Iron-rich lherzolitic xenoliths from Oahu: origin and implications for Hawaiian magma sources

Iron-rich lherzolitic xenoliths from Oahu: origin and implications for Hawaiian magma sources

Earth and Planetary Science Letters, 102 (1991) 45-57 45 Elsevier Science Publishers B.V., Amsterdam [CL] Iron-rich lherzolitic xenoliths from Oahu...

1MB Sizes 5 Downloads 34 Views

Recommend Documents

No documents
Earth and Planetary Science Letters, 102 (1991) 45-57

45

Elsevier Science Publishers B.V., Amsterdam [CL]

Iron-rich lherzolitic xenoliths from Oahu: origin and implications for Hawaiian magma sources G a u t a m Sen

a

a n d W i l l i a m P. L e e m a n b

a Department of Geology, Florida International University, Miami, FL 33199, USA I, Keith-Weiss Geological Laboratories, Rice Unioersity, Houston, T X 77251, USA

Received March 14, 1990; revised version accepted July 31, 1990

ABSTRACT Petrographic studies and mineral chemical analyses support the hypothesis that garnet lherzolites and websterites from Oahu, Hawaii were produced by mechanical disintegration (stoping), chemical alteration (metasomatism) and mechanical mixing of spinel lherzolite wall rock (i.e., lithosphere) with intrusive clinopyroxenite veins. This conclusion is inconsistent with two prevalent hypotheses concerning the origin of these rocks, one of which proposes that the garnet lherzolites represent the upper mantle source for Hawaiian tholeiitic magmas, and the other that invokes origin of these xenolith lithologies by high-pressure crystallization of Hawaiian alkalic magmas. Although the undepleted mantle source of Hawaiian tholeiitic magmas may be chemically similar to some of the garnet lherzolite xenoliths, it is physically impossible for these xenoliths to represent actual source rocks.

1. Introduction

composition of the shield-building tholeiitic magmas.

The nature of the suboceanic mantle is largely understood from petrologic studies of its partial melts, as represented by oceanic island (e.g., Hawaii) and ocean floor basalts. Geochemical studies of such rocks indicate that the upper mantle is heterogeneous (e.g., [1-3,7]), but the volume scale of melt formation and extraction is p r o b a b l y large enough so that they provide only limited spatial resolution of such heterogeneities. Fragments of the uppermost mantle, included as xenoliths in alkalic basaltic magmas, provide direct and c o m p l e m e n t a r y constraints on smaller-scale variations in the lithosphere as well as an insight into interaction between ascending m a g m a s and the lithosphere. Here we describe a group of mantle xenoliths from the island of O a h u and emphasize textural and mineral-compositional evidence for c o m p l e x m a g m a - l i t h o s p h e r e i n t e r a c t i o n processes beneath Oahu. We also discuss their relevance to current petrologic models for the origins of these xenoliths and the source mantle 0012-821X/91/$03.50

© 1991 - Elsevier Science Publishers B.V.

2. Background Isotopic and trace element chemical studies of Hawaiian lavas have shown that at least two (and possibly more) isotopically distinct e n d - m e m b e r mantle source c o m p o n e n t s must be participant in generating the magmas, one of which is a primitive plume and the other is the lithosphere (see reviews in [2,3]). However, varied interpretations of analyses of Hawaiian tholeiitic basalts have led to controversy over the major element composition and mineralogy of mantle sources of these magmas. O n one hand, if olivine-rich to picritic tholeiites (_> 15% M g O ) are considered as ' p r i m a r y m a g m a s ' (cf. [4]), their high M g - n u m b e r s ( = 100 x M g / ( M g + Fe)) imply a relatively refractory source ( M g - n u m b e r > 90) similar to depleted garnet or spinel lherzolites c o m m o n l y f o u n d in kimberlites and alkalic basalts worldwide. Alter-

46 natively, if entrainment of accumulated olivine from tholeiitic magma chambers explains the high and varied olivine concentration in some lavas (cf. [5-7]), then the primary liquids could resemble aphyric or sparsely porphyritic basalts with lower MgO contents (ca. 7-10%); such liquids could have equilibrated with relatively fertile, iron-rich mantle (Mg-number ca. 85; [4,16,35,36]). This latter view is weakened by the occurrence of relatively Fo-rich olivine [Fo86_89] as phenocrysts and cumulates in Hawaiian tholeiitic magmas [22,34]; the parental magmas of these lavas must have equilibrated with olivines at least this magnesian. At any rate, as it is difficult to discern how much olivine fractionation (accumulation) occurred in most Hawaiian tholeiites, the choice of a parental melt composition remains subjective. Past studies of Hawaiian xenoliths have added further to the controversy of an Fe-rich vs. Mg-rich mantle source for Hawaiian basalts [2,8]. In particular, relatively Fe-rich garnet bearing websteritic to lherzolitic inclusions (termed F E L G here, for Fe-rich lherzolitic group, defined below) from the Salt Lake Crater of Oahu have been interpreted by some workers to represent an unusually Fe-rich mantle (presumably the hotspot or plume) which partially melts to form the Hawaiian tholeiitic magmas [7,9]. Others have implicitly or explicitly stated that these F E L G xenoliths are olivine-rich lithologic variants of the pyroxenite suite, which represents late vein crystallizations in the lithosphere [8,10-12]. Some of the authors who suggest that the undepleted (i.e., prior to removal of partial melts) Hawaiian upper mantle is relatively Fe-rich with Mg-numbers ( = 100 × M g / ( M g + Fe)) of 84-85 [7,9,13,16], as compared to the more 'normal' (Mg = 87-89) upper mantle depicted in the models of Carter [17] and Ringwood [18], based their inferences largely on these F E L G xenoliths. This small but significant difference in Mg-number has important implications for the temperature at which melting may initiate, the compositions of primary magmas, and details of magma production. To clarify these issues, we have studied some of the F E L G xenoliths which comprise about 2% of the inclusions found in late-stage (i.e., post-erosional) alkalic lavas of the Honolulu Volcanic Series at Salt Lake crater on Oahu [13,19]. F E L G

G . S E N A N D W.P. L E E M A N

OI

Cpx

Fig. l. Modal compositionof FELG xenoliths studied. is generally defined here to include all garnet bearing inclusions with greater than 15% olivine and less than 20% orthopyroxene, i.e., in modal terms they include garnet bearing lherzolites, websterites and wehrlites (Fig. 1). Their constituent minerals (e.g., olivine: Fo_<87) are relatively Fe-rich compared to those of the spinel lherzolites of Hawaii (Fo86_92; most commonly Fo88_91 ; [12]) and garnet lherzolites from continental kimberlites (Fo > 88; [39]) and from Malaita (Fo> 88; [39]). The nomenclature of FELGs has been a subject of some debate: whereas Kuno [26] preferred to call them olivine eclogites, Anderson [35] named at least one of them (i.e., 66SAL-1) a 'piclogite'. In Anderson's mantle model, piclogite forms an important source mantle reservoir from which basalt magmas are extracted. Most other authors describe them as garnet lherzolite or websterite (cf. [2]). 3. Lavas and xenoliths of Oahu

Two shield volcanoes, namely Waianae and Koolau, form the western and eastern parts, respectively, of the island of Oahu [19]. The Koolau shield volcano erupted tholeiitic lavas around 1.82.6 Ma ago [20]. Following a long hiatus, the alkalic lavas of the Honolulu Volcanics erupted about 0.3-0.6 Ma ago, carrying with them a variety of xenoliths from a wide range of depths in the crust and upper mantle [8,13,21]. Three major xenolith lithologies are found in Oahu; namely, dunite, spinel lherzolite and pyroxenite suites [13,21,22,26]. The dunite suite xenoliths contain relatively Fe-rich olivines (Fo82.6_89.7) and chromites whose compositions are analogous to the

IRON-RICH

XENOLITHS:

ORIGIN

AND

IMPLICATIONS

FOR

HAWAIIAN

chromite microphenocrysts in Hawaiian basalts [23]. Therefore, Sen and Presnall [23] concluded that these dunites are cumulates. The spinel lherzolites contain chrome-diopside, magnesian olivine (Fo86_92), spinel, and orthopyroxene whose modal proportions, textures, and mineral compositions are similar to the so-called Group I or Cr-diopside group inclusions found all over the world [2,12]. Based on their trace element compositions, Frey and Roden [2] concluded that the spinel lherzolites represent metasomatically modified lithosphere which originally formed as residues of partial melting related to the generation of mid-ocean ridge basalts (MORB). Trace element and isotopic data ([24,38,41]; M.

MAGMA

47

SOURCES

Tatsumoto, pers. commun., 1989) suggest that significant variation exists in the spinel lherzolites, one end member of which is similar to Koolau and the other is similar to Honolulu Volcanics (also MORB source) in terms of their isotopic (Nd, St, Pb) compositions. Widespread heterogeneity in terms of Na and Cr in the spinel lherzolite suggests the possibility of multiple modes of origin (e.g., variably metasomatized lithosphere, plume) of this xenolith suite [12]. Furthermore, intrusive relationships in composite xenoliths clearly show that spinel lherzolites were wall rocks into which pyroxenite-parent magmas intruded [22]. The pyroxenite suite is composed dominantly iiii¸ ii ¸

~!iiiii~ii¸.... i!i,

Fig. 2. Hand specimen photographs showing olivine clusters in FELG xenoliths. (a) 68SAL-5 This rock is a composite xenolith. The FELG part of it has olivine clusters dispersed in a clinopyroxene-rich groundmass, and the contact between the foliated spinel lherzolite and the FELG is clearly visible. It is interpreted that the olivine clusters are broken remnants of the wall rock lherzolite. The groundmass pyroxenes in the FELG part represent intrusive veins. (b) Olivine + spinel cluster in a pyroxene-rich groundmass in 70SAL-76. The spinel is compositionally akin to those of the spinel lherzolite suite (cf. [12]).

48

G . S E N A N D W.P. L E E M A N

of clinopyroxenites, m a n y of which contain garnet, and appears to have originated due to high-pressure vein crystallization of Hawaiian alkalic magmas [11,22]. These authors concluded that garnetbeating lherzolites, websterites and wehrlites are actually olivine-rich variants of the garnet clinopyroxenite suite and that all have a c o m m o n petrogenesis. However, because whole-rock bulk compositions of the F E L G samples are distinct from both the clinopyroxenites and the spinel lherzolite suite (e.g., [14]), we here reevaluate their petrogenetic significance.

4. Petrography F E L G xenoliths are composed of clinopyroxene, olivine, orthopyroxene, spinel and garnet. The proportions of these minerals vary considerably (Fig. 1). The overall texture of these rocks varies from porphyroclastic to coarse granoblastic. In larger ( - 17 cm) hand specimens, the most striking feature of the F E L G xenoliths is the clustered occurrence of olivine (Fo84_87) grains dispersed in a clinopyroxene-rich groundmass (Fig. 2). In smaller ( < 5 cm) xenoliths, such clustering of olivine grains is not immediately apparent because of the coarse grain size of the constituent minerals. The occurrence of olivine aggregates in F E L G rocks is distinct from the more randomly distributed olivine grains in spinel lherzolites and is considered to be a key to their origin. Kuno's [26, p. 218,219] observations on F E L G xenoliths (his "olivine eclogites") are pertinent in this context. He divided these into two types on the basis of olivine morphology: "one with euhedral olivine and the other with anhedral olivine, although there is some transition between the two habits of olivine". Of particular interest is Kuno's description of his specimen number 14, "which is unusually rich in olivine, and in hand specimen, encloses a mass of lherzolite about 5 cm across". 69SAL-97, a key sample whose composition will be described later, has the morphological characteristics of Kuno's second type of 'olivine eclogites'; in fact, it appears to be very similar to his specimen no. 14. The olivine of F E L G xenoliths has a greenish yellow color reflecting its Fe-rich composition (Fo83_87) relative to the more magnesian (Fo86_92), transparent, light green colored olivines in spinel

Fig. 3. Photomicrograph of an orthopyroxene xenocryst (5 mm long) poikilitically enclosed inside a large clinopyroxene in Smithsonian sample 114923-111B. Note that (1) the outline of the opx grain is irregular, and (2) the fine lamellar (100) exsolutions of cpx in the opx are oriented at a right angle to the exsolvedblebs of opx in the host cpx. The orthopyroxene is interpreted as a broken fragment of the wall rock (spinel lherzolitic lithosphere) which was later enclosed by a single clinopyroxene that crystallized from an intrusive pyroxehite vein. The annealed fracture, marked by fluid inclusion trail, cuts through both the opx and the host cpx. Such fracturing must have occurred during cooling and emplacement of the last remaining fluids along a cooling crack.

lherzolites. Microscopic examination reveals that the olivine grains form large (25 mm) porphyroclasts as well as smaller (0.25 mm) neoblasts; and that these grains often contain subgrain boundaries. Clinopyroxene crystals can be both very large (up to 1 cm long) and small (0.25 mm) in the same xenolith. The large crystals contain abundant exsolved lamellae of spinel, orthopyroxene and garnet; however, not all three exsolved phases occur in the same grain (see Sen and Jones [25] and Sen [12] for further details). Petrographically, these large clinopyroxene crystals appear very similar to those which constitute the clinopyroxenites.

The very large crystals with abundant exsolution lamellae often poikilitically enclose broken crystals (xenocrysts) of olivine, orthopyroxene and clinopyroxene (Fig. 3), that are probably remnants of the wall rock which itself may have been either spinel lherzolite (lithosphere) or a previously crystallized pyroxenite vein intrusive into the lithosphere. Interestingly, garnet has not been observed as a xenocryst in these large clinopyrox-

IRON-RICH XENOLITHS: ORIGIN AND IMPLICATIONSFOR HAWAIIAN MAGMA SOURCES

49

TABLE la Chemical and mineral composition of FELGs 69 SAL-97

SiO2 TiO 2 A1203 Cr203 FeO * MnO MgO CaO Na20 K20

Ol

Opx

Cpx

Gt

Phi

Amp

Green Sp

Brown Sp

Rock

Ol

Opx

Cpx

Gt

Phi

Sp

44.5 0.29 3.04 n.d. 11.1 0.15 37.3 3.25 0.35

39.70 n.d. n.d. n.d. 12.30 n.d. 47.86 0.07 n.d.

54.04 0.28 4.15 0.0 11.07 n.d. 29.50 0.97 0.0

51.63 0.75 6.82 0.18 4.69 n.d. 15.25 18.10 1.82

41.13 0.17 23.14 0.09 10.44 0.30 19.13 5.12 n.d.

38.95 4.04 16.13 0.0 6.10 n.d. 21.57 0.12 0.68 8.93

41.02 6.58 14.59 0.0 6.87 n.d. 14.59 10.69 2.49 1.48

n.d. 0.25 62.96 2.07 14.61 0.10 20.56 0.0

n.d. 0.37 50.84 13.02 16.73 0.12 18.00 0.08

47.8 0.57 7.75 n.d. 9.52 0.17 24.8 8.41 0.95 n.d.

40.93 n.d. n.d. n.d. 15.14 n.d. 44.53 0.12 n.d. n.d.

53.66 0.23 5.43 0.08 10.08 n.d. 30.12 1.00 0.19 n.d.

52.92 0.69 6.96 0.02 5.10 n.d. 14.19 17.68 2.28 n.d.

42.70 0.01 21.77 0.00 10.75 n.d. 19.01 5.06 n.d. n.d.

39.44 4.17 15.72 0.20 6.81 n.d. 19.65 0.26 0.66 8.39

0.16 0.34 59.45 1.69 17.06 n.d. 19.75 0.18 n.d. n.d.

99.93

100.01 99.24

99.12

96.53

98.31

100.55

99.16

100.72 100.79 99.84

99.30

95.30

98.63

Total V La Sm Eu Tb Yb Lu Ba

66SAL-1

Rock

81 0.86 0.41 0.16 0.07 0.16 0.02

184 1.78 1.08 0.44 0.30 1.05 0.16 1.45

70SA1-52

SiO 2 TiO 2 A1203 Cr203 FeO * MnO MgO CaO Na20 K20

Rock

O1

Cpx

Gt

Rock

O1

Opx

Cpx

Amp

Sp

45.0 0.66 7.56 n.d. 10.5 0.16 25.9 9.31 0.95 n.d.

39.36 0.02 n.d. n.d. 16.48 n.d. 43.43 0.08 n.d. n.d.

50.56 1.03 7.30 0.20 4.99 n.d. 13.92 19.83 1.81 n.d.

40.88 0.48 21.68 0.47 12.70 n.d. 17.54 5.02 n.d. n.d.

49.4 0.54 5.13 n.d. 10.3 0.16 25.4 8.37 0.74 n.d.

39.85 n.d. n.d. n.d. 15.69 44.37 0.12 n.d. n.d.

53.11 0.20 5.09 0.00 9.73 n.d. 30.17 0.75 0.00 n.d.

51.56 1.07 7.58 0.23 4.42 n.d. 14.13 19.88 1.60 n.d.

42.87 4.88 12.52 0.05 5.55 n.d. 16.90 10.71 2.45 1.48

0.65 0.78 57.12 5.11 16.13 n.d. 19.43 0.19 n.d. n.d.

99.37

99.64

98.77

100.03

99.05

100.47

97.41

99.41

Total V La Sm Eu Tb Yb Lu Ba

69SAL-89

195 2.49 1.58 0.59 1.58 0.72 0.11 30

172 1.73 1.12 0.42 0.24 0.51 0.05 35

Whole-rock analyses (all elements except Si) were determined by WPL using instrumental neutron activation analysis at the Radiation Center, Oregon State University; analytical uncertainties are generally better than 5% relative [7]. SiO 2 was calculated by difference assuming all major element oxides total 100%; K 2 0 is negligible for this purpose based on published analyses [2,10,11,13,16,41]. Where direct comparison can be made with U.S. Geological Survey analyses [13; E.D. Jackson, pers. commun., 1980], SiO2 values agree within 0.5 wt.%. Mineral analyses were performed with an electron microprobe; methodology as in ref. [12]. n.d. = not determined.

50

G. SEN AND W.P. LEEMAN

TABLE lb Whole-rock analyses of other Hawaiian xenoliths Spinel lherzofite suite

SiO2 TiO 2 AI203 FeO * MnO MgO CaO Na20 V La Sm Eu Tb Yb Lu

Pyroxenites

69SAL-53

69SAL-118

69SAL-2

69SAL-214

68SAL-6

69SAL-93

69SAL-43

69SAL-29

69SAL-80

46.0 0.04 2.40 7.31 0.12 42.0 1.95 0.14 50 0.11 0.092 0.033

44.8 0.14 2.23 9.22 0.12 41.3 1.92 0.27 57 0.46 0.33 0.121 0.053 0.1 0.014

46.1 0.08 21.6 8.14 0.13 40.8 2.26 0.32 48 0.61 0.35 0.116

49.4 1.1 10.2 9.49 0.17 16.8 11.8 1.01 355 2.10 1.99 0.78 0.43 1.28 0.174

51.5 0.64 8.26 7.09 0.15 18.6 12.9 0.83 230 0.72 1.15 0.43 0.30 0.78 0.111

50.9 0.57 8.81 7.78 0.15 18.7 12.0 1.06 244 1.75 1.35 0.52 0.29 0.67 0.101

50.6 0.78 9.65 8.22 0.16 17.6 11.9 1.09 253 1.93 1.59 0.61 0.34 0.98 0.144

45.5 1.48 12.6 12.9 0.20 16.2 9.71 1.37 335 2.05 2.10 0.75 0.48 1.47 0.208

44.6 2.77 12.3 14.1 0.14 14.7 10.2 1.20 419 1.42 1.71 0.69 0.42 0.51 0.043

0.16 0.023

0.12

enes. Orthopyroxene crystals are generally medium sized (20 mm) and carry abundant fine exsolved lamellae of clinopyroxene and less abundant, tiny tablets of spinel. Garnet and spinel crystals vary greatly in abundance and size. Garnet commonly occurs as a rim around spinel, which may be a result of the reaction spinel + magma = garnet or Al-rich pyroxene + spinel = garnet + Al-poor pyroxene + olivine [10,12,26]. Although garnet most commonly occurs as reaction rims, spinel-free garnet crystals (primary) are also found to occur interstitially [see also [26]). The spinel crystals vary significantly in size, color and shape. The dominant variety of spinel has a light to deep greenish gray color, varies from euderal to anhedral in habit, and is similar to the spinels that occur in Hawaiian pyroxenites [8]. A somewhat less common variety is a medium-to-coarse brown anhedral spinel. The FELG sample 69SAL-97 is an example in which both types of spinel occur. FELG xenoliths often contain minor amphibole and phlogopite, as well as distinctive pockets of a Na-rich hydrous glass with olivine + clinopyroxene + / - spinel + / - amphibole phenocrysts (Sen, in preparation). Such glasses are uncommon in Hawaiian spinel lherzolites. 5. Chemical and mineral composition

Bulk rock and mineral analyses of FELGs, along with some clinopyroxenites and spinel

lherzolites are presented in Table 1 and plotted in Figs. 4-8. Note that these figures include additional analyses published by others [11,13,26]. Figure 4 shows that whereas the spinel lherzolites and the pyroxenites form two distinct trends, the F E L G xenoliths are scattered: one sample, 69SAL-97, plots amongst the spinel lherzolites, a few overlap the pyroxenites and the remaining xenoliths plot between the spinel lherzolites and pyroxenites. The trends of simultaneous decrease of A1203 and F e O * / M g O with CaO, shown by the spinel lherzolite suite, is expected of partial melting residues [8]. On the other hand, the trends shown by the pyroxenites in Fig. 4 are typical of clinopyroxenes (the dominant constituent of these rocks) that crystallize from alkalic mafic magmas at high pressures [26, p. 225]. The scatter shown by the FELGs and the area in which they plot are not typical of mantle xenoliths from other areas. For example, a comparison between the Hawaiian xenoliths and spinel lherzolites (i.e., Cr-diopside group; [28]), websterites and pyroxenites (Al-augite group) from 68 localities of the western United States (WUS; data from ref. [28]) shows that while fields of WUS and Hawaiian spinel lherzolites overlap, the websterites and pyroxenites of WUS are generally dissimilar to the Hawaiian pyroxenite and F E L G fields (Fig. 4 a,b). The olivine-poor FELGs that compositionally overlap the Hawaiian pyroxenites are equivalent to the 'euhedral olivine eclogites' of Kuno [26] and probably mostly repre-

IRON-RICH

XENOLITHS:

ORIGIN

AND IMPLICATIONS

6

I

FOR HAWAIIAN

U

MAGMA

51

SOURCES

I

U

I

I

(b)

5

fo~jX!o

(a)

A

!.'.

.::i:/i":::':::

0 ,i-

0 • O

0 ~3



• II•

0

==



..°.°

69SAL-97 a 5

0 0

n 10

I 15

,

0

I

o

20

5

10

20

CaO

CaO 6

15

6

n

(c)

0

0

0

0

¢>O O 0 0

0

0

M

Oo

[]

0

;,f'o .

[]

69SAL-97

M

O O (DO ",~- 69SAL-97

[]

3

O

0(0

e,i

O 2 I 2

,

I 3

n

I

4

CaO

,

2

2

'

3

,

CaO

I

4

n

Fig. 4. (a and b) Chemical comparison between spinel lherzolites (circles), pyroxenites (triangles), and FELG xenoliths (filled squares) from Oahu. Data include those presented here and Kuno [26]. Errors are less than or equal to the size of the points. The compositional fields for mantle xenoliths from the western United States (WUS [28]) are also shown for comparison purposes: shaded area--pyroxenites; ruled area--websterites; dashed area--spinel lherzoLites. Note that (1) the Hawaiian spinel lherzoLites and pyroxenites form generally well-defined trends and whereas (2) some FELGs plot on the Hawaiian pyroxenite trend, six of them scatter between the spinel lherzolites and pyroxenites. 69SAL-97 plots within the spinel lherzolite field. See text for further discussion. (c and d) Chemical comparison between proposed models of source upper mantle compositions and 69SAL-97. The circles represent model compositions by a number of authors, including Hart and Zindler, Palme and Nickel, Carter, Ringwood, Maaloe and Aoki, Jagoutz et al, Boyd, and McDonough and Frey (references Listed in [39,40]). Asterisk--Sen [8]; diamond--Wilkinson [16]; unfilled squares--Wright [15].

52

G. SEN AND W.P. LEEMAN

sent vein crystallization products. The FELGs plrtting between 5-10% CaO, of which 69SAL-97 is an example, are unlikely to have been generated by any simple partial fusion or crystallization process but are inferred to be a result of random mixing (accompanied by extremely variable Fe, Na, and REE enrichment) between wall rock lherzolites and intrusive pyroxenites (discussed further later). Figures 4c and d show a comparison between 69SAL-97 and various models of the undepleted upper mantle proposed by previous workers ([17,18,39,40, and refs. in [39,40]). Most of the model compositions are more magnesian and aluminous than 69SAL-97. Of particular relevance are the compositions of the source mantle for Hawaiian tholeiites proposed by Wright [15], Wilkinson [16], and Sen [18]. Sen's model composition is richer in A1 a n d Mg than 69SAL-97. Although the compositions estimated by Wright and Wilkinson are significantly higher in A1, they are reasonably close in M g O / F e O * to 69SAL-97. Therefore, the major element composition of 69SAL-97 is not too different from model compositions proposed by a number of workers for basalt source regions.

I

Figure 5 shows that FELGs largely overlap the pyroxenites in Na and Ti. 69SAL-97 again plots with the spinel lherzolites. One F E L G xenolith (Kuno's specimen no. 29) has exceptionally high Na (and also unusually high Fe203, K 2 0 , and H20; [26, table 2]), suggesting that this specimen may be significantly altered a n d / o r has a high content of hydrous phases. Unfortunately, the extent of alteration and modal composition were not indicated by Kuno [26]. In addition, some compositional variation in F E L G xenoliths likely reflects relative abundance of a Na, LREE-rich interstitial glass in these rocks (see section 4). Figure 6 shows that the FELGs are characterized by relatively flat REE (chondrite-normalized) pattern, as compared with the 'common' LREE-enriched patterns shown by spinel lherzolite suite and convex upward patterns shown by the pyroxenites. Note that the fields shown for the pyroxenites and spinel lherzolites are based on very limited data (refs. in [8] and present data) and it is certain that both fields are much larger in shape and size. This statement is substantiated by recent ion microprobe determination of REEs in clinopyroxenes of these xenoliths by the first author [Sen et al., in prep.], which suggests that (1)

I

I

u

I

(a)

@

I

(b)

A

A

if nmu

l N

09sAL 97 o 0

,

,

1

2

TiO2

I

3

0.0

0.2

,

I

n

0.4

I

0.6

n

!

0.8

1.0

TiO2

Fig. 5. (a) Na-Ti variation in the Hawaiian xenoliths compared with that of the xenoliths from the western United States. Symbols as in Fig. 4a. (b) F E L G s compared with model source upper mantle compositions. Symbols as in Fig. 4c and d.

I RO N - R IC H XENOLITHS: O R I G I N A N D I M P L I C A T I O N S FOR HAWAIIAN M A G M A SOURCES

100

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

..-

3

1

z I

I

I

I

~

i

i

t

t

{

I

I

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Fig. 6. Chondrite-normalized REE patterns of six FELG xenoliths compared with some spinel lherzolites (ruled) and pyroxenites (shaded) from Oahu. Filled squares--69SA1-97.

absolute abundances of the REEs vary tremendously, and (2) the spinel lherzolites vary from LREE-depleted to LREE-enriched types. In a general way, it is possible to generate the flat REE patterns of FELGs by mixing of a convex upward REE pattern of a pyroxenite and a LREE-enriched pattern of a spinel lherzolite. In detail such mixing calculations are premature because of uncertainties in compositions of the spinel lherzolite and pyroxenite end members as well as higher La abundance in some of the FELGs than the endmember suites. 69SAL-97 is somewhat unusual in that its H R E E and LREE contents are similar to the spinel lherzolite and pyroxenites, respectively; but its MREE are intermediate (Fig. 6). Figure 7 shows variation in Yb amongst Hawaiian xenoliths as a function of CaO. Four of the six F E L G samples plotted were analyzed in this study (Table 1), the remaining data are from Frey [11]. The pyroxenites show a wide variation in Yb for a small change in CaO, which is a reflection of the wide variation of clinopyroxene/garnet modal ratios in this suite. The FELGs generally plot between the pyroxenites and spinel lherzolites; 69SAL-97 plots with the spinel lherzolites. The field in which the FELGs plot is rather unusual for xenoliths from other areas (e.g., WUS xenoliths). Considering the petrographic evidence, we infer that the intermediate composition of F E L G s in Fig. 7 originated by mixing between spinel lherzolites and pyroxenites.

53

In terms of mineral composition, olivines in F E L G s (Fo83_87) partially overlap pyroxenite olivines (Fo79_83: [12]) and are generally more Fe-rich than the spinel lherzolite olivines (Fo86_92" [12]). Generally, they do not appear to show any correlation between mineral composition and mode: for example, 66SAL-1 (spinel+ garnet websterite) with 29% olivine and 69SAL-121 (spinel + garnet lherzolite) with 55% olivine have essentially the same olivine composition (Fo84; Table 1). Interestingly, 69SAL-97 contains olivine (Fo87) whose composition is the same as that of the model undepleted upper mantle proposed by Carter [17]. However, we suggest that 69SAL-97 is a mixed rock whose anhedral olivines were once more magnesian ( > Fo88 ) and part of the wall rock spinel lherzolite but became more Fe-rich upon reaction with magmas which precipitated the clinopyroxenite veins partially engulfing the wall rock fragments. Such vein-wall rock reactions have been previously documented for xenoliths from Hawaii [12,27] and other areas of the world [28,33]. As mentioned earlier, significant variation exists in the color of the spinel grains of the F E L G xenoliths, the extremes of which are brown and grayish green varieties. The grayish green spinel is poor in C r / ( C r + A1 + Fe 3+) and is similar to the spinels in the clinopyroxenites (Fig. 8). The brown variety is Cr-rich and is similar to the spinel in

20

,~

,gl

A

10

"

A A •

a 0

A

., -, L~il~ A •A A •

I

10

,

20

CaO Fig. 7. CaO vs. Yb variation between FELGs, spinel lherzolites and pyroxenites from Oahu. Symbols as in Fig. 4a and b.

54

G, S E N A N D W.P. L E E M A N

0.5 I |

,

,

,

69SAL-99 (Brown

[

sp)

0.4 ~" I~, Sp. Lherzolite ~ . "T field ~ 0.3 / ~ ~ + 4-

0.2

. 0.0

.....~

0.5

~&mll~ ~Green sp ~

0.6



,

0.7

,

~

0.8

.

J

0.9

Mg/[Mg+Fe2+] Fig. 8. Spinel compositions of FELGs (filled squares), pyroxenites (triangles) and spinel lherzolites from Oahu. Data are from Sen [12]. The compositions of brown and green spinels in 69SAL-97 are shown. 69SAL-99 also has green and brown spinels; but only composition of the brown spinel is labelled. Its green spinel plots amongst the Cr-poor FELGs. See text for further explanation.

spinel lherzolites. A wide range in spinel composition exists in individual xenoliths: for example, in 69SAL-99 (modal composition--O155Cpx24Opx]8 Gar2Sp]Phltr; [12]) most of the spinels are greenish and have C r / ( C r + A1 + Fe 3+) between 0.03-0.04 but one grain (brown) has a value of 0.4 (Fig. 8). 69SAL-97 also contains two types of spinels and their compositions are identified in Fig. 8. In our view, at least these high-olivine FELGs must be hybrids in which the brown spinels (and much of the anhedral olivine and orthopyroxene) represent relict fragments of the wall rocks (spinel lherzolites) in a matrix composed dominantly of spinel + garnet clinopyroxenites (vein materials). Sen [12] further pointed out that in individual xenoliths a spinel rim around garnet is always more Cr-rich than the one without a rim. Other minerals of the FELGs, i.e., garnets, amphiboles and phlogopites, are compositionally similar to those of the pyroxenites [12]. 6. Discussion

6.1. Summary of observations In summary, the following features of the garnet-bearing F E L G xenoliths of Oahu are

noteworthy: (1) Compared to other xenolith lithologies they are rare. (2) There is a lack of correlation between modal content of olivine and olivine composition: a websterite and a lherzolite with very different olivine contents may have the same olivine composition. (3) Garnet occurs interstitially, commonly forming rims around spinel (and not vice versa), and also as exsolution blebs in clinopyroxene. (4) A relatively Fe-rich olivine (Fo83_87) occurs as clustered aggregates in a clinopyroxene-rich groundmass. (5) The olivine crystals are of two morphologic types: the anhedral variety predominates in the olivine-rich samples and the euhedral variety is present in FELGs of all modal varieties. Transitional types also occur. (6) Large clinopyroxene crystals in F E L G xenoliths sometimes contain xenocrysts of wall rock minerals (i.e., olivine, ortho- and clinopyroxenes). (7) In rare olivine-rich examples (e.g., 69SAL-97) two distinct types of spinels occur, one of which is Cr-rich and appears to be relict from the wall rock (spinel lherzolite) and the other is Cr-poor and is similar to pyroxenite spinels. (8) In CaO vs. F e O * / M g O and A1203 plots, the spinel lherzolites and pyroxenites form relatively smooth trends, the olivine-poor FELGs overlap the field of pyroxenites and the olivine-rich FELGs mostly plot between the spinel lherzolite and pyroxenite fields. 69SAL-97 is particularly interesting as its composition is similar to proposed model mantle sources for oceanic basalts. (9) The FELGs show relatively flat REE patterns. 69SAL-97 plots in an intermediate area between spinel lherzolite and pyroxenite fields.

6.2. Origin of FELG xenoliths Any successful hypothesis concerning the origin(s) of the xenoliths must satisfactorily explain the above observations. Additional constraints must come from the recently published isotopic data on some of these xenoliths [24,29,38,41] discussed below. As outlined earlier, the two previously proposed hypotheses are that: (1) F E L G xenoliths represent undepleted plume material (i.e., ' piclogite' of Anderson [35]) which is the major contributing source component of the tholeiites, and (2) they are magmatic vein materials and share a common petrogenesis with the pyroxenites and therefore are a part of a lithologic

I RO N - R IC H XENOLITHS: O R I G I N A N D IMPLICATIONS FOR HAWAIIAN M A G M A SOURCES

continuum with the garnet pyroxenites [10,11,12, 26]. The evidence against hypothesis (1) m a y be summarized as follows. First, experimental phase equilibrium studies have strongly demonstrated that in ultramafic rocks the boundary separating the garnet and spinel peridotite stability fields is curved in P - T space, with garnet being stable at higher pressures than spinel [12]. If the F E L G s were plume material brought up from the deeper mantle, then one would expect to see spinel rims around garnets, but not the converse. So far, not a single garnet-bearing xenolith has been found in the Hawaiian islands in which spinel-rimmed garnets occur. On the other hand, the commonly observed rimming of spinels by garnet can be explained assuming near-isobaric cooling of a spinel bearing primary (solidus) ultramafic assemblage [10,12]. Second, as magmas are progressively extracted during the partial melting of lherzolite, the residual minerals become increasingly refractory and modal contents of olivine and orthopyroxene increase (cf. [30]). However, no such correlation exists amongst the F E L G rocks. Third, the observed scatter amongst the low CaO F E L G s between pyroxenites and spinel lherzolites in Fig. 4 cannot be expected of a series of residues that have undergone variable extraction of melt from a common source. Fourth, the available isotopic data suggest that the F E L G s are distinct from the Koolau tholeiites and therefore cannot be the source for the tholeiites. For example, the three F E L G xenoliths analyzed by Shimizu [41] and two F E L G xenoliths analyzed by Basu and Tatsumoto [29] and Vance et al. [38] all have very different Sr and Nd isotopic composition from the tholeiites: 66SAL-1 and 69SAL-96 have an eNd of about + 8, whereas Koolau tholeiites have an eNd of < + 4 [31]. Taken together, the evidence is overwhelmingly against the proposal that the F E L G xenoliths are physical representatives of undepleted plume material involved in Hawaiian tholeiitic m a g m a generation. The second hypothesis that the F E L G s are lithological variants of the pyroxenites is addressed as follows. As discussed earlier, it was K u n o [26] who first distinguished between two subtypes of FELGs: the "euhedral olivine bearing" and the "anhedral olivine bearing .... olivine eclogites". We observed that this first-order distinction also exists

55

in terms of modal ohvine/clinopyroxene ratio, M g O / F e O * , CaO, and REEs. It was also noted in an earlier section that there are m a n y compositional features c o m m o n between the clinopyroxenites and olivine poor F E L G s (or euhedral olivine bearing olivine eclogites) which support a comm o n petrogenesis. Therefore, we agree with the previous workers [10-12,26] that these olivine-poor F E L G s likely formed by crystallization from mafic magmas with clinopyroxene and spinel appearing early and olivine appearing late in the crystallization sequence [12,26]. It is the petrogenesis of the olivine rich F E L G s that is of broader significance in our discussion here inasmuch as these rocks compositionally resemble proposed upper mantle source model compositions (discussed earlier). The petrographic and

(a)

(b)

Fig. 9. Sketch illustrating our preferred model for the origin of FELG xenoliths. Dotted pattern--pyroxenite vein; blocky pattern--spinel lherzohte wall rock (lithosphere). In the first step (a) intrusion and stoping of wall rocks by intrusive pyroxenite veins occur. In more advanced stage (b), the stoped fragments disintegrate and are thoroughly disseminated in the vein. Significant chemical reactions between wall rock minerals and the intrusive vein magmas (or exsolved fluids) may accompany the mechanical mixing process. As the pyroxenites consist only of spinel + / - garnet and clinopyroxene, most of the olivine and orthopyroxene in such mixed rocks (i.e., FELGs) must be reacted relicts from spinel lherzolite wall rocks ( = lithosphere).

56

mineral chemical features, such as the occurrence of xenocrysts of olivine and orthopyroxene in a clinopyroxene-rich groundmass, clustered olivines, two compositionally distinct spinels (spinel lherzolitic and pyroxenitic), all support a hybrid origin of these rocks. These rocks appear to be products of mixing between garnet clinopyroxenite intrusives and spinel lherzolite wall rocks. An examination of composite xenoliths, in which pyroxenites and spinel lherzolites are juxtaposed, suggests that the olivine clusters in these xenoliths represent dispersed remnants of wall rock (spinel lherzolite) intruded, reacted, and distintegrated by intruding pyroxenite-parent magmas during stoping (Fig. 9). Study of such composite xenoliths further suggests that the chemical reactions accompanying this mixing may have resulted in Na, REE, and Fe-enrichment of the phases [12,27, Sen et al., in prep.]. A later-stage metasomatic process may have further reset Na, REEs in these composite xenoliths. In some cases, new minerals (e.g., garnet rims around spinel; [12]) crystallized as a result of such reactions. On the other hand, brown spinels from the wall rock reacted more slowly and often survived by developing only a 2-5 micron black rim that is richer in Cr, Fe than the brown core. Black rims around spinels were not formed always this way, however, and in many cases such rims may have formed by reaction between the enclosing lava and the xenolith spinel or due to decompression melting of spinel during ascent of the xenolith from great depth [12,42]. Our conclusion that olivine-rich F E L G xenoliths are products of physical mixing and chemical reaction between pyroxenite and spinel lherzolite, together with the limited isotopic evidence cited earlier, contradicts the view that the physically represent source rock for Hawaiian tholeiitic magmas. This conclusion does not necessarily rule out a relatively Fe-rich source for such magmas (discussed earlier), but it does cast doubt on petrologic models involving undepleted (or fertile) Hawaiian upper mantle similar to the F E L G suite.

7. Condusions The Fe-rich garnet lherzolites of Oahu are mixed rocks composed of stoped and reacted spinel lherzolite fragments embedded in pyroxenites which were precipitated by the parental magmas

G. SEN AND W.P. LEEMAN

of the Honolulu Volcanics. These rocks cannot represent the plume component or the source mantle rock for the Koolau tholeiites. Although it is possible that the olivine-rich FELGs are chemically similar to the source rock for Hawaiian tholeiites, it is concluded that the FELGs cannot be physical representatives of the undepleted source. It is possible that these magma-added lherzolites (which have been depleted in basaltic components in previous melting episodes) eventually become favorable sites of future magma generation when the appropriate thermal, chemical and tectonic conditions are created.

Acknowledgements Portions of this research were supported by NSF grants EAR 85-12167 (WPL) and EAR 8903879 (GS). The xenoliths studied were collected by the late E.D. Jackson and D.C. Presnall. The authors thank Sorena Sorensen and the Smithsonian Institution for providing access to the Dale Jackson Collection. G.S. appreciates the continued support of D.C. Presnall who introduced him to studies of Hawaiian xenoliths. We are grateful to C.H. Langmuir, F.A. Frey, J. Ryan and T.L. Wright for their thorough reviews which helped the authors in revision of the manuscript. Thanks are particularly due to M.F. Roden for discussion, incisive review and important comments that led to a clearer (hopefully) presentation of our thesis. The ideas presented here are entirely ours and may differ significantly from those of the reviewers.

References 1 M.D. Feigenson, Constraints on the origin of Hawaiian lavas: J. Geophys. Res. 91, 9383-9393, 1986. 2 F.A. Frey and M.F. Roden, The mantle source for the Hawaiian islands: constraints from the lavas and ultramafic inclusions, in. Mantle metasomatism, M.A. Menzies, and C.J. Hawkesworth, eds., pp. 423-463, Academic Press, London, U.K., 1987. 3 H.B. West, D.C. Gerlach, W.P. Leeman and M.O. Garcia, Isotopic constraints on the origin of Hawaiian lavas: evidence from the Maui Volcanic Complex, Hawaii, Nature 330, 216-220, 1987. 4 T.L. Wright, Chemistry of Kilauea and M a n n a Loa lavas in space and time, U.S. Geol. Surv. Prof. Pap. 735, 40 pp., 1971. 5 D.H. Richter and K.J. Murata, Xenolithic nodules in the

IRON-RICH XENOLITHS: ORIGIN AND IMPLICATIONS FOR HAWAIIANMAGMASOURCES

6

7

8

9

10 11

12

13 14

15 16 17

18 19

20

21 22

23

24

1800-1801 Kaupulehu flow of Hualalai volcano, U.S. Geol. Surv. Prof. Pap. 424B, B215-B217, 1961. W.P. Leeman and K.F. Scheidegger, Olivine/liquid distribution coefficients and a test for crystal-liquid equilibrium, Earth Planet. Sci. Lett. 35, 247-257, 1977. W.P. Leeman, J.R. Budahn, D.C. Gerlach, D.R. Smith and B.N. Powell, Origin of Hawaiian tholeiites: trace element evidence, Am. J. Sci. 280-A, 794-819, 1980. G. Sen, Xenoliths associated with the Hawaiian hot spot, in: Mantle xenoliths, P.H. Nixon, ed., pp. 359-375, John Wiley & Sons, London, 1987. H.R. Shaw and E.D. Jackson, Linear island chains in the Pacific: result of thermal plumes or gravitational anchors?, J. Geophys. Res. 78, 8634-8652, 1973. D.H. Green, The origin of "eclogites" from Salt Lake crater, Hawaii, Earth Planet. Sci. Lett. 1, 414-420, 1966. F.A. Frey, The origin of pyroxenites and garnet pyroxenites from Salt Lake Crater, Oahu, Hawaii: Trace element evidence, Am. J. Sci., 280-A, 427-449, 1980. G. Sen, Petrogenesis of spinel lherzolite and pyroxenite suite xenoliths from the Koolau shield, Oahu, Hawaii: implications for petrology of the post-eruptive lithosphere beneath Oahu, Contrib. Mineral. Petrol. 100, 61-91, 1988. E.D. Jackson and T.L. Wright, Xenoliths in the Honolulu volcanic series, Hawaii, J. Petrol. 11, 405-430, 1970. W.P. Leeman, A.V. Murali, M.-S. Ma and R.A. Schmitt, Mineral constitution of mantle source regions for Hawaiian basalts--rare earth evidence for mantle heterogeneity, in: Magma genesis, H.J.B. Dick, ed., Oreg. Dep. Geol. Miner. Ind. Bull. 96, 169-183, 1977. T.L. Wright, Origin of Hawaiian tholeiites: a metasomatic model, J. Geophys. Res. 89, 3233-3252, 1984. J.F.G. Wilkinson, Undepleted mantle composition beneath Hawaii, Earth Planet Sci. Lett. 75, 129-138, 1985. J.L. Carter, Mineralogy and chemistry of the earth's upper mantle based on the partial fusion-partial crystallization model, Geol. Soc. Am. Bull. 81, 2021-2034, 1970. A.E. Ringwood, Composition and Petrology of the Earth's Mantle, 618 pp, McGraw-Hill, New York, NY, 1975. G.A. Macdonald and A.T. Abbott, Volcanoes in the Sea: The Geology of Hawaii, 441 pp., Univ. of Hawaii, Honolulu, 1970. R.R. Doell and G.B. Dalrymple, Potassium-argon ages and paleomagnetism of the Waianae and Koolau Volcanic series, Oahu, Hawaii, Geol. Soc. Am. Bull, 84, 1217-1242, 1973. R.W. White, Ultramafic inclusions in basaltic rocks from Hawaii, Contrib. Mineral. Petrol. 12, 245-314. G. Sen, A petrologic model for the constitution of the upper mantle and crust of the Koolau shield, Oahu, Hawaii, and Hawaiian magmatism, Earth Planet. Sci. Lett. 62, 215-228, 1983. G. Sen and D.C. Presnall, Petrogenesis of dunite xenoliths from Koolau volcano, Oahu, Hawaii: implications for Hawaiian volcanism, J. Petrol. 27, 197-217, 1986. O. Okano, M. Tatsumoto and W.P. Leeman, Sr, Nd, and Pb isotopes in Hawaiian basalt xenoliths: inferences for the origin of Hawaiian basalts, EOS 68, 1521, 1987.

57

25 G. Sen and R.E. Jones, Exsolved silicate and oxide phases from clinopyroxenes in a single Hawaiian xenolith: impfications for the oxidation state of the Hawaiian upper mantle, Geology 16, 69-72, 1988. 26 H. Kuno, Mafic and ultramafic nodules in basaltic rocks of Hawaii, Geol. Soc. Am. Bull. 115, 189-234, 1969. 27 A. Goto and K. Yokoyama, Lherzolite inclusions in olivine nephelinite tuff from Salt Lake Crater, Hawaii, Lithos 21, 67-80, 1987. 28 H.G. Wilshire, C.E. Meyer, J.K. Nakata, L.C. Calk, J.W. Shervais, J.E. Nielson and E.C. Schwarzman, Mafic and ultramafic xenoliths from volcanic rocks of the western United States: U.S. Geol. Surv. Prof. Pap. 1443, 179 pp., 1988. 29 A.R. Basu and M. Tatsumoto, Nd isotopes in selected mantle-derived rocks and minerals and their implications for mantle evolution, Contrib. Mineral. Petrol. 75, 43-54, 1980. 30 A.L. Jaques and D.H. Green, Anhydrous partial melting of peridotite at 0-15 kb pressure and the genesis of tholeiitic basalts, Contrib. Mineral. Petrol. 73, 287-310, 1980. 31 M.F. Roden, F.A. Frey and D.A. Clague, Geochemistry of tholeiitic and alkalic lavas from the Koolau range, Oahu, Hawaii: implications for Hawaiian volcanism, Earth Planet. Sci. Lett. 69, 141-158, 1984. 32 M.A. Lanphere and G.B. Dalrymple, Ages and strontium isotopic composition of the Honolulu volcanic series, Oahu, Hawaii, Am. J. Sci. 280-A, 736-751, 1980. 33 S.Y. O'Reilly, Volatile-rich mantle beneath eastern Australia, in: Mantle Xenoliths, P.H. Nixon, ed., pp. 661670, John Wiley & Sons, London, 1987. 34 R.T. Helz, Diverse olivine types in lava of the 1959 eruption of Kilauea volcano and their bearing on eruption dynamics, U.S. Geol. Surv. Prof. Pap. 1350, 691-722, 1987. 35 D.L. Anderson, Theory of the Earth, 366 pp., Blackwell Scientific, Boston, 1989. 36 C.H. Langmuir and G.L. Hanson, An evaluation of major element heterogeneity in the mantle sources of basalts, Phil. Trans. R. Soc. London Ser. A297, 383-407, 1980. 37 P.H. Nixon (ed.) Mantle Xenoliths, Wiley, London, 1987. 38 D. Vance, J.O.H. Stone and R.K. O'Nions, He, Sr, and Nd isotopes in xenoliths from Hawaii and other oceanic islands, Earth Planet. Sci. Lett. 96, 147-160, 1989. 39 W.F. McDonough and F.A. Frey, Rare earth elements in upper mantle rocks, in: Geochemistry and Mineralogy of Rare Earth Elements, B.R. Lipin and G.A. McKay, eds., Mineral. Soc. Am., pp. 99-145, 1989. 40 F.R. Boyd, Compositional distinction between oceanic and cratonic lithosphere, Earth Planet. Sci. Lett. 96, 15-26, 1989. 41 N. Shimizu, Geochemistry of ultramafic inclusions from Salt Lake crater, Hawaii, and from southern African kimberlites, Phys. Chem. Earth 9, 655-669, 1975. 42 H.J.B. Dick and T. Bullen, Chromian spinel as a petrogenetic indicator in abyssal and alpine-type peridotites and spatially associated lavas, Contrib. Mineral. Petrol. 86, 5476, 1984.