Roadmap | The review process | Home
   Samoa is cool
Why olivine control arguments cannot be used to infer high temperature at Samoa

James H. Natland

Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149



Olivine-liquid geothermometry based on FeO-MgO equilibrium of picritic basalt usually does not work because picrites almost always are hybrid rocks, the result of mixing between differentiated liquids on a low-temperature two- or three-phase cotectic, and crystal sludges of more primitive magmas rife with accumulated olivine. The mixing results in higher FeO (here taken to be ~ 0.9 x total iron as FeOT) in the resultant picrite than would otherwise have been the case, producing artificially high estimates of parental MgO, eruptive temperature, and potential temperature from backtrack techniques. In the case of Samoa, an additional complication is the presence of clinopyroxene together with olivine phenocrysts in almost every picrite; the rocks cannot be described as “olivine-controlled”, the essential assumption for backtrack techniques. The influence of clinopyroxene on liquid compositions is readily shown on simple major-oxide variation diagrams. The effects of mixing are not so pronounced at Vailulu`u, which is a small, active, zero-age volcano at the eastern end of the chain with an as-yet poorly developed flanking rift system where differentiates would normally extensively develop. Magma is thus more directly tapped at Vailulu`u from mantle sources, but picrites there have the least FeO at given MgO contents in all of Samoa, and most of them still carry some clinopyroxene phenocrysts. This means that they have at least some contribution from a cotectic liquid in their makeup. Thus no estimates of the temperature of parental liquids in equilibrium with the Samoan mantle can yet be made, even at Vailulu`u.


A number of studies use FeO-MgO olivine-liquid partitioning (e.g., Roeder & Emslie, 1970) to estimate compositions, eruptive temperatures and potential temperatures of primitive magmas (e.g., Larsen & Pedersen, 2000; Breddam, 2002; Herzberg & O’Hara, 2002; Putirka et al., 2007; Falloon et al., 2007a, b; Herzberg et al., 2007; Putirka, 2008). The studies all identify basaltic magmas of picritic composition, that is, with abundant olivine phenocrysts, as the only ones liable to be descended from likely primary magmas by means of simple olivine-controlled differentiation beginning at elevated pressure, as first inferred by O’Hara (1968). Since O’Hara (1968) also said that all eruptive basalts have experienced at least some olivine differentiation, the primary magma is invariably a back-calculated or derived, rather than an eruptive, composition that is presumed to be in equilibrium with mantle olivine (~Fo91). Of the several studies, only Falloon et al. (2007a,b) take into account pressure in influencing FeO-MgO partitioning (e.g., Ford et al., 1983); this helps provide the likely range of primary compositions at relatively low potential temperature. From experiments in a six-component mantle-analog system Presnall & Gudfinnsson (2008) also strongly advocate small differences in depths and temperatures of partial melting between basalts from islands and those from spreading ridges. All others calculate high potential temperatures for derived primitive compositions from islands, which mainly have MgO contents > 20%, and this implies convective overturn (Putirka et al., 2007, Putirka, 2008). Convective overturn in turn supports plume models for mantle convection.

One place outside of Hawaii that gives particularly high potential temperatures, 223° above that of mantle beneath spreading ridges (Purtirka et al., 2008), is Samoa, a linear volcanic chain in the southwestern Pacific near the Tonga Trench (e.g., Daly, 1924; Stearns, 1944; Kear & Wood, 1959; Natland, 1980a). Background on Samoa and a tectonic model for its origin are given in Natland (2003a). An alternative view is given in Hart et al. (2004). For this discussion, I take the example of Samoa to provide a general criticism of the overall approach taken in the olivine-melt equilibria studies, asserting from the perspectives of petrography and bulk compositions that no basaltic liquid so far identified in the entire Samoan chain is truly “olivine-controlled”. Furthermore, no stage of simple olivine control can even be inferred from the study of ultramafic xenoliths, many of which crystallized from basaltic magma at depth. Instead almost all Samoan picrites have at least clinopyroxene together with olivine in phenocryst assemblages; this is a typical low-pressure assemblage of alkalic olivine basalt, producing “picritic basalt of the ankaramite type” (Macdonald, 1949).

Data from the eastern end of the chain presented by Workman et al. (2004) indicate that Samoan picrites are usually mixtures between strongly differentiated magmas with low MgO content on the one hand, and, on the other, porphyritic, mainly olivine-rich, but also clinopyroxene-bearing, “sludges” from high-level magma chambers that were entrained in the differentiated magmas during their ascent and passage through the magma conduit system. Such mixing was first inferred for Hawaiian picrites of Kilauea volcano and Puna Ridge (Clague et al., 1991, 1995) and this was later applied to picritic tuffs of the Nu`uanu submarine landslide off of Oahu (Natland, 2007a,b). Helz (1987) had long since provided convincing evidence for polygenetic (mixed) origin of olivine in Hawaiian tholeiite, including presence of a significant proportion of deformed dunite with subgrains – these essentially are xenoliths – and broken, not faceted, olivine taken from crystal mats or conduit walls. Of greatest importance is that mixing gives the hybrid magmas unusually high FeO at a given MgO content, but this has nothing to do with the primitive liquid line of descent or high potential temperature. Nevertheless, this is the erroneous basis for inferring high potential temperatures among Samoan picrites as well as those of other linear volcanic chains.

Picrites are Mixed Magmas

The difficulty with FeO-MgO olivine-liquid relationships is that picrites, lavas charged with olivine phenocrysts and which have ~12-30% MgO contents, are mixed magmas, with olivine derived from many different primitive mafic magma stems that mixed together (e.g., Natland, 1989; 2003b; 2007a,b). A given magnesian olivine in a rock thus might not be related to very many of the other olivine crystals in the same rock on a liquid line of descent, nor to the "host" liquid, which is itself a differentiate of a hybrid primitive mafic magma. The simplest way to test claims of "olivine-controlled liquid" is to check whether the rock in question is "olivine-controlled". That is, does it only have olivine phenocrysts or does it also have others? If it has others, then the bulk rock is necessarily a mixture that includes some proportion of a cotectic liquid. Then it cannot be used for backtrack calculation, even if some interval of olivine-only crystallization is evident from phenocryst compositions.

This is also true even if the hybrid mesostasis, the host glass or groundmass, is in an olivine-only primary phase volume at eruption; mixing may have been insufficient to saturate the hybrid with any other phase. Indeed, the curvature of the phase boundary may even have pulled the hybrid from the cotectic of the multiply saturated end-member into an olivine-only regime of crystallization (e.g., Natland, 2006). If an olivine-only episode of crystallization occurred, then it was only in some, perhaps merely one, of the parental liquids that contributed to the hybrid. A single faceted euhedral clinopyroxene crystal means that some portion of the hybrid moved beyond this point and crystallized that clinopyroxene on a cotectic. The crystal was in a liquid at the time mixing occurred. Thus this clinopyroxene and some of the melt from which it was crystallizing inevitably comprised this mixing component. The chances of scavenging a euhedral clinopyroxene crystal from a crystal sludge without incorporating some of the associated cotectic magma are extremely slim. If it were from mantle wall rock, the crystal would be broken, not euhedral. This likelihood makes the hybrid glass or groundmass with a euhedral clinopyroxene unsuitable for a backtrack calculation.

All that is necessary is look at thin sections. Together with olivine, spinel is always there, and electron probe microanalyses of both the spinel and the olivine typically indicate polymagmatic sources of both MORB and island picrites (Natland, 1989, 2003b). Otherwise, many picrites contain a few phenocrysts of plagioclase and/or clinopyroxene (Figure 1; e.g., from Iceland; Hansteen, 1991). In the absence of petrographic descriptions, say when considering data files from GeoRock (e.g., Putirka et al., 2007, Putirka, 2008), bulk-rock compositions can be compared to those of glasses in the same specimen, obtained by electron microprobe. This almost always reveals that the host glasses are low-temperature cotectic liquids – that they experienced some degree of high-iron differentiation involving plagioclase and clinopyroxene, with plagioclase fractionation in particular producing high FeO.

Figure 1: Scanned thin sections of ankaramites and picrites with clinopyroxene phenocrysts from Samoa. UF samples = those from Fagaloa volcano, Samoa; TP samples = those from Masefau volcano, Tutuila, American Samoa. MT sample = one sample from Manu`a, American Samoa, Ta`u volcano. Most samples contain olivine (light, with cracks and irregular outlines) and clinopyroxene (titanaugite = gray) phenocrysts. Sample 85MT-18 has two clinopyroxene phenocrysts, but the section is too thin to show them in the scanned image. Data for MgO contents and 3He/4He (R/Ra) are from Farley et al. (1992). Those for R/Ra from TP samples (Tutuila, Masefau volcanic series) were, until recently (Jackson, 2008) the highest known from Samoa. The sections show that the high R/Ra of Samoan samples occurs in olivine phencorysts of differentiated compositions.

In the Hawaiian case, the point of onset of plagioclase-clinopyroxene cotectic differentiation occurs at about 7% MgO contents, and corresponds to a sharp change in slope of trending data points on many variation diagrams (FeO-MgO, TiO2-MgO, CaO-MgO, Na2O-MgO, etc.). Of most importance, on a diagram of MgO versus FeO, cotectic differentiation of tholeiitic and alkalic basalt results in strong iron-enrichment (steep negative slope), something well known in igneous petrology for more than a century. This follows an interval of “olivine control” at higher MgO in which FeO does not increase. Basalt glasses from Hawaii’s Kilauea volcano and Puna Ridge with >7% MgO contents thus have a flatter slope, with MgO in the glass reaching 15% (Clague et al., 1995). Of some importance, however, Kilauea and Puna Ridge have the only olivine-controlled glasses known so far from anywhere in the ocean basins, and all of them are from sand grains in a single box core; there are no corresponding mineral analyses for these glasses. The supposition behind all FeO-MgO equilibria backtrack calculations is that similar olivine-controlled glasses must exist at every other ocean island or spreading ridge. If Hawaii has olivine-controlled glasses, in other words, every other island chain must have them as well. Whether they do or not, this assumption is usually extended to the notion that picrites generally have olivine-controlled glass compositions. I shall show that this is far from the case.

All glass analyses for mid-ocean-ridge basalts (MORB) fall on a cotectic. Presnall & Gudfinnsson (2008) emphasize the dearth of olivine-controlled glasses on spreading ridges and Iceland, even questioning whether picrites from Siqueiros Fracture Zone (Natland, 1989; Perfit et al., 1996) have quenched glasses that are olivine-controlled. Falloon et al. (2008), on the other hand, argue that there are at least some olivine-controlled MORB glasses including those from Siqueiros Fracture Zone.

The Siqueiros picrites cannot be discussed without recourse to petrography, which in my view is decisive (Natland, 2007a,b). Danyushevsky et al.  (2003) and Falloon et al. (2007a; 2008) assert that on eruption and quenching, this glass was crystallizing only olivine. Nevertheless, both olivine dendrites and plagioclase spherulites occur in the quenched margin, plus plagioclase spherulites in embayments of olivine phenocrysts (Natland, 1980b; 1989; Perfit et al., 1996). By analogy to programmed-cooling experiments (Kirkpatrick, 1979) the glass must be saturated in both minerals regardless of whether olivine is the only silicate phenocryst. Compositions of spinel, variously included in olivine and occurring as single crystals in glass, strongly support a mixing history among disparate primitive magma stems in the basalt (Natland, 1989), thus neither the spinel nor the olivine enclosing it can represent a single liquid line of descent.  This is evident in any case from data and discussion of Danhushevsky et al. (2002). More than one picrite type was obtained in the same dredge haul at Siqueiros Fracture Zone, and they are not magmatically related, even by mixing (Natland, 1989). Why has only one of them been selected as “the” ultimate parental MORB magma along the East Pacific Rise?  Instead, the several picrites and glass inclusions within these and plagioclase-phyric basalts dredged nearby likely indicate bulk lithologic heterogeneity of the mantle within the same melting domain (Natland, 1989). Despite theoretical or experimental approaches that are based on the assumption of a single, olivine-controlled liquid line of descent, the relationship of the most magnesian olivine to the host glass in the one picrite remains uncertain.

A cotectic trend thus is easy to spot from trends on variation diagrams, and if a picrite has a glass composition that lies along it, this means it cannot be used to backtrack to a parental value for MgO contents, or to eruptive or potential temperatures using FeO-MgO olivine-glass equilibria. The backtracked calculated composition will necessarily have higher FeO contents than it should for this purpose. Judging this should be standard operating procedure before embarking on a backtrack calculation.

I say, then, that picrites from islands and seamounts are usually mixes between differentiated liquids on a low-pressure, low-temperature cotectic, and one or more sludges of magma containing a high percentage of olivine phenocrysts (cf., Clague et al., 1995; Natland, 2003b; 2007a,b). The olivine in the sludges may be so abundant because of mechanical mechanisms of crystal sorting such as flowage differentiation (Drever & Johnston, 1958; Simkin, 1967; Komar, 1972, 1976; Natland, 2007a,b). The olivine, too, usually has a differentiated composition (Fo86-70); primitive olivine (Fo91) is only present in rare samples, even though all of the picrites from a given locality with their own iron-rich olivine phenocrysts are typically used in inferring the presence of “olivine-controlled” liquids. But it does not matter what the composition of olivine happens to be – that is governed merely by some happenstance of geology related to the detailed workings of an active magma plumbing system. If the host groundmass or liquid composition, or any mixing aliquot of it, was along the cotectic, any backtrack calculation must involve more than just olivine and cannot be uniquely determined.

Exhibiting Olivine and Other Controls on CaO-Al2O3 Variation Diagram

The simplest way to evaluate the geochemistry is to use a diagram that is sensitive to addition/subtraction of phases other than olivine. The usual FeO-MgO diagram is not always by itself sufficient. A plot of CaO vs. Al2O3 is effective (Figure 2). The advantage of this diagram is that olivine plots virtually at the origin (almost no Al2O3 or CaO is present in olivine). Thus a nominal "olivine-controlled" set of compositions, whether of glasses or whole-rocks, should plot along a trend pointing toward the origin. This is observed, for example in the olivine-controlled glasses dredged from Kilauea-Puna Ridge (Clague et al., 1995), as it is on other variation diagrams.

Figure 2: Al2O3 versus CaO. Symbols are from Workman et al. (2004). A. Eastern Samoa (Muli, Malumalu, and Vailulu`u Seamounts and submerged and emergent Ta`u volcano. Red dots = whole-rock analyses of samples with glass analyses; inverted purple triangles = samples described as having clinopyroxene in addition to olivine phenocrysts; blue dots = glasses; large black dots = no petrographic information. Small black dots = older analyses (GeoRoc data base). Orange triangle = calculated Samoan parent from Putirka (2008). Additional symbols are for clinopyroxene in Type 2 magmatic xenoliths (green triangles) and phenocrysts from Upolu (blue diamonds) and Tutuila (black circles). Xenolith data are from Wright (1986, 1987), Dieu (1995) and J. Natland (unpublished). A curving plagioclase-clinopyroxene differentiation control curve is shown, the curving indicating different proportions of the two minerals (more plagioclase) as differentiation proceeds. An add olivine control line extends toward the origin. The two diagonal lines give possible olivine + clinopyroxene control trends from either end of the array of glass compositions. B. Filled circles = lavas of the Masefau basalt-dike complex (Stearns, 1944); open squares = basalts of the Greater Pago volcano of Natland (2003a; the Taputapu, Pago, Pago-intracaldera, Alofau and Olomoana volcanoes of Stearns, 1944). Xenoliths are as in A. C. Blue diamonds = Fagaloa shield volcano; Xenoliths are as in A.

On the other hand, clinopyroxene, the usual second silicate mineral to form in many seamount/island basalts, plots at high CaO and moderately high Al2O3. Trends of basalt or liquid compositions exhibiting strong clinopyroxene control on the liquid line of descent thus should plot along trends that move away from the clinopyroxene field for phenocrysts as differentiation involving clinopyroxene proceeds, or the trends should lie between olivine at the origin (Figure 2, lower left) and clinopyroxene (Figure 2, mid- to lower-right).

Figure 2A, in which I plot lava compositions from eastern Samoa (Stice, 1969; Farley et al., 1992; Workman et al., 2004), illustrates the contrasting trends of differentiation well. These data include compositions that fall along a nominal olivine-control line pointing toward the origin (add-olivine, in Figure 2A), and another trend among basalt liquids almost orthogonal to this – the "Plag-CPX Control" line, which applies to rocks with <7% MgO contents. The latter are not "olivine controlled" but their trend corresponds to the low- to intermediate-pressure cotectic between plagioclase and clinopyroxene (shallow "gabbro" fractionation, as first applied to Reykjanes Ridge near Iceland by Schilling, 1973).

Figure 2, A-C, show data from Eastern Samoa (Ta`u Island, and Vailulu`u, Muli and Malumalu Seamounts), Tutuila, and the Fagaloa volcano of Upolu, respectively, which are progressively further west along the Samoan chain. Only Eastern Samoa has samples that plot along a potential "add-olivine" control line. These are bulk-rock samples. Tutuila and Upolu, which have another trend, can thus immediately be excluded from further consideration. All existing analyses of the rocks from these islands are multiply saturated in at least olivine and clinopyroxene (Figure 1), and thus any olivine-control-related assumptions do not apply. The "picrites", so called, are what Macdonald (1949) termed "picritic basalts of the ankaramite type" and all of them contain clinopyroxene. That is, in thin section they can be seen to contain abundant titanaugite phenocrysts along with olivine (Figure 1). No potential temperatures can be calculated from these rocks by an olivine backtrack procedure. Some proportion or aliquot of the rocks must have crystallized along the low-pressure cotectic, and host melt compositions would likely fall along a low-pressure cotectic, were glass analyses available. Unfortunately, for Tutuila and Upolu, there are no analyzed glasses. It does not matter when, or in what mixed fraction of a rock, clinopyroxene may have joined the liquidus. Its mere presence is sufficient to show that this happened, and that means there is no simple, let alone unique, way to backtrack to an initial parental composition using the bulk composition.

The minerals provide additional information on this score. At Samoa, the olivine is differentiated (Fo70-85); it contains titanian magnesiochromite with substantial iron as both FeO and Fe2O3; the clinopyroxene is titanaugite; it contains crystals of titanomagnetite. Individual samples usually contain more than one discrete population of spinel and clinopyroxene. These are unambiguously differentiated and hybrid rocks.

The only candidate for olivine control therefore is eastern Samoa, where some of the rocks may have only olivine phenocrysts. We shall test this. However, note that little difference exists for analyses from Eastern Samoa (Figure 2A) and basalts of either Tutuila (Figure 2B) or Upolu (Figure 2C). And for eastern Samoa, we need only evaluate those dozen or so samples that fall along the “Add olivine” control line in Figure 2A and are not so obviously, on this diagram, clinopyroxene-bearing. They are broken down by symbols as follows:

  • those in which petrographic information in papers listing the analysis states that there are both olivine and clinopyroxene phenocrysts in the rock (inverted purple triangles);
  • those with olivine phenocrysts only (red circles), but which have host cotectic (plagioclase-clinopyroxene) glass compositions (blue circles); and
  • those with only olivine phenocrysts but no corresponding glass analyses (black circles; four samples)

First note that all eastern Samoan glass samples (blue circles) and most whole-rock samples (small black dots) fall on or near the curving plagioclase-clinopyroxene control arrow. This means that no olivine-controlled Samoan glasses are known, even among bulk rocks, which fall on or near to the “Add-olivine” line.

This leaves just four porphyritic samples that cannot be evaluated using this diagram, and these, to represent olivine-controlled liquids, must lie significantly outside the domain of all known Samoan glass compositions. However, all glasses that are hosts for whole-rock compositions that plot along the "Add olivine" line in Figure 2A have cotectic compositions (blue circles); that is, bulk rocks with these known glass compositions are not on “add olivine” lines. Instead, the bulk rocks plot toward higher MgO because of addition of clinopyroxene and olivine together along a general trend bounded by the two diagonal lines in Figure 2A, in typical proportions of about 4:1 or 3:1 olivine:clinopyroxene. So, too, then, could the four similar black-dot samples for which we have no glass compositions.

TiO2 versus CaO and Clinopyroxene Crystallization

To examine the influence of clinopyroxene on these four remaining samples, CaO is plotted vs. TiO2 for Eastern Samoan lavas and glasses in Figure 3. In this diagram, olivine would plot at the origin down and to the lower left, and off the figure, and clinopyroxene off the figure to the upper left. The olivine could have any Fo value, since CaO and TiO2 are both very low in any olivine. Any liquid may have olivine added to it, as indicated by two "Add olivine" lines, one at either end of the data array for plagioclase-clinopyroxene cotectic liquids. Truly primitive mixing liquids, or a liquid line of descent from a very primitive starting composition, would lie along or near the left "Add-olivine" line. Olivine of any composition added to a more differentiated liquid composition would follow some "Add-olivine" line with a lower slope to the right. Bulk compositions with olivine added to some intermediate cotectic liquid composition would fall in between the two lines shown. Most of the samples in question do. The light blue field indicates these. Three of the four black-dot samples are in this field. None of these are suitable for backtrack calculations.

Figure 3: TiO2 versus CaO contents of lavas from Eastern Samoa. Lava symbols as in Figure 1A. Control and mixing lines are explained in the text. One picritic sample, VAI 71-22 has both a whole-rock and glass analysis; the red line connecting the two indicates that it contains both olivine and clinopyroxene phenocrysts in ~3:1 proportion. Field 1, light blue, bounds samples that either contain both phenocrysts that mixed with cotectic glasses (an array given approximately by the dashed arrow), or which can have either olivine or olivine plus clinopyroxene added to such differentiated compositions (black dots).

An unusual trend exists at the high TiO2 end of this diagram. Most glass samples there form an array that points back obliquely to lower TiO2 (blue arrow labeled “End-member mixing”). These are liquids (glasses) or aphyric whole-rock compositions that appear to be hybrids combining something with very high TiO2 and primitive picrite, with or without augite. This may represent “end-member mixing”, and be related to the physical isolation of magma bodies undergoing extreme differentiation, and their later injection with something straight out of the mantle, or nearly so. Samoan magma chambers were not very big, and the rate of eruption probably fairly low, to allow so much differentiation to occur between injections of primitive magma.

The black-dot samples that plot in the light blue field are probably hybrids between intermediate cotectic liquid and crystals of olivine, perhaps scavenged from sludges somewhere beneath the volcanoes, as inferred for picritic samples from Kilauea-Puna Ridge (Clague et al., 1995). This reduces to one the number of samples that could potentially be an olivine-controlled liquid. This, at the lower left of the diagram, is a primitive composition from the active seamount, Vailulu`u, VAI 68-11. Since it is a primitive end-member in both Figures 1A and 2, the identity of differentiated liquids that may have mixed with it cannot be determined; it could be the most primitive composition of all along the left add-olivine line, or it could lie within an extension of the light blue field and have an intermediate cotectic glass composition.

Matthew Jackson (MIT/Woods Hole Oceanographic Institution), who has studied these lavas, sent me slides of three eastern Samoan picrites (personal communication, 2008). Two of these turn out to have some clinopyroxene phenocrysts. The other, VAI 63-11, which has not been analyzed, contains fragments of strained dunite, a type of deformed cumulate. All three thin sections are thus of mixed or hybrid rocks, the dunite being a type of xenolith. VAI 63-11 has enough dunite and olivine phenocrysts so that it could plot about where VAI 68-11 plots. Nevertheless, it is still a mixed rock. Whether or not all the olivine in VAI 68-11 crystallized along a single liquid line of descent is not yet established, but if it is, the sample would uniquely not be a mixed rock among Samoan picrites, and the only one suitable for FeO-MgO olivine-liquid calculation of a parental composition.

FeO-MgO Relationships Reconsidered

In Figure 4, MgO is plotted versus FeO for the same three localities as in Figure 2, eastern Samoa, Tutuila and Upolu. Here, FeO is taken to be 0.9 x FeOT (total iron as FeO; Herzberg & O’Hara, 2002), to allow for the likely oxidation state of the rock. Regardless, at all three localities, basalts have a range of FeO values at given MgO contents. The samples with higher FeO are the ones used to weight regressions and estimate or reconstruct “mantle” conditions (e.g., Putirka, 2008). In Figure 3A, for Eastern Samoa, the same symbols are used as previously. The range of the “Plag-cpx" cotectic liquids (steep, heavy arrow) spans the range of FeO contents of the “picrites”. This high FeO is still taken by Putirka (2008) to represent a liquid line of descent derived from a nominal mantle with Fo91 olivine. For Samoa the orange star (Samoa parent) shows this in Figure 4A. Total iron as FeOT in this estimated parent is 12%. This is an artificially high value, since everything that went into its estimation is a hybrid rock.

Figure 4: MgO versus FeO (= 0.9* FeOT = total iron as FeO) for A. Eastern Samoa; B. Tutuila; and C. Upolu. Symbols are as in Figure 1. In A, samples from Vailulu`u seamount are highlighted by the light blue field, which is repeated in B and C. Fine arrows indicate trends of addition of olivine phenocrysts with average compositions of Fo84 (upper ) and Fo86 (lower). Cotectic differentiation of plagioclase and clinopyroxene is shown by the bold arrow. The average Hawaiian calculated liquid line of descent (Putirka, 2008) is shown in B and C.

Alternatively, I argue that the “picrites” are differentiated liquids to which olivine and augite have been added by mixing. They have high FeO only because the host liquids are not in equilibrium with any of the olivine, and instead are differentiated cotectic liquids. Alternatively, I argue that the “picrites” are differentiated liquids to which olivine and augite have been added by mixing. They have high FeO only because the host liquids are not in equilibrium with any of the olivine, and instead are differentiated cotectic liquids (blue dots in Figure 4a). Eruptive temperatures (Beattie, 1993) calculated for these glasses are low, 1151ºC ± 21° (1118° – 1228°C; n = 96). The values of “average olivine” that need to be added are shown for two mixing arrows in Figure 4A. These are not very primitive olivine compositions, but are in accord with electron probe determinations of the actual olivine in the picrites (Matthew Jackson & Stan Hart, personal communication, 2007), and also with data on picritic basalts of the ankaramite type from Tutuila and Upolu. They have compositions ranging from Fo86-90. This range of olivine compositions supports the idea that the parent for Vailulu`u seamount (samples enclosed in the light blue field) is a hybrid rock as just described, with Fo86-87 being the likely composition of average “added olivine” to a host liquid with about 7% MgO and 9.5% FeO (Figure 3A). This has total iron as FeOT of 10.55%, nearly 1.5% less than the Samoan “parent” suggested by Putirka (2008).

Why is Iron Low at Vailulu`u?

Why is Vailulu`u so much lower in FeO? It may have to do with it being a small volcano with, as yet, an insignificant amount of cotectic differentiated material in the magma column. Ta`u, on the other hand, is both bigger and older, and thus has a lot of cotectic material in its magma column. Further west, Ofu Island, which is now extinct, may have had even more differentiated magmas its upper reaches when it was active (Natland, 2008). The system, as it develops, thus mulches on itself, and develops extensive and extended high-iron differentiation near its summit over time. Primitive melts derived from the mantle necessarily interact with this material in the magma column, and wind up having enhanced total iron as FeOT as a result of mixing. Such mixing has not occurred yet at Vailulu`u.

Vailulu`u is thus a better indication of temperature at Samoa than anywehere else, and it must be at least 100-150°C lower than claimed by Putirka (2008) for Eastern Samoa. However, this is still guesswork because none of these samples represent olivine-controlled liquids.

The Source of ~Fo91

The most magnesian olivine found at Vailulu`u is Fo90.2 (Matthew Jackson, personal communication, 2008). What is its source? Phenocrysts of >Fo88 are rare among Samoan picrites. It is likely that differentiation begins in the mantle and proceeds to shallow levels. Thus, it is important to consider ultramafic xenoliths, where such early differentiation may be recorded. This has been done for xenolith suites from Tutuila, Upolu, and Savai`i (Natland, 2008), of which there are two types (Figure 5):

  • (Type 1) residual xenoliths, representing material, mainly harzburgite, from which basalt was extracted at some point, and;
  • (Type 2) high-P magmatic segregations or cumulates, chiefly belonging to a series comprising dunite, wehrlite, clinopyroxenite and rare harzburgite. The clinopyroxene is green, not pink or brown, so it is not titanaugite.


Figure 5: Olivine Fo versus NiO content from Samoan ultramafic xenoliths. Type 1 = residual harzburgites, Type 2 = magmatic or cumulus xenoliths. Data are from Wright (1986, 1987), Dieu (1995) and J. Natland unpublished. The nominal mantle value of olivine in equilibrium with the Samoan parent of Putirka (2008) is Fo91, given by the vertical line. This is more magnesian than any known magmatic xenolith, and not as magnesian as the typical or average olivine in residual harzburgite that underlies most of the Samoan chain.

Although the olivine compositions in Type 2 xenoliths extend back to most magnesian compositions of about Fo90, even among such xenoliths these are rare. In addition, almost all Type 2 xenoliths attest to the early and high-P cotectic segregation of olivine and clinopyroxene together, which violates the assumption of olivine control required for the model of Putirka (2008) and studies based on similar approaches. The overwhelming majority of olivine crystals having Fo90-92 compositions, however, are in Type 1 xenoliths.

Given the frequency with which xenocrysts and xenoliths, particularly dunite, occur in Samoan shield basalts, and without petrographic or mineralogic proof to the contrary, the Fo90.2 in the Vailulu`u sample could be a xenocryst scavenged from either type of xenolith. Furthermore, given the great frequency of mixing between low-P cotectic liquids and other magmas with abundant accumulated olivine and clinopyroxene, the likelihood is that the Fo90.2 in the one sample from Vailulu`u is not on a unique line of descent linking it and its host liquid. This could also be the case for any sample claimed to represent a picritic mantle precursor related by simple olivine control to all the other olivines in the same sample, and the host liquid composition.

Conclusion: Conditions for use of FeO-MgO olivine-liquid equilibria

The essential point is that mixing confounds this subject. The evidence for mixing at Samoa (and, indeed, at Hawaii and the Juan Fernandez Islands; Natland 2003b, 2007a,b) is so pervasive that it should be assumed to have occurred unless explicitly disproven before proceeding to model FeO-MgO olivine-liquid equilibrium. Backtracking is viable only if the following general conditions are met:

  • The rocks are picrites, not ankaramites; they have NO euhedral clinopyroxene or plagioclase phenocrysts;
  • All olivine is faceted (euhedral) the picrites do not contain olivine xenoclrysts or dunite indicative of a mixing history;
  • The olivine phenocrysts represent a single range of compositions rather than clumping into two or more discrete sets of composition;
  • The same applies to spinel in olivine, which is probably a better test;
  • A bona fide Fo91 phenocryst is found, not strongly zoned, euhedral, with an analyzed inclusion of the unmodified (or at least easily reconstructed) melt (glass) from which it precipitated;
  • Host groundmass or glass compositions do not lie along a plagioclase-clinopyroxene cotectic differentiation trend;
  • Proof (experimental or otherwise) that neither clinopyroxene nor plagioclase were in equilibrium with that crystal as well.

If these conditions are met, the technique is likely to be valid and parental MgO and Tp can be obtained. If any of these conditions fail, they cannot.



I thank Matthew Jackson and Stan Hart for providing thin sections and information on mineral compositions from eastern Samoa; Don Anderson, and Gillian Foulger for comments on an initial draft; Dean Presnall, Trevor Falloon, and Yaoling Niu for reviewing the manuscript; and Gill for shaping it up for the web page.


  • Beattie, P., 1993. Olivine-melt and orthopyroxene-melt equilibria: Cont. Min. Pet., 115, 103-111.
  • Breddam, K., 2002. Kistufell: Primitive melt from the Iceland mantle plume: J. Petrol., 43, 345-373.
  • Clague, D.A., Moore, J.G., Dixon, J.E., and Friesen, W.B., 1995. Petrology of submarine lavas from Kilauea’s Puna Ridge, Hawaii. J. Petrol., 36, 299-349.
  • Clague, D.A., Weber, W.S. and Dixon, J.E., 1991. Picritic glasses from Hawaii. Nature, 353, 553-556.
  • Daly, R.A., 1924. The geology of American Samoa. Carnegie Institution of Washington, Publication 340: 5-145.
  • Dieu, J., 1995. On the formation of cumulates, characteristics of oceanic lithosphere, and the process of carbonatite metasomatism; a study of ultramafic xenoliths from South Pacific Islands. University of California, San Diego (Ph.D. Dissertation), 393 pp.
  • Drever, H.I., & Johnston, R., 1958. The petrology of picritic rocks in minor intrusions – a Hebridean group. Trans. Royal Soc. Edinburgh, 63, 459-499.
  • Falloon, T.H., Green, D.H., Danyushevsky, L.V., and McNeill, A.W., 2008. The composition of near-solidus partial melts of fertile peridotite at 1 and 1.5 GPa: Implications for the petrogenesis of MORB. J. Petrol., 49: 591-613, doi:10.1093/petrology/egn009.
  • Farley, K.A., Natland, J.H., and Craig, H., 1992. Binary mixing of enriched and undegassed (primitive?) mantle components (He, Sr, Nd, Pb) in Samoan lavas. Earth Planet. Sci. Lett., 111, 183-199, doi:10.1016/0012-821X(92)90178-X.
  • Francis, D., 1985.The Baffin Bay lavas and the value of picrites as analogues of primary magmas. Cont. Min. Pet., 89: 144-154.
  • Gurenko, A.A., and Chaussidon, M., 1995. Enriched and depleted primitive melts included in olivine from Icelandic tholeiites: Origin by continuous melting of a single mantle column, Geochim. et Cosmochim. Acta, 59, 2905-2907, doi:10.1016/0016-7073(95)00184-0.
  • Hansteen, T.H., 1991. Multi-stage evolution of the picritic Maelifell rocks, SW Iceland: constraints from mineralogy and inclusions of glass and fluid in olivine: Cont. Min. Pet., 109, 225-239.
  • Hart, S.R., Coetzee, M., Workman, R.K., Blusztajn, J., Johnson, K.T.M., Sinton, J.M., Steniberger, B., and Hawkins, J.W., 2004. Earth Planet. Sci. Lett., 227, 37-56.
  • Helz, R.T., 1987. Diverse olivine types in lava of the 1959 eruption of Kilauea volcano and their bearing on eruption dynamics. In: Decker, R.W., Wright, T.L., & Stauffer, P.H. (eds), Volcanism in Hawaii, U.S. Geological Survey Professional Paper 1350. Washington: U.S. Government Printing Office, 691-722.
  • Herzberg, C., Asimow, P.D., Arndt, N., Niu, Y., Lesher, C.M., Fitton, J.G., Cheadle, M.J., and Saunders, A.D., 2007. Temperatures in ambient mantle and plumes: Constraints from basalts, picrites, and komatiites. Geochemistry, Geophysics, Geosystems, 8, Q02006, doi:10.1029/2006GC001390.
  • Herzberg, C. and O’Hara, M.J., 2002, Plume-associated magmas of Phanerozoic age, J. Petrol., 43, 1857-1883.
  • Jackson, M.G., Hart, S.R., Saal, A.E., Shimizu, N ., Kurz, M.D., Blustajn, J.S. and Skovgaard, A.C., 2008. Globally elevated titanium, tantalum and niobium (TITAN) in ocean island basalts with high 3He/4He, Geochemistry, Geophysics, Geosystems, 9: doi:10.1029./2007GC001876, 1-21.
  • Kear, D., and Wood, B.L., 1959. The geology and hydrology of Western Samoa. N. Zeal. Geol. Surv. Bull., 63, 92 p.
  • Kirkpatrick, R.J., 1979. Processes of crystallization in pillow basalts, Hole 396B, DSDP Leg 46. In: Dmitriev, L., Heirtzler, J., et al., Initial Reports of the Deep Sea Drilling Project 46. Washington: U.S. Government. Printing Office, 272-282.
  • Komar, P.D., 1972. Mechanical interactions of phenocrysts and flow differentiation of igneous dikes and sills. Geol. Soc. Am. Bull., 83, 973-988.
  • Larsen, L.M., and Pedersen, A.K, 2000. Processes in high-Mg, high-T magmas: Evidence from olivine, chromite and glass in Paleogene picrites from West Greenland, J. Petrol., 41, 1071-1098.
  • Macdonald, G.A., 1949. Hawaiian petrographic province, Bull. Geol. Soc. Am., 6, 1541-1596.
  • Natland, J.H., 1980a. Progression of volcanism in the Samoan linear volcanic chain. In: Irving, A.J. and Dungan, M. (eds.), The Jackson Volume, Amer. J. Sci., 280-A, 709-735.
  • Natland, J.H., 1980b. Crystal morphologies in basalts dredged and drilled from the East Pacific Rise near 9°N and the Siqueiros Fracture Zone. In Rosendahl, B.R., Hékinian, R., et al., Initial Reports of the Deep Sea Drilling Project: Washington (U.S. Government Printing Office), 605-633.
  • Natland, J.H., 1989. Partial melting of a lithologically heterogeneous mantle: Inferences from crystallization histories of magnesian abyssal tholeiites from the Siqueiros Fracture Zone. In: Saunders, A.D. and Norry, M. (eds.), Magmatism in the Ocean Basins, Geol. Soc. London, Spec. Publ., 42, 41-77
  • Natland, J.H., 2003b. Capture of mantle helium by growing olivine phenocrysts in picritic basalts from the Juan Fernandez Islands, SE Pacific. J. Petrol. 44, 421-456.
  • Natland, J.H., 2008. Will the real Fo91 please stand up?, in preparation.
  • O’Hara, M.J., 1968. The bearing of phase equilibria studies on the origin and evolution of igneous rocks. Earth-Science Reviews, 4, 6-133.
  • Perfit, M.R., Fornari, D.J., Ridley, W.I., Kirk, P.D., Casey, J., Kastens, K.A., Reynolds, J.R., Edwards, M., Desonie, D., Shuster, R., and Paradis, S., 1996. Recent volcanism in the Siqueiros ransform fault: picritic basalts and implications for MORB genesis. Earth Planet. Sci. Lett., 141, 91-108.
  • Presnall, D. C., and Gudfinnsson, G.H., 2008. Origin of oceanic lithosphere. J. Petrol., 49, 615-632, doi:10.1093/petrology/egm052
  • Putirka, K., Perfit, M., Ryerson, F.J., and Jackson, M.G., 2007. Ambient and excess mantle temperatures, olivine thermometry and active vs. passive upwelling, Chem. Geol., 241, 177-206, doi: 10.1016/j.chemgeo.2007.01.014.
  • Robillard, I., Francis, D., and Ludden, J.N., 1992. The relationship between E= and N-type magmas in the Baffin Bay lavas. Cont. Min. Pet., 112: 230-241.
  • Roeder, P.L., and Emslie, R.F., 1970. Olivine-liquid equilibria. Cont. Min. Pet., 29, 275-289.
  • Schilling, J.-G., 1973a, Iceland mantle plume - geochemical study of Reykjanes ridge, Nature, 242, 565-571.
  • Simkin, 1967. Flow differentiation in the picritic sills of north Skye, In Wyllie, P.J. (Ed.), Ultramafic and Related Rocks, New York (John Wiley & Sons), 64-69.
  • Stearns, H.T., 1944. Geology of the Samoan Islands. Geol. Soc. Am. Bull., 55, 1279-1332
  • Stice, G.D., 1968. Petrography of the Manu`a Islands, Samoa, Cont. Min. Pet., 19, 343-357.
  • Workman, R.K., Hart, S.R., Jackson, M.G., Regelous, M., Farley, K.A., Blustajn, J, Kurz, M., and Staudigel, H., 2004. Recycled metasomatized lithosphere as the origin of the Enriched Mantle II (EM2) end-member: Evidence from the Samoan volcanic chain, Geochemistry, Geophysics, Geosystems, 5, Q04008, doi:10.1029/2003GC000623.
  • Wright, E., 1986. Petrology and geochemistry of shield-building and post-erosional lava series of Samoa; implications for mantle heterogeneity and magma genesis. University of California, San Diego (Ph.D. Dissertation), 305 pp.
  • Wright, E., 1987. Mineralogical studies of Samoan ultramafic xenoliths: implications for upper mantle processes. In Fryer, P., Batiza, R., and Boehlert, G.W. (eds.), Seamounts, Islands, and Atolls, Geophysical Monograph, 43: Washington (American Geophysical Union), 221-234.
last updated 1st September, 2008