Contributions to

Contrib. Mineral. Petrol. 74, 7-12 (1980) Mineralogy and

Petrology
9 by Springer-Verlag 1980

Melt Density and the Average Composition of Basalt

Edward Stolper l and David Walker 2, 3
i Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA
z Department of Geological Sciences, Harvard University, Cambridge, MA 02138, USA
3 Lunar and Planetary Institute, Houston, TX 77058, USA

Abstract. Densities of residual liquids produced by The essence of the dilemma is summarized well
low pressure fractionation of olivine-rich melts pass by Carmichael et al. (1974), who contends that the
through a m i n i m u m when pyroxene and plagioclase restricted and relatively uniform compositions of con-
joint the crystallization sequence. The observation tinental tholeiites indicate that they are unfractionat-
that erupted basalt compositions cluster around the ed, primary melts of mantle peridotite:
degree of fractionation f r o m picritic liquids corre-
"Continental tholeiite flood basalts and related diabases, more
sponding to the density minimum in the liquid line than any other class of volcanic rocks, satisfy the two criteria
of descent m a y thus suggest that the earth's crust postulated previously for magmas generated directly by fusion -
imposes a density filter on the liquids that pass great volume and compositional homogeneitywithin each province.
through it, favoring the eruption of the light liquids How else can one explain the uniform character of more than
at the density minimum over the eruption of denser 100,000 km3 of Yakima basalts erupted within 5 m.y. in the Colum-
bia River province, northwestern United States? To derive that
more fractionated and less fractionated liquids. magma from picritic basalt of deep-seated origin by low pressure
fractionation (cf. O'Hara 1965, Table 1) requires that again and
again each successive draught of magma must rid itself cleanly,
while still largely liquid, of the same fraction of crystalline olivine
along some identical course of ascent. This seems highly improba-
ble."
Introduction
In this note, we offer what may be part of a solu-
Many petrologists believe that continental, ocean tion to this dilemma. We show that during fractiona-
island, and mid-ocean ridge tholeiites develop by frac- tion of a picritic liquid, or indeed of any liquid crystal-
tionation of perhaps as much as 30 wt.% olivine from lizing only magnesian olivine at its liquidus, the densi-
high-MgO, picritic primary liquids ( O ' H a r a 1968a, ties of residual liquids decrease continuously until
b; Clarke 1970 ; Irvine 1977 ; Elthon 1979 ; Green et al. pyroxene and plagioclase join the crystallization se-
1979). However, it has been difficult to understand quence. After this, residual liquid density increases
in the f r a m e w o r k of this model why most of the with fractionation. Thus, there is a m i n i m u m in the
tholeiites within each petrographic province and from liquid density vs. fractionation trend of picritic com-
province to province appear to have been produced positions. The preferred or average compositions of
by roughly similar amounts of olivine fractionation basalts all over the world coincide with this minimum
from these primary liquids. F o r example, although in density; this m a y suggest that the low density of
they are not uniform in composition, most mid-ocean residual liquids at this stage of the fractionation se-
ridge tholeiites fall in a restricted compositional quence leads to a bias in favor of their eruption rela-
range, with few primitive or very highly fractionated tive to denser more fractionated and less fractionated
samples (Engel et al. 1965; Walker et al. 1979). In liquids. Other factors m a y contribute to the rarity
continental flood basalt provinces, there is a world- of picritic m a g m a s a m o n g erupted samples ; for exam-
wide uniformity in composition that, although not ple, m a g m a chambers beneath the mid-ocean ridges
absolute, has been considered striking by petrologists could intercept rising picritic magmas en route to
for fifty years (e.g., Kennedy 1931 ; Walker and Pol- the surface, preventing eruption of these m a g m a s be-
dervaart 1949; Turner and Verhoogen 1960, Chap. 8). cause they mix with the fractionated m a g m a already

0010-7999/80/0074/0007/$01.20
8 E. Stolper and D. Walker: Melt Density and the Average Composition of Basalt

in the magma chamber (O'Hara 1977). Sparks et al. I I I f

(1980) have independently noted the variation of melt I MORB SUITE I
density with fractionation that forms the basis of this 2.8 [i-t- 6
+7 XS*~
note. They suggest that the dense picritic melts do +s +s x/
not mix with the fractionated magma in the chamber, 2++
I,, I-~.ll...f I X~ X "~ IiX
but rather are trapped beneath it and thus are pre- ~ x
2.7 + +4
3 ,,
vented from erupting.
,,
i
cpx - in (a)
P 2.6
Liquid Density vs. Fractionation (g/cc) ol-in ,\plag-in ilm-in~

Figure 1 shows the densities of silicate liquids at their

liquidi, calculated using the partial molar volumes 2.5
9 1 aim. e~perimenls(Bender et el.)
given by Bottinga and Weill (1970), following the x 1 elm. experiments(Walker et el.)
dry liquid lines of descent of a picritic mid-ocean ridge *
+
residual glasses(Shibata el el.)
MORE~,picritic b~salts, mod~l parenl, liquids
\
tholeiite parent liquid and of a basanite from Tahiti. 2~
The mid-ocean ridge tholeiite line of descent was com-
piled from several sources: Picritic compositions are 2.9 [ I i
I TAHITIAN SUITE I
based on actual rock compositions (Clarke 1970; E1-
thon 1979) and on model parent liquid compositions
, X ~ II
(Irvine 1977; Elthon 1979). The liquidus temperatures 2.8 ,_lh~.....
of these compositions were calculated using Equa-
tions 4 and 6 in Roeder (1974). The liquid line of
descent for Fe/(Fe+Mg) [molar]=0.3-0.8 is based
2.7 i,- 'i
on glass compositions from 1 atm melting experi- i

ments on mid-ocean ridge tholeiites (Bender et al.

1978 ; Walker et al. 1979). Liquidus temperatures were (g/cc) , , , (b)
ol-in ix ilm-i
taken from the conditions of the experiments. The
plag-!n \
highly fractionated end of the fractionation sequence CpX-
Ill. X ~
is based on interstitial glasses in natural mid-ocean
ridge tholeiites (Shibata et al. 1979); the liquidus tem- 2.5 9 1 atm. experiments \
perature was assumed to be 1,050~ C. The Tahitian (Stolper and Tracy) \
basanite line of descent is based on lava compositions • lavos(McBirney and Aoki) x \
(McBirney and Aoki 1968) and on glass compositions -x, residual gfass(Tracy) \
2.4 = L L I
from 1 atm melting experiments on a nodule-bearing 92 .4 .6 .8 1.0
basanite (Stolper and Tracy, unpublished); the liqui- Fe/(Fe + Mg)rnol
dus temperature at each stage in the sequence was Fig. 1 a and b. Liquid density versus fractionation [Fe/(Fe + Mg)
calculated from the results of Roeder (1974) or taken (molar)] for liquid lines of descent of (a) picritic tholeiites and
from the temperatures of the experiments. One highly mid-ocean ridge basalts (MORB); and (b) a Tahitian basanite.
fractionated composition is based on an analysis of Densities were calculated using the method of Bottinga and Weill
(1970). All Fe was taken as FeO. Liquidus temperatures used in
interstitial glass in the basanite (R.J. Tracy, personal the calculations are explained in the text. Vertical dashed lines
communication) with an assumed liquidus tempera- show entry and departure of various phases in the crystallization
ture of 1,050~ C. sequences, ol, olivine; cpx, diopsidic clinopyroxene; plag, plagio-
Both liquid lines of descent display the same fea- clase; ilm, ilmenite, a Numbered + ' s correspond to the following
tures. The liquid densities decrease when olivine is compositions: 1 hypothetical primitive mid-ocean ridge picrite (Ir-
vine 1977); 2 hypothetical average oceanic crust (Elthon 1979);
the only crystallizing phase. When pyroxene and pla- 3 average F A M O U S basalt (Bryan and Moore 1977); r average
gioclase join the crystallization sequence, the liquid oceanic tholeiite glass (Melson et al. 1976); 5 high-MgO dikes from
density increases, due primarily to the FeO enrich- Tortuga (Elthon 1979); 6 Svartenhuk picrite (Clarke 1970) ; 7 Baffin
ment at this stage of fractionation and to the removal Bay picrite (Clarke 1970). Filled circles are glass compositions from
1 atm crystallization experiments on a primitive F A M O U S glass
of "light" plagioclase components from the liquid. (Bender et al. 1978); x's are glass compositions from 1 atm melting
Density again begins to decrease when ilmenite begins experiments on an Oceanographer Fracture Zone basalt (Walker
to crystallize, depleting the residual liquids rapidly et al. 1979); *'s are residual glasses from natural crystallization
in FeO. The buildup of SiOz and the other "light" of Oceanographer Fracture Zone basalts (Shibata et al. 1979).
oxides as the residual liquids move towards petro- b Filled circles are glass compositions from 1 atm melting experi-
ments on a Tahitian basanite (Stolper and Tracy, unpublished);
geny's residua system late in the crystallization se- x's are Tahitian lava compositions (McBirney and Aoki 1968);
quences also contributes to a continuous decrease in * is interstitial glass in Tahitian basanite (R.J. Tracy, personal
density at this stage of fractionation. Addition of communication)
E. Stolper and D. Walker: Melt Density and the Average Composition of Basalt 9

rocks with SiO2 between 45 and 54wt.% SiO2,

2.8 H20<4wt.%, CO2+SO3<3wt.%, and FeO/
0 (FeO + Fe203) < 0.5 in each of these environments are
(g/co) plotted.
--12.7 Although basalts are multicomponent systems and
we have only shown one compositional variable in
Fig. 2, it nevertheless appears that mid-ocean ridge
and oceanic island basalt compositions worldwide
MID-OCEAN
RIDGE cluster preferentially around the density minimum in
the liquid line of descent from picritic magmas. The
histogram for continental basalts has two maxima.
One at Fe/(Fe + Mg) ~ 0.45 corresponds closely to the
density minimum; the other, at Fe/(Fe+Mg)~0.58
is displaced from the density minimum but still lies
tu
within the density "well". We note that the precise
location of the density minimum will differ from rock
< type to rock type, reflecting differences in bulk com-
position, including Fe/(Fe + Mg) ratio, crystallization
u_

O sequence, and liquid line of descent of the primitive

liquids. This is illustrated in Fig. 1 by the differences
u~
in the shapes of the curves and the locations of the
density minima for the MORB and Tahitian suites.
7" The greater complexity of the histogram for conti-
nental basalts shown in Fig. 2 and the displacement
of the histogram from the density minimum may thus
reflect the diversity of rock types that have been
lumped together in the continental basalt histogram
and the fact that the density vs. fractionation curve
shown in Fig. 2 is for MORB rather than continental
0 0.5 1
basaltic compositions.
Fe/( Fe + Mg)Mo t
Fig. 2. Comparison of density versus Fe/(Fe+Mg) ratio in mid-
ocean ridge tholeiites and histograms showing the Fe/(Fe+Mg) Discussion
ratios of basalts from three geological environments. The density
vs. Fe/(Fe + Mg) ratio trend is from Fig. I a. Histograms were com-
piled from the data base of Mutschler et al. (1976) using the filters Figure 2 shows that if primary magmas are in general
described in the text. Total numbers of samples plotted are: mid- picritic, then the degree of fractionation from olivine-
ocean ridge (2004), ocean island (2970), continental (3774). Note rich parent magmas reached by most basalts prior
the correspondence between the Fe/(Fe-Mg) ratio (a function of
the degree of fractionation) at the density minimum and at the
to their eruption is not random, but corresponds clo-
maxima in the histograms sely to a local minimum in the density of residual
liquids as a function of fractionation. We propose
that this density minimum may be a factor in control-
water to the system would lower the density vs. frac- ling which basalt compositions will erupt and in ex-
tionation curves shown in Fig. 1 but would have little plaining why, if primary magmas are picritic, most
effect on their shapes, assuming that the amount of observed basalts are produced by similar amounts
water is sufficiently small that it behaves essentially of fractionation from these high-MgO parent liquids.
as an incompatible species and has little influence The compositions at the density minimum are also,
on the phase relations (Sparks et al. 1980). for the most part, approximately saturated with oli-
vine+diopsidic clinopyroxene+plagioclase at their
1 atm liquidi. Some crustal rocks, both in continents
Basalt Compositions and the upper levels of oceanic crust, have densities
near the 2.7 2.8 g/cc range of liquid densities between
Figure 2 shows, superimposed on the trend of density picrites and the density minimum (Daly et al. 1966;
vs. fractionation in a picritic mid-ocean ridge tho- Salisbury and Christensen 1978). Lighter liquids
leiite, histograms of numbers of analyzed basalts in would typically rise more readily through crustal
continental, ocean island, and mid-ocean ridge envi- rocks than would denser liquids. Thus, both primary
ronments. These histograms were compiled from the liquids and liquids that fractionate past the density
PETROS data file (Mutschler et al. 1976). All igneous minimum would be less likely to erupt than liquids
10 E. Stolper and D. Walker: Melt Density and the Average Composition of Basalt

2.9 in preventing the eruption of) primitive magma com-

positions. In this note, we support this emphasis while
adding to it the potential implications of a minimum
2.8 in the density versus fractionation trend of basic liq-
LIQUID uids in biasing the compositional range of erupted
p ~ / DENSIT'
basalts. We recognize, of course, that factors other
(g/cc) CRUSTAL
2.7 "~OFWINDCW ,,,el- DENSITY
than the local density contrast between melt and
country rocks also influence whether or not a melt
will rise through the crust and erupt. Volatile content
ERUPTB L TY"
of the melt, melt viscosity, volatile overpressure in
2.6 i I I I the magma chamber, and the widths of the conduits
0 .2 .4 .6 .8 through which the melt must flow are examples of
F e / ( Fe + Mg)Mo L other important variables (G.P.L. Walker 1974).
Also, densities of magmas need not necessarily corre-
Fig. 3. Schematic picture showing how low density crustal rocks
could act as a density filter preventing eruption of primitive and spond to densities of liquids. Thus, for example, mag-
highly fractionated basaltic melts, biasing the range of observed mas laden with olivine or pyroxene phenocrysts will
basalts to near the degree of fractionation from primary magmas be denser than the liquid component of the magma.
corresponding to the density minimum Even though the liquid component of a magma may
be lighter than the overlying crust, the magma may
not be and only after the crystals settle from the liquid
at the density minimum. This effect would be en- willthe liquid be able to pass through the overlying crust.
hanced if the average crustal density were near 2.70- We also note that if the melt is present in continuous
2.75 g/cc (Fig. 3). In this case, primary magma from conduits down to its source region (or if the magma
the mantle, denser than the crustal rocks, would be chamber is at some depth below the crust) it is the
trapped below or within the crust. As the trapped integrated density contrast, rather than the local den-
magma cools and crystallizes olivine, the density of sity contrast (i.e., the hydrostatic head), that controls
the residual liquid decreases. When the density dips whether melt will rise relative to its surroundings and
below that of the crustal rocks, the liquid is able that in this case, melt could rise through lighter
to move to the surface and erupt. Any magma that country rocks.
does not erupt will continue to cool and fractionate Sparks et al. (1980) have noted another way in
in crustal intrusions; when pyroxene and plagioclase which the density minimum shown in Fig. 1 could
begin to crystallize, the density of the residual liquid account for the rarity of picritic lavas. They suggest
will increase and eventually exceed that of the enclos- that in mid-ocean ridge environments, fractionated
ing crustal rocks and again be prohibited from erupt- magma in the sub-ridge magma chamber with a den-
ing. After considerable further fractionation, the re- sity near the density minimum provides a density cap
sidual liquid density will again drop below that of under which primitive picritic liquid spreads, without
the crust, but by this point, the volume of liquid substantial mixing, when it enters the magma
available for eruption is small and the viscosity of chamber. Only after crystallizing substantial amounts
residual liquids will have increased by almost an order of olivine is this deep layer of liquid able to mix
of magnitude over its value at the density minimum with the overlying magma and become available for
(from 400 to 3,000 poise; based on the calculation eruption. There are thus at least three possible expla-
scheme of Bottinga and Weill 1972). These factors nations of the rarity of picritic lavas in oceanic regions
could account for the relative rarity of the eruption and of the clustering of mid-ocean ridge basalt com-
of lavas corresponding to these intermediate to high positions, assuming that essentially unfractionated
degrees of fractionation. Eruption of "late-stage" picritic liquid is delivered to the crust and is the parent
rhyolites could be facilitated by a viscosity decrease to mid-ocean ridge tholeiites: (1)Ascending picritic
late in the fractionation sequence due to a buildup liquid could be intercepted by and mixed into a so-
of volatiles in the fractionated liquids or to fusion called steady state magma chamber (O'Hara 1977).
processes favoring rhyolitic melt compositions (Yoder (2) Picritic liquid could be intercepted by a magma
1973; Chayes 1977). chamber containing fractionated liquid and due to
The concept of the crust as a density filter is not its greater density, spread beneath the fractionated
new (e.g., G.P.L. Walker 1974; Neumann 1980). Ear- liquid with minimal mixing (Sparks et al. 1980). Or
lier studies have emphasized the decrease in density (3), the glassy, altered, fractured, and vesicular rocks
expected during differentiation and the importance forming the roof of the magma chamber could provide
of low density crustal rocks in filtering out (that is, a density trap for picritic liquid in the chamber. These
E. Stolper and D. Walker: Melt Density and the Average Composition of Basalt 11

three e x p l a n a t i o n s are n o t m u t u a l l y exclusive a n d all OCE 79-09699, EAR 79-06321, EAR 79-23977, NASA grant
could p l a y roles in m i d - o c e a n ridge systems. NGL-22-007-247, and the Committee on Experimental Geology
and Geophysics of Harvard University. Contribution
I f picritic liquids are also p a r e n t a l to c o n t i n e n t a l Number 3392, Division of Geological and Planetary Sciences, Cali-
basalts, then the sialic crust w o u l d p r o v i d e a low fornia Institute of Technology, Pasadena, CA 9 1125. A portion
d e n s i t y b a r r i e r to the rise o f these melts until their of this research was done while D. Walker was a visiting Scientist
densities decrease sufficiently b y olivine f r a c t i o n a t i o n at the Lunar and Planetary Institute which is operated by the
Universities Space Research Association under Contract No.
to a l l o w t h e m to rise to the surface. This w o u l d i m p l y
NASW-3389 with NASA. This paper is Lunar and Planetary Insti-
t h a t relatively large a m o u n t s of c u m u l u s olivine c o u l d tute Contribution No. 411.
be l o c a t e d in the crust o r at the base o f the crust.
This m i g h t also i m p l y t h a t b a s i c i n t r u s i o n s are c o m - References
m o n d e e p in the crust o r at the c r u s t - m a n t l e b o u n d -
ary. The p r e f e r e n t i a l c o m p r e s s i o n o f b a s a l t i c liquid Bender JF, Hodges FN, Bence AE (1978) Petrogenesis of basalts
from the Project FAMOUS area: Experimental study from
relative to crystalline silicates w o u l d raise the d e n s i t y 0 to 15 kbars. Earth Planet Sci Lett 41:27%302
o f the picritic p a r e n t a l liquids relative to c o n t i n e n t a l Bryan WB, Moore JG (1977) Compositional variations of young
crustal r o c k s at deep crustal levels ( S t o l p e r et al. 1980), basalts in the Mid-Atlantic-Ridgerift valley near 36~ ' N. Geol
increasing the l i k e l i h o o d t h a t the p a r e n t a l liquids Soc Am Bull 88:556-570
Bottinga Y, Weill DF (1970) Densities of liquid silicate systems
w o u l d be t r a p p e d d e e p in the c o n t i n e n t a l crust until
calculated from partial molar volumes of oxide components.
f r a c t i o n a t i o n l o w e r e d the d e n s i t y sufficiently for m e l t Am J Sci 269:169-182
rise to continue. Bottinga Y, Weill DF (1972) The viscosity of magmatic silicate
K o m a t i i t e s are e x t r e m e l y dense (2.82-2.84 g/cc) liquids: A model for calculation. Am J Sci 272:438 475
at their liquidi. These w o u l d likely be t r a p p e d b e l o w Carmichael ISE, Turner F J, Verhoogen J (1974) Igneous Petrology,
McGraw-Hill, p 739
light crust u n d e r m o s t c i r c u m s t a n c e s . This c o u l d con- Chayes F (1977) The oceanic basalt-trachyte relation in general
t r i b u t e to the r a r i t y o f their e r u p t i o n . T h e y w o u l d and in the Canary Islands. Am Mineral 62:666-671
t y p i c a l l y be t r a p p e d within the crust o r n e a r its base, Clarke DB (1970) Tertiary basalts of Baffin Bay: Possible primary
p r e s u m a b l y f r a c t i o n a t i n g t o w a r d tholeiitic r e s i d u a l magma from the mantle. Contrib Mineral Petrol 25:203-224
Daly RA, Manger GE, Clark SP Jr (1966) Density of rocks. Geol
liquids t h a t w o u l d be light e n o u g h to m o v e t o w a r d
Soc Am Mere 97:19 26
the surface. Elthon D (1979) High magnesia liquids as the parental magma
for ocean floor basalts. Nature 278:514 518
EngeI AEJ, Engel CG, Havens RG (1965) Chemical characteristics
Conclusions of oceanic basalts and the upper mantle. Geol Soc Am Bull
76:719-734
W e p r o p o s e t h a t the o b s e r v a t i o n t h a t the c o m p o s i - Green DH, Hibberson WO, Jaques AL (1979) Petrogenesis of
tions o f m o s t e r u p t e d b a s a l t s can be p r o d u c e d b y mid-ocean ridge basalts. In: MW McElhinny (ed) The Earth:
r o u g h l y similar a m o u n t s o f f r a c t i o n a t i o n o f olivine Its Origin, Structure and Evolution. Academic Press, London
Irvine TN (1977) Definition of primitive liquid compositions for
f r o m picritic liquids m a y be related to the fact t h a t basic magmas. Carnegie Inst Washington Yearb 76:454-461
the densities of residual liquids p r o d u c e d b y f r a c t i o n a - Kennedy WQ (1931) The parent magma of the British Tertiary
tion of olivine-rich m a g m a s pass t h r o u g h a m i n i m u m Province, Summary of Progress for 1930, Pt 2. Mem Geol Surv,
when p y r o x e n e a n d p l a g i o c l a s e j o i n the c r y s t a l l i z a t i o n pp 61-73
sequence. The light liquids at this d e n s i t y m i n i m u m McBirney AR, Aoki K (1968) Petrology of the island of Tahiti.
GeoI Soc Am Mem 116:523-556
can be expected to pass t h r o u g h the crust a n d r e a c h Melson WG, Vallier TL, Wright TL, Byerly G, Nelson J (1976)
the surface m o r e easily t h a n the d e n s e r m o r e f r a c t i o n - Chemical diversity of abyssal volcanic glass erupted along Pacif-
ated a n d less f r a c t i o n a t e d liquids. A l t h o u g h factors ic, Atlantic, and Indian Ocean sea-floor spreading centers. In:
o t h e r t h a n the d e n s i t y c o n t r a s t b e t w e e n m e l t a n d The Geophysics of the Pacific Ocean Basin and Its Margin
(AGU, Washington), p 351-367
c o u n t r y rocks are involved in the m o v e m e n t o f m a g -
Mutschler FE, Rougon DJ, Lavin OP (1976) PETROS a data
m a to the surface, the o b s e r v a t i o n t h a t e r u p t e d b a s a l t bank of major-element chemical analyses of igneous rocks for
c o m p o s i t i o n s cluster a r o u n d the degree o f f r a c t i o n a - research and teaching. Comput Geosci 2:51-57
tion c o r r e s p o n d i n g to the d e n s i t y m i n i m u m in the Neumann E-R (1980) Petrogenesis of the Oslo region larvikites
liquid line o f descent suggests t h a t the e a r t h ' s crust and associated rocks. J Petrol (in press)
O'Hara MJ (1965) Primary magmas and the origin of basalts.
m a y indeed i m p o s e a d e n s i t y filter, albeit an i n c o m - Scot J Geol 1 : 19~t0
pletely efficient one, on the liquids t h a t pass t h r o u g h O'Hara MJ (1968 a) Are any ocean floor basalts primary magma?
it. Nature 220:683 686
O'Hara MJ (1968b) The bearing of phase equilibria studies in
Acknowledgements. Soon after the preparation of this manuscript, synthetic and natural systems on the origin and evolution of
we discovered that Sparks and his coworkers had independently basic and ultrabasic rocks. Earth Sci Rev 4:69 133
noted the density vs, fractionation trend discussed in this manu- O'Hara MJ (1977) Geochemical evolution during fractional crystal-
script. We thank them for the opportunity to see their manuscript lization of a periodically refilled magma chamber. Nature
prior to publication. This work was supported by NSF grants 266 : 503-507
12 E. Stolper and D. Walker: Melt Density and the Average Composition of Basalt

Roeder PL (1974) Activity of iron and olivine solubility in basaltic Turner FJ, Verhoogen J (1960) Igneous and Metamorphic Petro-
liquids. Earth Planet Sci Lett 23:397~410 logy. McGraw-Hill, p 694
Salisbury MH, Christensen NI (1978) The seismic velocity structure Walker F, Poldervaart A (1949) Karoo dolerites of the Union
of a traverse through the Bay of Islands ophiolite complex, of South Africa. Geol Soe Am Bull 60 : 591-706
Newfoundland, an exposure of oceanic crust and upper mantle. Walker GPL (1974) Eruptive mechanisms in Iceland. In: L
J Geophys Res 83:805-817 Kristjansson, D Reidel (eds) Geodynamics of Iceland and the
Shibata T, DeLong SE, Walker D (1979) Abyssal tholeiites from North Atlantic area. pp 189 201
the Oceanographer Fracture Zone. I. Petrology and fractiona- Walker D, Shibata T, DeLong SE (1979) Abyssal tholeiities from
tion. Contrib Mineral Petrol 70:89 102 the Oceanographer Fracture Zone. II. Phase equilibria and mix-
Sparks RSJ, Meyer P, Sigurdsson H (1980) Density variation ing. Contrib Mineral Petrol 70:111-125
amongst mid-ocean ridge basalts : implications for magma mix- Yoder HS Jr (1973) Contemporaneous basaltic and rhyolitic mag-
ing and the scarcity of primitive lavas. Earth Planet Sci Lett mas. Am Mineral 58:153-171
46:419-430
Stolper E, Walker D, Hays JF (1980) Melt segregation from partial-
ly molten source regions: the importance of melt density and
source region size. J Geophys Res (submitted) Received January 31, 1980; Accepted June 24, 1980