanish ultrapotassic clinopyroxene (Mg# 90). Xenoliths/xenocrystsof
rocks are the only lamproites containing magmatic cnrstal origin may also occur (mica schistsand "graorthopyroxeneas a consequenceof the high silica nitic" fragments, quartz, K-feldspar, plagiocontent. Magmatic orthopyroxene, with Mg# [00
clase,kyanile, andalusite,limestones).
69
PETROGENESIS OF SPANISH LAMPROITIC ROCKS
ANDCLINOPYROXENE
TABLE1. REPRESENTATIVE
MICROPROBE
DATAFOROLIVINE,OR1HOPYROXENE
0l ivine
lr
si02
lc
--r
AF
ssc
40.50 40.69 40.00 40.47 N.40
--r
sc
54.32
rt02
0.60
0.6.|
0.46
A1203
1,38
0.80
0.04
FeZ03
Feo
12-06
l4n0
0.24
Cao
t{aZ0
0..|8
Nio
nd
0.38
cr203 nd
0.04
0.08
0.i6
0.07
0.13
'1.87
0-29
nd
0.08
sx
54.78
53.00
-
0.21
53.45 54.60 5 3 . 1 0 5 2 . 0 0
0.84 0.94 0.80 0-29
0.48
0.86
4.05
0.07
0.46
0.32
0.14 0.36 8.22 13.36 5.54
1.38
1.19
3.13
2.66
0.35
3.67
0.20
0-12
0 . 11
33.09. 18.51
16.32
0 . 1 4 0.20
17.59. 't7.70 17.40
17.94
19.85 22.15
22.59 22.50 22.40
22.33
0 .t 8
0.73
0.16
nd
nd
0.15
12
??
0.36
27 aA
0. 13
1.56
1.55
1.27
0.58
0.05
0.08
0.14
0.07
0.58
0.74
0.04
0.06
0.02
0.30
nd
0.07
0.17
0.15
0-07
0.31
0.55
0.32
-
0.18
29.90
,]F
FAF
xlrS
5 5 . 0 0 5 5 . 0 0 55.75
54.25
12.30 8.53 13.67 n.u, ,.r, 12-1't
0.33 0.13 0.29 0.15 0.09 0.32
r,190 45.30 49.83 45.57 49.02 4t).60 28.98
*
Cl lnopyroxene
0rthopyroxene
JCL
xx
0.10
0.33
u.50
4.20
0.16
0.33
nd
-
Total 98.61 99.68 99.87 99-76 98.54
t9.tz
9l-t5
98.t1
98-79
99.92
99.45
97.88
99.30 r00.59 99-U
xs S
81.0
81.5
85.5
78.9
9 1. 4
9t .3
9'1.6
89.4
86.8
Juntlla, cL
J
91.2
85.6
90.0
Aljorra, F
Calaspaffa, A
l, s . Iarge, small crystal,
90.3
Forluna, Cx
z.ot
1.00
88.2
93.0
0.07
1.05
98.93
93.4
Cancarix. values expressedln wejght g.
x ; xenocryst, c r core, r : rrm, nd " not determined,Mg, - lOOxMg/(Mg+Fe2+)
Fe203for pyroxenehas been calculated accordingto Cawthorna Collerson (1974).
r
after venturelli eL al.(1984).
A detailed description of the mineralogy is
reported in the papers cited above. Here we confine ourselvesto mineralogical data that will be used
for petrogeneticinterpretations,and also describe
minerals that have not been described carefully in
previous papers.
Olivine
Phenocrystsof olivine are widespreadand have
a magnesiancore(Mg# ca.87-91)and a magnesiumdepletedrim (Mg# ca. 83-88; Table l). The matrix
olivine is usually enriched in iron. During the late
stagesof crystallization, some olivine was rimmed
AMPHIBOLE,
SANIDINEANDAPATI1E
TABLE2. REPRESENTATIVE
MICROPROBE
DATAFORPHLOGOPITE,
Amphibole
Phlogopite
FJCX
poik
1r
lc
--r
cLz
Lc
Sm
PM
I
st02
37.10
39.57
40.90
40.23
42.04
39.97
40.51
Ti02
7.68
8.55
2.27
4.58
1.92
3.96
41203
12.14
9.78
12.24
11.70
11 . 9 4
12.89
1.87
'13.54
FeZ03
2-03
1.56
0.67
l-13
0.60
1.94
0.87
5 3 . 0 2 5 3 . 1 5 52.43
6.3't
7.90
5.15
1.05
0-22
0.77
,l
64.40
64.28
4.95
0.36
0.43
16.60
17.42
3.06
1.40
-
Feo
7.78
5.95
2.56
4.34
2.31
7.44
a.9z
o.vJ
4.50
18.96
0.09
0.i0
0.07
0.02
0.04
0.04
o.02
0.10
0 .l 0
0. I t
0.60
It1g0
16.49
19.33
23.88
21.89
24.74
19.19
23.12
18.19
15.42 18.90
6.28
7.57
-
0.08
0-07
0.36
0.30
0.16
0.09
0.1,
o.*
5.39
4.86
4.93
7-51
Kzl
9.49
10.30
10.42
10-46
9.65
9.09
10-20 3.63
4.44
3.09
4.34
1.81
0.48
14.10 .|6.10
PeoS
nd
nd
nd
nd
nd
nd
nd
nd
nd
3.04
2.09
4.40
3.35
2.27
6 A'
':'u
o:t'
cl
0.10
0.05
2.94
-
nd
1 .t 6
nd
F
0.03
0-02
96.30
97.58
96.36
97.18
94.80
97.57
97.87
nd
Lt
:
99.27
99.2'l
ZJPM
0.38
0.71
0.35
--':*
0.26
0.37
0.12
0.51
0-27
0.47
53.53
52.63
53.93
40.94
40.72
41.57
2.85
3.04
2.47
1.09
0.18
0.30
99.67
97.72
99.82
0.09
0.15
Na2Q
rotal
o.,o
cx
53.01
Mno
Ca0
Apattte
Sanldln€
clil
sc--rls
nd
-
9 7 . 5 1 9 7 . 0 5 1 0 0 . 5 5 1 0 0 .1 9
FFortuna,JJumilla,CXCancarix,CLCalasparra,ZZenetaiPMPuebladeMula,CMLasMinasdeHellin'
Fe2O3for phlogopite has been calculated from Feo tot. considering Fe203/Fe0= 0.26 (Fuster et al. 1967)Values expressed in weight g'
I'lhere Fe2O3is not reported, Feo = Feo tot.
70
THE CANADIAN MINERALOGIST
AI
AI
I
Eas
sld
,r-----
Nbppm
30
40
50
1a
60
Frc.2. Top part of the diagramshowsthe relation betweenphlogopiteand host-rock
chemistry.GroupA includestlle outcropsof Jumilla, Cancarix,Ias Minas de Hellin,
Calasparra;group-B the outcropsof Vera, Aljorra, Zeneta,Fortuna, Puebla de
Mula, Barqueros. At lower left, arrows in diagram (Al).-(Al). indicate the pattern of zoning toward the rim and in the microcrysts. (Al). and (Al)' rrpresent
the All(Al + Mg + Fe) ratio (atomic) in phlogopiteand in the rock, respectively.
At lower ight, Zr, M, a1rdMgO are bulk-rock contents. Data after Venturelli
et al, (1984).
by orthopyroxeneand phlogopite. Kinked xenocrysts
(Mg# 90-94) occur in places.The uranium content
of olivine has beendeterminedby the fission-track
method. The phenocrystscontain lessuranium than
xenocrysts(e.g., ll-22 ppb and 58-l15 ppb, respectively, for samplesfrom Calasparra).
ELE}JENTS
IN fiE PHLOGOPITE
TABLE3 . BEHAVIOR
OF SOI'48
Mtcagroup
A
B
ATi/A(l,1g,Fe)
-0.97
-0.99
.
(a) lver + vlAl * 3(l,rs,Fe)
un " no llgnlflcant
corelation
Ali/AAI
-1.6
un
ATi/ASI
un
-1.5
lvTj * zVlti t zvtu
AAI/ASi
un
0.3- 0.4
Brown mico
Phlogopite(Table 2), presentas phenocrystsand
microcrysts, commonly showsa Fe-Ti-enriched rim
(FeOrora
and TiO, up to 11.7 and ll.39o, respectively). Two main typesof phlogopitemay be recognized (Fig. 2). Group A (Jumilla, Cancarix,
Calasparra, Las Minas de Hellin) exhibits a low
All(Ca+Fe+Mg) atomic ratio and a trend subparallel to the phlogopite-annite join. Group B
(Vera, Aljorra, Fortuna, Barqueros, Zeneta,Puebla
de Mula) hashielrer AV(AI + Fe + Mg) ratios and follows the phlogopite-siderophyllitejoin. The host
rocks differ chemically (Fig. 2). The rocks with a
lower All(Al + Fe + Mg) ratio contain mica of group
A, whereasmicas of group B occur in rocks with
ratio, suggesting that
higher All(Al+Fe+Mg)
7l
PETROGENESISOF SPANISH LAMPROITIC ROCKS
moderate variations in rock chemistry exert an
important control on mica composition. In most
casesthe mica has (Si + Al) lessthan 4 and exhibits
a correlationbetweenTi and (Mg + Fe) with a slope
= I (A TilA(Mg,Fe) = 1). This trend differs from
that usually found in mica from K-rich rocks (Arima
& Edgar l98l). ConsideringTi astetravalentand disregardingthe role of Fe3*, the behavior of groupA micas may be explained through reaction (a)
shown in Table 3; the behavior of group-B micasis
more complicated and difficult to interpret.
All samplesof mica investigated are rich in fluorine (l-4.490), reflecting the high fluorine content
of the host rocks (cc. 300-2600ppm; R. Vannucci,
pers. cornm., 1983).High contents of fluorine in
mica of lamproites and similar rocks have been
reported by Cross (1897), Prider (1939), Jones &
Smith (1983),Scott-smith & Skinner(1984),Jaques
(1986)and Foley el o/. (1986b).
Oxides
Representativebulk compositions of the oxide
minerals are shown in Table 4.
Spinel. Five different types of spinel have been
identified (Fig. 3); more than one type may appear
in the samesample: (i) Chromite (CHR, Cr2O3in
the range 39-63t/o) exhibits low to very low Al
(O.9-3.40/o
Al2O) and a Cr/Fe3+ ratio greaterthan
2:l
1.8
.9
1.6. The chromite occurs mostly in olivine
phenocrystsand may be consideredas anear-liquidus Frc.3. Chemicalfeaturesof the spinels(atomsin the forphase.(ii) Cr magnetite(CrMT) hasintermediateCr,
mula on the basisof 4 atomsof oxygenand 3 cations).
low to very low Al (sample34 from Aljorra, AlrO,
CHR: chromite, CrMT: Chromian magnetite,TiMTI
Titanian magnetite,AISP: Al-rich spinel, SPH: spinel5.8390, is an exception)and a CrlFe3* ratio less
hercyniteseries.34, samplefrom Aljorra.
than 1. (iii) Ti macnetite(TMT) hasvery low Cr and
OF OXIDE.PHASES
TABTE4. REPRESENTATIVE
CO}IPOSITIONS
t&PB
AISP
J
inol
2.76
Tt02
A1203 0.94
Cr203 39.09
24.71
hz0:
24.71
FeO
0.95
lfno
5.39
l'lg0
0.23
l{10
CX
inolr
9.99
0.08
32.31
17.99
33.90
0.71
3.68
0.44
Cl.l
lnol
1.A
2.15
63,62
4.63
17.85
-
Total
98.78
99.10
99.80 100.36 98.89
r"$ s
Cr f
36.8
61.0
34.6
65.1
6-2
89.3
J Junilla,
inol
10.14
0.12
CL
inol
1.37
2.50
63.35
3.s3
21.23
0.4u
7.90
,t
gn
8.38
0.21
10.65
42.04
34.20
0.69
2.@
4.8
89.9
7g.5
20.9
Cx Cancarix, Cl'l|Las l,llnas de Hellln,
gm groundnass.
F
x
0.U
48.68
14.24
2.16
10.15
0.25
18.86
0.34
JJCXFfi
gm
0.29 50.32
62.65
0.10
4.30
7.76
20.84 35.67
r.13
0 .1 6
4.77
16.54
0.02
cn
52.13
0.02
0.77
3.68
36.97
0.60
5.24
gn
66.12
1.32
0.33
8.45
12;97
0.15
8.00
gn
65,61
99.77
99.41
97.U
98.16
98.7
1.3
82.0
17.9
77.A
3.2
98.3
1.7
98.91 100.95
2.2
19.6
4.2
-
0.i9
12.00
12.12
4.24
CL Calasparra, F Foriuna.
ln ollvine,
* = too"r"*/(ellcrrFe}),
cr, = looxcr/(Al+cr+Fe}), atmtc rattos.
Fe203 calculaled consideflng: total cations - 3 and valsces .8 for splnel, total cations = 2 and
total catlons = 3 and valences = 10 for iig-pseudobrooklte.
ralences ' 6 for ilrenlte,
MgPB llgseries, lL ilrentte'
SPHsplnel-hercyntte
AISP Al-spinel,
CHRchrcnite, CrilT Cr-mgnetite,
r"*
-Dseudobrcok lte .
72
THE CANADIAN MINERALOGIST
Al (<0.290 Cr2O3and ca,0.130/oAlrOr). Along the
sequenceCHR*CrMT-TiMT,
the Fe3+,/Feratio
increases(Fig. 3). (iv) Al spinel (AISP), very low in
Ti (0.1-0.2590TiO), is eitherincludedin xenocrystic olivine or occurs as separateanhedral crystals;
it is very similar to the spinel of lherzolite noA,rt.i
common in alkali olivine basalts.(v) One category
of spinel(SPH) belongsto the spinei-hercyniteseries;SPH is very poor in Cr (<0.1590 Cr2O3),poor
in Ti (ca. 0.390 TiO) and has a very low Fe3*,/Fe
ratio, suggestinglow /(O) conditions. A similar
type of spinelhasbeendescribedin WestKimberley
(Jaques& Foley 1985)and Leucite Hills lamproites
(Kuehner et ol. L98l). In the West Kimberley lamproite, SPH has beenrelatedto exsolutionof nonstoichiometricleucitesupersaturatedin Mg, Al and
Fe (Jaques& Foley 1985). In the Spanish rocks,
spinelSPH hasdifferent features:(a) SPH may occur
as tiny emerald-greengrains contained in spongy
brown mica (usuallyunstablebiotite) along with sanidine(?)+ ilmenite; spinelgrainsmay be arrangedin
fine-grainedaggregatesmimicking the morphology
of previous mica flakes. (b) SPH may constitute
aggregatesrimmed by phlogopite, generatingcharacteristic snow-ball textures. Since the miscibility
between MgCrrO. and MgAlrOa end membersis
complete(Muan et al. 1972),spinelof types(a) and
(b) did not crystallizefrom the samemelt that generated spinelCHR. Spinelof types(a) and (b) probably representthe product of tire Urea'tOownat diffetent stagesof a former unstablebiotite that wasricher
in Fe and Al compared to the newly formed
phlogopite'
A pseudobrookite-likemineral, identified in the
groundmassof somesamples,appearsassmall subhedral deepreddishgrainswith a low birefringence.
In termsof the end membersMgTirOr,-FeTirO,and
fgrTlor, the composition is about 46, 37-42 and
12-17 moleo/0.The mineral is very similar to the
Fe3*-bearingarmalcolite found by Velde (1975)in
the Smoky Butte lamproites.
ctinopyroxen
Clinopyroxenecommonly occurs as microcrysts
and rarely as phenocrysts(Table l). Approximate
calculations of the end memberr 1Cu*tnoro e
Collerson l9?4) suggestthat the clinopyroxeneii
generally cnaracteriiid by high f"r- .o"i""t iop to
i.ofo fero;. The aluminum content is low, as ii
usual foi l6w-Al ultrapotassicrocks. Xenocrystic
pvroxene
(Me#
e0)is&countered
onrv
rarerv.
Sonidine
Sanidine(Table 2) is presentin the groundmass;
in some coarse-grainedrocks from Jumilla it may
constitute large poikilitic crystals. It may have high
contents of Fe3* (up to 3.570 FerOr), Ti (up to
0.43V0TiOJ and Ba (up to 1.990BaO, Venturelli
et al. 1984)-.
Apatite
Apatite is alwayspresentand is locally abundant
as euhedralcrystals(Jumilla, Puebla de Mula). It is
very rich in fluorine (Table 2)'
EvaluarroN oF TEMrERATURE
oF EeurLrBRArroN
Olivine-melt equilibrium has been tested in the
Calasparra samplesin which (i) olivine is the only
large phenocrystpresentand (ii) its weight proportion may be correctly evaluated(ca. ?.0t/owt.). To
test for equilibrium and to determinethe temperature of crystallization, we usedequation (3) of Ford
et al, (1983) and equation (17) of Ghiorso el a/.
(1983).The former is an empirical equation that considersthe influenceof compositionand structureof
the melt on the componentactivities;the latter uses
an iterative calculation taking into account the
mineral-liquid equilibrium constant and the activities of the liquid and mineral components. Both
equations,however, call for FqO3 concentrations
in the melt. Since chromite crystallized contemporaneously with the olivine phenocrysts at
Calasparra, the Fe3+,/Fd+ ratio of the melt has
value obtained
been calculated using the f$S;J+
by Foley (1985)for the Caussberglamproite, which
has a chemistry very similar 19 t[at oi the dpanish
ultrapotassicrocks and contains comparableir-rich
spinei (Sheraton& Cundari 19g0,Foley l9g5). The
timperature of crystallizationof olivine so obtained
@a. 1200.C),is not far from that obtained during
ihe experimentson Gaussbergmaterial (ln\"C;
n'oleyiSAS)with no volatile pr6sent.The calculated
t..p".utui, of olivine crystallizationmust be consideiedwith caution becausethe Spanishlamproitic
magmas contained volatiles (mainly H2o) and had
ffi
fr.'?fJ"i,r#fflri,i?:,i?,ilttr
becausewater depresses
the temperatureof crystaluzatron'
Jumilla,
.Inth.ecoarse-grainedsubvolcanicrocksof
whereilmenite and magnetiteoccur as late phases,
the temperaturehasbeenestimatedusing a Buddington & Lindsley type of geothermometer(Stormer
l983,.pers..comm., Spencer& Lindsley l98l). As
["*i'iil:l5,Tl#?ifr$r1'3ff#::nff*.t*
late-magmatic temperatures.
Fluorine and hydroxyl distribution betweenapatite and coexistingmica is a potential geothermometer. Stormer & Carmichael (1971) obtained an
approximatetemperature-compositionrelation using
fluorides and magnesiumand calcium hydroxides.
73
PETROGENESISOF SPANISH LAMPROITIC ROCKS
More recently Chernyshevael al. (1976)and Ludington (1978)'obtainedmore reliable equationsfor the
equilibria:
TEMPERATURE-OXYCNN
FUCACTTV PATHS
The state of oxidation of the iron in the silicate
melts dependson/(O), temperature,pressure,and
(l) composition(structure)of the liquid. Assumingthe
biotite(oH) + HF : biorite(F) + H2O
other parametersto be constant, the Fe3+/Fe2*
aparire(oH) * HF = apatite(F) + H2O
A) ratio increaseswith decreasingtemperature (Sack el
ql. L980,Kilinc et al, 1983, and referencestherein)
and decreasingpressureMysen & Virgo 1978).The
apatite(oH) + biotite(F) : apatite(F) +
(3) relationship betweenFe3*/Fe2* and composition
biotire(oH)
hasbeendiscussedby Douglaset ol. (1965),who conUsing different data relative to reaction (2) and sidereda redox reaction of the type:
assumingthe mica to be phlogopite, Chernysheva
t/r or: Fe3* + V2o2-,
(4)
et al. (1976)and Ludington (1978)obtained equa- Fd* +
tions for the dependenceof loeJ(3) on the reciprocal of absolute temperature l(l3U/T)-0.16 and Theseauthors found that K(4) dependslargely on
1100/T, respectivelyl.Using the equation reported composition. Fe3+/Fe* ratio, oxygen-ionactivity
by Ludington & Munoz (1975),temperaturemay be and K() increasewith the basicity of the melu
correctedfor the influenceof mica componentsother however,it is not possibleto evaluateseparatelythe
than phlogopite. The temperatureevaluation is only role of the individual parameters.More recently,
approximate becauseof (i) the heterogeneousdistri- Yirgo et al. (1981)found that the composition of
bution of F in mica and (ii) the difficulty in calculat- melts and their degreeof polymerization influence
ing correctly the mole fractions of OH and F and the Fe3*/Fd* ratio. Sack et al, (1980)and Kilinc
of the phlogopite, anniteand siderophyllitecompo- et ql. (1983)haveproposedempirical equationslinknents in the mica. The low values obtained ing Fe3+/Fd*, temperature, composition of the
(580-730'C) suggestthat F+ OH exchangebetween silicatemelts and/(O). Theseempirical equations
apatite and mica continued to temperaturesbelow are poorly calibrated for alkalis (alkaline rocks form
only a small part of the database)and yet the comthe solidus.
J,A
a
---NNo-cx
cil
CL
800
1000
1200
T,oc
Frc.4. Paths of oxygenfugacity ver,$rstemperaturefor somerocks investigated.J
Jumilla, A Aljorra, CX Cancarix, CM Las Minas de Hellin, CL Calasparra' The
equationsfor the buffers HM, QFM, WM, IW are after Myers & Eugster(1983),
and for MnH (Mn1-rO - Mn3O) and NNO after Huebner & Sato (1970).
Alogf(Ot=logil"(O) (sample)-loe/(O) (QFM).Thin lines : rocks, thick lines
= buffers. Filled circles refer to data obtained for the coarse-erainedrocks of
Jrrmilla using maguetite-ilmenitepairs.
74
T}IE CANADIAN MINERALOGIST
positional terms for alkalis are very significant. It
is encouragingto note, howgver,that thereis a good
agreementbetween the data obtained using such
equations and wet-chemical measurements of
Fe3*/F&+ ratio in experimentalproducts obtained
in a lamproitic system(Foley 1985).
Becausethe Fe3*/Fe2+valuesofthe melt and of
the crystallizing phasesare related in someway, the
value in the phaseswill also dependon the liquid
composition.This, of course,placeslimits on the use
of the geothermometer-geobarometer
proposedby
Buddington & Lindsley Qge). Moreover, the
ilmenite-magnetite pair is found only rarely in these
lamproitic rocks. Consequently,we relied mainly on
the empirical equation proposed by Kilinc et al,
(1983) to determine potential temperature-/(O2)
paths for the rocks investigated.
The Fe3*,/F*+ ratio of the melt may be calculated approximately using the value of KB#J"'reported by Foley (1985)and the data for Cr spinel
in the Spanishrocks. Consequently,.f(O) may be
calculated at constant F€+ /F&+ rado for different
temperaturesusing the empirical equation proposed
by Kilinc et ol. (1983):
ln[X@e2o)/X(Feo)] :
olnf(O) + b/T + c + Ddixi
(5)
wherea, b and c are constants,d are constantsfor
different oxide componentsin the melt from which
the Cr spinel crystallizes,and X are the mole fractions. The temperature-fo) paths obtained are
Mst{CP)0
IL
a't
OH
SP
.y
fH
reported in Figure 4. Samples from Calasparra
exhibit the lowest /(O) trend (close to the MW
buffer) and the rocks from Aljorra and Jumilla, the
highest /(O). The redox conditions of the Las
Minas de Hellin, Cancarix and Fortuna materialsare
similar to those of West Kimberley and Gaussberg
lamproite (Foley 1985). The temperature-/(Or)
paths reported are strictly correct only if the following assumptionsare fulfilled: (i) the distribution
coefficient for Cr spinel is not significantly dependent on temperaturein the range 1000-1300"C;(ii)
the redox state of Cr spinel in olivine did not change
significantly after crystallization; and (iii) extrapolation of equation(5) from FMQ to MW conditions
does not induce seriouserrors.
The rocks from Jumilla arethe only samplescontaining Ti magnetite(besidesCr spinel) and iknenite,
which crystallized during the intermediate to late
stages under subvolcanic conditions. The
temperature-/(Ot pattern obtained using a Buddington & Lindsley type of geothermometergeobarometer(Stormer 1983,pers. cotrun.,Spencer
& Lindsley 1981),is reported in Figure 4; f(O) is
in the range 10.14.8
to lQ-t2.eats1.
TABLE5. REPRESEN1ATIVE
DATAONMICA ANDAPATITEANDON
HF, HCl, Hro FUGACITY
MTI05
Localiti.es
!&g(1)
x (F )
x (0 H )
x(Phl)
x (Ann)
X( Sld+East)
loq(fHF/fH20)
Jum!I la
(i)
Puebia de Mula
(it)
0.240
0.758
0.764
0.'103
0.240
0.716
0.764
0.132
-4.1
(i)
(ri)
0.132
0.866
0.732
0.073
0.'t59
-4.0
0.132
0.843
0.730
0.060
0.159
-4.0
a ^ o +| + . ( 2 )
e
x(F)
x ( c )l
x(oH)
0.875
0.02r
0.t04
log(fHF/fH20)
los (fHcl /fHZo)
+cx
cilo'
no-
CL
a7
.2
.4
.6
Ftc. 5. Mg# [ = Mg x 100/(Mg+ Fd+)] for clinopyroxene
versw Fe3+lFe ratio in spinel and ilmenite. Symbols
asin Figure 4 with the addition of F Fortuna; Sp spinel;
IL ilmenite; open circle: Cr-spinel in olivine; square:
Cr spinel not included in olivine; filled circle: titanian
magnetite.
-4.3
-4.9
-4.3
U./JJ
0.043
0.2Q4
-4.0
-4.0
-? A
-4.0
Apatlte/mlca
Toa
611
605
781
749
Fugacltyrat,ioscalculated (1) accordlngto Ludtngton&
M u n o (z1 9 7 5 )a n d ( 2 ) d c c o r d i n tgo K o r z h i n s k i(yt g g t ) .
Temperatures
calculatedaccordingto Ludtngton(1979).
X - molefraction. lhe nole fractions havebeencalculated in the following way:
a. Mica. (i) 0n the basls of a formulawhereall iron
as beenconsideredas divalent, valences. 22 and(F+
+ C l + o H=) 2 . ( i i ) F e 3 r =0 . 1 8 7F e t o t . ( t . e . , F e r 0 r / F e 0 .
= 0 . ? 6 , c f . , F u s t eert a l . 1 9 6 7 )a n d0 H. (2-F-Cl-0)!,her€
o . Fe3+'
b. Apatite. 0n the basis of ( F + C l + 0 H )= 1 .
PETROGENESISOF SPANISH LAMPROITIC ROCKS
RrlartoNs BrrwssN Oxrpr,s ANDSILIcATEPrusrs
Comparable oxygen fugacity - temperaturepaths
were obtained using Cr spinel included in olivine,
and Cr spinel occurring as isolated crystals; the
resultssuggestthat the spinelsare of samegeneration and that their Fe3+/Fe2+ ratio was not
influenced greatly by olivine-spinel subsolidusreequilibration. The Fo content of olivine does not
increase with /(O) in the rocks investigated;
instead, it decreasesroughly with an increasing
Fe3+/Fd+ ratio in the coexistingCr spinel. This
suggeststhat temperatureand melt compositionwere
probably more important in controlling the Fo content of olivine than /(OJ. In contrast, the composition of clinopyroxene, which began to crystallize
later than olivine, was strongly influenced by the
value of/(O), as suggestedby the good correlation
between its Mg# and the Fe3*/Fd* ratio of the
spinel(Fig. 5) This correlation also suggests(i) that
the composition of the Cr spinel reflectsthe redox
conditions when clinopyroxene crystallized, and (ii)
that the oxidation state of the Cr spinel did not
changesignificantly during the latest stagesof crystallization and at subsolidusconditions.
Acttvtttss
or VolarnB
Con4poNeNrs
ratios
The /(HF)f(H2o)
and l(HCl)/f(HzO)
may be calculatedfrom HF+H2O and HCI+H2O
exchange equilibria involving mica and apatite.
Ludington & Munoz (1975)and Korzhinskiy (1981)
reported useful equationsdeterminedexperimentally
assumingideal solution betweenthe componentsof
the solid phaseand ideal mixing of the components
in the gas phase.The data obtained (Table 5) are
approximate for the s€rmereasonswe have pointed
out in the section concerningthe temperafureevaluation. In spiteof the high contentof fluorine in biotite and apatite,the fugacityratios obtainedsuggest
that H2O was dominant over HF.
That CO2 also played somerole during petrogenesis is suggestedby the presenceof magmatic carbonates(Fusteret al. 1967, Pellicer 1973).The carbonatesmay be related to CO2 incorporated from
country rocks (e.9., limestones)after magmasegregation, or to authigenic CO2. In the latter case,
sincethe high-SiO2Spanishlamproitic magmascannot havebeengeneratedin a CO2-richsource,CO2
could be due to oxidation, during emplacementand
crystallization,of CHa presentin a reducedsource
(c/ Foley et al.1986b).
Frerunss oF Tr{E MaNrrB Souncs
AND THE ONTCNI OF THE MEI-TS
Nelson et al. (1986) used lead isotope data and
Foley et al. (1987)usedmainly chemicalfeaturesto
75
infer a depleted mantle source for the Spanish
magmas. However, the high content of "incompatible" elements suggestsenrichment of the
early-depletedmantle source.Sr, Nd and Pb isotopes
are in agreementwith a metasomaticcomponentthat
had the "characteristics of continental crust or sediments derivedfrom the continental crust" (Nelson
et ql. 1986,p.239),Furthermore,the unusuallyhigh
content of uranium in the olivine xenocrystsalso supports mantle metasomatism.The high to very high
K/Na ratio and low Na values,and the high P content, suggest that phlogopite and apatite were
involved during the mantle melting.
That the distribution of trace and somemajor elementsin the most magnesian(MgO > 990)Spanish
lamproitesgivesno evidenceof fractionation from
a singlemagma is suggestedby the following: (i) Me
is not correlatedwith many "incompatible" elements
(e.9., Zr, Nb, Rb); (ii) taking into accountthat olivine is usually the first major mineral which crystallized during the emplacementof the magmas,the
variation of the Ni content (67G-380ppm) in the fineto medium-grainedrocks is too low if comparedwith
the olivine fractionation required (ca. 200/o) to
explainthe largevariation in Mg (ca. l3-9Vo MgO)
(Venturelli et al. 1984); (iii) the 87511865r
ratio is
largely variable (Nelson et al. 1986).
The high Ni (> 380 ppm) and Cr (> 610 ppm)
contents and the high Me# 67-78) in the most
magnesianrocks (MgO > 990) suggestthat they
could be regardedas near-priniary,mantle-derived
magmas.
In a previous study of Spanishlamproites (Venlurelh et al. l9M), a potential loss of Na asNaF during crystallizationwassuggested.Owing to the high
capacity of the alkaline melts to dissolvevolatile
components,the NaF loss could have beensignificant only if F wasvery abundantin the magmas.This
is not supported by any geological and petrographic
evidenceas, for instance,the presenceof products
ofinteraction betweenF-rich fluids and the country
rocks. Thus we believethat at leastthe composition
of the fine- to medium-grainedrocks is not far from
that of the melts before crystallization and nearsurfaceemplacement.
The data of Luth (1967)and unpublisheddata of
A.K. Gupta and D.H. Green,quotedbyFoley et al.
(1986a),for the systemKs-Fo-Qtz (Fig. 6), suggest
that the PHL + EN + FO + L peritecticpoint moves
toward lower Fo./Ks values and higher Qtz content
as pressure decreases(Fig. 6, top), whereasthe
F/H'O ratio affects the position of the peritectic in
a direction transverseto the pressuretrend. A shift
of the EN+FO+L peritecticline away from Qtz
with increasingpressureis also seenin the dry experiments of Kushiro (1980).
Let us considera sourceconsistingof a phlogopitebearingperidotite with a compositionfalling in the
76
THE CANADIAN MINERALOGIST
Ks
Fo
Fta.6. a. Qtz(silica)-Ks(kalsilite)-Fo(olivine)
diagramfor the investigatedrocks with
MgO>990. (l) Fonuna, (2) Las Minas de Hellin (3) Cancarix, (4) Calasparra,
(5) Jumilla (fine-grainedrocks), The coarse-grainedrocks from Jumilla havebeen
omitted becausethe K/Na ratio may havebeensignificantly modified by late crystallization of analcime. Normative data are averagevalues obtained from chemical analysesreportd by Venturelli et al. (1984). phase fields (capital letters) for
warer-rich and for volatile-free system at 28 kbar are after Foley et al. (1986a,
b), basedon unpublisheddata of A.K. Gupta & D.H. Green. D. point A and
point B representthe PHL + FO + EN + L peritecticpoint at 3 kbar (Luth 196?)
and 28 kbar (A.K, Gupta & D.H. Green,unpubl.), respectively,for a water-rich
system.C is the peritectic point for a F-bearingwater-freesystemat 28 kbar (Foley
et al. 1986a).Thus, as pressuredecreases,the peritectic point moves away from
the Fo comer, whereasthe variation of the F/H2O ratio changesthe position of
the peritecticpoint in a direction transverseto the trend for increasingpressure.
otT
16.8
f,6',
1.25
-/
\
13-4
%T;'-';
-
1or
9,1
Pzos
'1.7
.5
.9
1.3
2l
Frc.7 . O/T ratios versasP2O5contentsfor the investigated
rocks with MgO > 990. Symbolsof the localitiesare as
in Figure 6. MgO contents(underlined)are alsoreported
= oxyfor the different outcrops. O/T
gen,z(Si+Ti+Al+Fe3+) atomic ratio calculated
assuminga dry basis and Fe2O3/FeO:0.32, a value
found in the most fresh rocks. Data after Venturelli al
al. (1984). The tie-line connects the samples from
Jumilla (6 and 6'are coarse-grainedrocks).
triangle Phl-En-Fo (Fig. O. The compositionsof the
rocks from Fortuna (l), Las Minas de Hellin (2),
Cancarix (3), Calasparra(4) and Jumilla (5, finegrained rocks) may correspondto peritectic melts
generatedfrom such a sourceunder hydrated conditions at different depths.Assuming that fractional
crystallization was not significant, the Spanish
magmas could have been generated at pressures
lower than 30 kbar.
If we assumethat the sourcedid not contain abundant phlogopite,the amountsof liquid formed near
the peritecticpoint would representa low to moderate degreeof partial melting. Thus the low contents
of Sc (13-15 ppm), Na (0.64-1.90/oNa2O), Al
(8-ll.89o Al2Or) and Ca Q.6-4.90/oCaO) would
reflect a source which, before the enrichment in
"incompatible" elements,underwenta strong deplgtion (oss of clinopyroxeneand garnet).
If, as an alternative, we assumethat the source
was composedmainly of phlogopite and orthopyroxene,and lower amounts of olivine and clinopyroxene, the proportion of melt may be very high; the
7',|
PETROGENESISOF SPANISH LAMPROITIC ROCKS
low contentsof Sc, Na, Al and Ca may reflect a dilution effect rather than a strongly depletednature of
the source.A sourcevery rich in phlogopitemay be
unusual,but is possible.The rarity of sucha source
would be consistent with the scarcity of lamproitic
rocks world-wide. It is unlikely, however,that high
proportions of melt can exist in the mantle without
moving off (McKenzie 1984).The proposal of a high
degreeof melting, therefore, would require fractional
melting followed by coalescence
of the batchesof
melt.
For the outcrops investigated,lhe O/ T ratio (oxygen/tetrahedral sites)andthe MgO contentsare well
correlatedwith PrO, (Fig. 7), suggestingthat apatite solubility decreaseswith increasing degree of
polymerization and with decreasingtemperature, in
agreementwith experimental data (Kogarko et al.
1981).The trend (1)-(2)-(3)-( F(5) (Fie. 7) suggests
that apatite was a potential residual phase, or that
it only just melted during partial melting of the
source.
Granted that the melts were saturated in PrO,
and that apatite in the mantle dissolveseasily into
the melts, the secondmelting hypothesisreported
above requires a large amount of apatite at the
source.The scattereddistribution of Zr, Nb, Ce and
Th in the rocks may be explainedby melting of a
source that had a heterogeneousdistribution of
accessoryminerals.
The Jumilla outcrop is a petrologicaloddity. The
rocks are mainly coarse-grainedand exhibit unusual
chemical features: the MgO content and the O/T
ratio correlate positively not only with P2O5(Frg. 7)
but also with Zr, Nb, Ce and Th (Venturelli el aL
1984).This may be explainedin two different ways:
(i) the rocks formed from a magmaproducedby prograde partial melting of a sourcecontaining residual
minor and accessoryminerals (apatite included); (ii)
the trend was generatedby processesof fractional
crystallization and accumulation involving not only
olivine, but also phasescontaining the high-valency
elementslisted before. For theseaccessoryphaseis,
petrographic evidence suggeststhat only apatite
formed large crystals during crystallization.
Moreover, the observed variation of Ni content
(from 740 to 670ppm) suggeststhat significant fractionation of olivine has not occurred.
REFERENCES
Anrue, A. & Epcen, A.D. (1981): Substitution
mechanismand solubility of titanium in phlogopites
from rocks of probable mantle origin. Contr.
Mineral, PetrologY77, 288'295.
BrrLoN, H., Bonoet, P. & MorrrNar' C. (1983):
Chronologiedu magmatismeN6ogdnedes CorM€ridionale).Bull. Soc.
dilldresBdtiques(Espagne
G6ol.France25, 205-217.
Bonmv,G.D. (1967):Potash-richvolcanicrocksfrom
southernSpain,Mineral. Mag. 36,364-379.
D.H. (19@):IronA,F. & LtNosI-ev,
BunorNcroN,
titanium oxide mineralsand syntheticaquivalents.
J. Petrologlt5, 310-357.
K.D' (1974):The
Cewlront, R.G. & CoLLEnson,
parameters
recalculationof pyroxeneend-member
and the estimationof ferrous and ferric iron content from electron microprobe analyses,Amer.
Mineral. 59' 1203-1208.
Cuemvsurve, Yr.A., Prrnov, L.L. & CuelNYsnnv,
L.V. (1976):Distributionof fluorinebetweencoexisting phlogopiteand apatitein carbonatitesand the
behaviourof fluorine in the carbonatite-forming
processes.
Geochem.Int. 13, 14-22.
Cnoss,W. (189?):Igneousrocks of the LeuciteHills
andPilot Butte,Wyoming.Amer.J. Sci.4'll5-141.
Douclas, R.W., Naru, P. & Peul, A. (1965):Oxygen
ion activity and its influence on the redox
equilibrium in glasses.Phys. Chem, Glasses6,
216-223.
Pacurco, A. (1972):Las
FrnNeroEz,S. & Hnnr.raNoez
rocaslamproiticasde CabezoNegro,Zeneta(Murcia). Estudios Geol. (Madrid) 28,267'276.
ForBv, S.F. (1985):The oxidationstateof lamproitic
magmas. TschermaksMineral. Petrog. Mitt. 34'
2r7-238.
-,
Tevlon, W.R. & Gnenw,D.H. (1986a):The
effectof fluorine on phaserelationshipsin the sysat 28 kbar and the
tem KAlSiOn-Mg2SiOa-SiO2
solution mechanismof fluorine in silicate.melts.
Contr. Mineral. Petrology 93' 4655.
&-
(1986b):The role of fluorine
and oxygen fugaci8 in the genesisof the ultrapotassic rocks. Contr. Mineral. Petrology 94, 183'192.
ACKNOWLEDGEMENTS
Vrrt'unsllt, G., GnmN, D.H. & ToscaNt, L.
(1987): The ultrapotassic rocks: characteristics, clas-
We are grateful to the refereesfor their useful sugsification, and constraints for petrogeneticmodels.
gestionsand to Dr. R.F. Martin for his careful review
Earth Sci. Rev. M,81-134.
of the manuscript. Financial support was from
Ministero della Pubblica Istruzione, Rome (grant n. Fonp, C.8., Russrn, D.G., CnevnN, J.A. & Ftsr,
M.R. (1983): Olivine-liqrlid equilibria: temperature,
027202-A5]320,40t/oM.P.I.).
78
THE CANADIAN MINERALOGIST
pressureand compositiondependence
of the crystal,zliquidcationpartition coefficientsfor Mg, Fe,
Ca and Mn. J. PetrologltU,256-265.
& Muxoz, J.L. (1975): Application of
fluorhydroxyl exchangedatato naturalmicas.Ceol.
Soc.Amer. Abstr. Programs7,lt79.
I
Fusmn, J.M., Gesresr,P., SecRnoo,J. & Frnuos<_r, Lurn, W.C. (1967): Studies in the system
M.L.'(1967): Las rocas lamproiticasde SE de
K4SiOa-Mg2SiO4-SiO2-H2O:
I, Inferred phase
Espafla.EstudiosGeol, (Madrid) ?A,35-69.
relationsand petrologicalapplications.J. petrolog;t
8,372416,
GHronso,
M.S., Carmcuenl, I.S.E., RrveRs,
M.L. &
Sacr, R.O. (1983): The Gibbs free energy of McKeNzrc,D. (1984):The generationand compaction
natural silicateliquids; an expandedregularsoluof partially moltenrock. "I. Petrolog1t25,713-765.
tion approximationfor the calculationof magmatic
intensivevariables. Contr, Mineral. Petrology t4, MuaN, A., Haucr, J. & Lorer4 T. (197.2):r07-t45.
Equilibriumstudieswith a bearingon lunar rocks.
Proc. third Lunar Sci. Conf., 185-196,
HuraNen,J.S. & Saro, M. (1920): The oxygen
fugacity-temperaturerelationships of manganese Mrens, J. & Eucsrrn, H.P. (1983):The systemFeoxideandnickeloxidebuffers.Amer. Mineral. SS,
Si-O: oxygenbuffer calibrationto 1,5@K. Cont.
934-952.
Mineral. Petrology 82, 75-90.
Jaqu'ns,A.L. (1986):F-rich micasin the WestKimber- Mvsru, B.O. & Vrnco, D. (1978):Influenceof presIey lamproites;contrastswith kimberlitesand other
sure, temperatureand bulk compositionon melt
micacdous
alkalineultramaficintrusions.Fourth Int,
structuresin the systemNaAISi3OB-NaFe3+Si2O6.
- Kimberlite Conf. Geot. Soc.Austrqlia 16, 45_47
Amer. J. Sci. 27E, 1307-1322.
(abstr.).
NeLsoN,
D.R., McCwroclr, M.T. & SuN,S.-S.(1986):
Orrgins
of ultrapotassicrocksas inferredfrom Sr,
(1985):
& Forev, S.F.
The origin of Al-rich
Nd and Pb isotopes.Geochim.Cosmochim.Acta
spinelinclusionsin leucitefrom the leucitelam50,23t-245.
proites of WesternAustralia. Arner. Mineral. 70.
I 143-1
150.
NtxoN,P.H., THIRLwILL,
M.F., BucxrEv,F. & Devrs,
C.J. (1984):Spanishand WesternAustralianlamJoxes,A.P. & Surru, J.V. (1983):Petrologicalsigproites:aspectsof wholerock geochemistry..ln
Kimnificanceof mineral chemistryin the Agathla peak
berlites.I: KimberlitesandRelatedRocks.Elsevier.
and the Thumb minettes,Navajovolvanicfield.
Amsterdam.
"/.
Geol. 91, 643-656.
Norrl, F.A., AronmssrN,P.A.M., Itrnrroa, E.H.,
KtLrNc,A., Cenurcnan, I.S.E., Rrvens,M.L. & Sacr,
Pnrclr,H.N.A. & RoNnenr,H.E. (1981):Isotopic
R.O. (1983):The ferric-ferrousratio of naturalsilidating of the post-AlpineNeogenevolcanismin the
cate liquids equilibratedin air. Contr. Mineral.
Betic
Cordilleras,southernSpain. Geol.Mijnbouw
Petrology83, 136-140.
60,209-214.
Kocenro, L.N., Pnrnova, E.N. & Vrvrunnrrr, G.
PELLrcER,
M.J. (1973):Estudiopetrologicoy geoqui(1981):Phosphate
contentof magmaticmelts,
Dokl.
mico
de
un nuevoyacimientode rocaslamproiticas
Akad. Nauk SSSR261. 1430-1432.
situadoen las proximidadesde Aljorra (Murcia).
EstudiosGeol, (Madriil 29.99-106,
KonznrNsrly,M.A. (1981):Apatite solid solsutionsas
indicators of the fugacity of HCI and HF in PnrorR,
R.T. (1939):Somemineralsfrom the leucitehydrothermal flutds. Geochem.Int. 18, 44-60.
rich rocksof the WesternKimberleyarea,Western
Australia.Mineral. Mag. 25, 273-387.
KueHxrn,S.M., Eocan,A.D. & ARlMa,M. (l9Sl):
Petrogenesis
ofthe ultrapotassicrocksfrom the Leucite Hills, Wyoming.Amer. Mineral, 66, 663-677. Puca, E. (1980):Hypoth0sesur le gen€sedes magmatismes calcoalcalins, intra-orog6nique et
Kusnno, I. (1980):Changes
with pressure
postorog6nique
of degreeof
alpins,danslesCordilldresBdtiques.
partial melting and K2O content of liquids in the
Bull. Soc.Geol. France22,243-250.
system Mg2SiO4-KAlSiO4-SiO2.
CarnegieInst.
lil'ash.YearBook 19,267-271.
Sacr, R.O., Canlucuanl, I.S.E., Rrvens,M. &
GHronso,M.S. (1980):Ferric-ferrousequilibriain
Lorrz Rurz, J. & RoonrcuEzBaoroLA,E. (1980):La
natural silicate liquids at I bar. Contr. Mineral.
region volcanicaNeogenadel surestede Espaia.
Petrology75, 369-316.
Estudios Geol. (Madrid) 36, 5-63.
Scorr-SurrH,B.H. & SKrNNen,
E.M.W. (1984):A new
LuorNcroN,S.D. (1978):The biotite-apatitegeotherlook at PrairieCreek,Arkansas.-Jn
Kimberlites.I:
mometerrevisited.Amer. Mineral. 63. 551-553.
Kimberlitesandrelatedrocks.Elsevier,Amsterdam.
PETROGENESISOF SPANISH LAMPROITIC ROCKS
79
SuEnatoN,J.W. & CuNoaru,Al{1980): Leucitesfrom Vrr.ou, D. (1975): Armalcolite-Tiphlogopitediopside-analcite-bearinglamproites from Smoky
Contr. Mineral. Petrologlt
. ' 7Gaussberg,,Antarctica.
1,417427.
Butte, Garfield County,
60, 56G573.
Sreucen,K.J. & Lntoslrv, D.A. (1981):A solution
S., ot BarnslNt, G., CRaw'
model for coexistingiron-titanium oddes. AmeL Verrunnrt, G;, CepEpRt,
S. (1984):
Mineral. 66, I 189-1201.
FoRD,A., Kocanro, L.N. & CsLEsrrNr,
The ultrapotassicrocks from southeasternSpain.
Sronurn, J.C. (1983):The effectsof recalculationon
Lithos ljl,37-54.
estimatesof temperatureand orygen fugaoity from
analysesof multicomponentiron-titanium oxides. Vnco, D., MvsEN,B.O. & SuFnnt,F. (1981):Rela'
Amer. Mineral. 6E, 586-594.
tionship betweenthe oxidation stateof iron and the
structureof the silicatemelts.CarnegieInst. llash.
YearBook E0, 308-310.
& Canurcuaet.,I.S.E. (1971): Fluorinehydroiyl exchangein apatite and biotite: a potential igneous geothermometer. Contr. Mineral. Received,Ianuary6, 1987;rertisedmanuseriptaccepted
Petrologlt3tr, 121-131.
March 27, 1987.
i
I
I