Vaugnerite.pdf

Earth and Environmental Science Transactions of the Royal Society of Edinburgh http://journals.cambridge.org/TRE Additional services for Earth

and Environmental Science Transactions of the Royal Society of Edinburgh: Email alerts: Click here Subscriptions: Click here Commercial reprints: Click here Terms of use : Click here

Shoshonites, vaugnerites and potassic lamprophyres: similarities and differences between ‘ultra’-high-K rocks J. H. Scarrow, F. Bea, P. Montero and J. F. Molina Earth and Environmental Science Transactions of the Royal Society of Edinburgh / Volume 99 / Issue 3-4 / December 2008, pp 159 - 175 DOI: 10.1017/S1755691009008032, Published online: 06 November 2009

Link to this article: http://journals.cambridge.org/abstract_S1755691009008032 How to cite this article: J. H. Scarrow, F. Bea, P. Montero and J. F. Molina (2008). Shoshonites, vaugnerites and potassic lamprophyres: similarities and differences between ‘ultra’-high-K rocks. Earth and Environmental Science Transactions of the Royal Society of Edinburgh, 99, pp 159-175 doi:10.1017/S1755691009008032 Request Permissions : Click here

Downloaded from http://journals.cambridge.org/TRE, IP address: 128.6.218.72 on 13 Apr 2015

Earth and Environmental Science Transactions of the Royal Society of Edinburgh, 99, 159–175, 2009 (for 2008)

Shoshonites, vaugnerites and potassic lamprophyres: similarities and differences between ‘ultra’-high-K rocks J. H. Scarrow, F. Bea, P. Montero and J. F. Molina Department of Mineralogy and Petrology, Campus Fuentenueva, University of Granada, 18002 Granada, Spain Email: [email protected] ABSTRACT: A comparative study of three main igneous rock associations that plot in the K2O–SiO2 diagram shoshonite field: shoshonite series absarokites–shoshonites–banakites (henceforth referred to as shoshonites s.l.), vaugnerites, and potassic lamprophyres, reveals that similarities between the associations are superficial. Vaugnerites and lamprophyres are more magnesian, richer in large ion lithophile and high field strength elements and have higher light rare earth/heavy rare earth ratios than shoshonites. Furthermore, shoshonites have low radiogenic heat production, typical of subduction-related rocks, but most vaugnerites and some lamprophyres are highly radioactive. Relative to bulk-Earth, shoshonites have depleted, asthenospheric mantle-like Sr and Nd isotope signatures, whereas vaugnerites and potassic lamprophyres have enriched, crust or lithospheric mantle-like compositions. Though vaugnerites and some lamprophyres show evidence of crustal contamination, the contaminated magma was not originally shoshonitic. Their composition is consistent with derivation from a metasomatised upper mantle source enriched long before melting, thus precluding an active subduction setting. In conclusion, the term shoshonite, implying late-stage arc magmas, cannot be applied to a rock series simply because it plots into the K2O–SiO2 diagram shoshonite field. Shoshonites with a subduction-related source may, however, be identified by discriminant function analysis. KEY WORDS: Discriminant function analysis, late-orogenic magmatism, rock classification, subduction, Variscan Recent IUGS subcommission recommendations (Le Maitre et al. 2002) restrict the use of shoshonite to potassic varieties of rocks which project in the basaltic trachyandesite field of the total alkalis versus SiO2 (TAS) diagram. Originally defined as potassic orthoclase-bearing basalts in Yellowstone National Park, Wyoming, USA (Iddings 1895), the term has since been used, in a general sense, to refer to potassic basalts and andesites. Joplin (1968) defined the broader ‘shoshonite association’ as K-rich trachybasalts to trachyandesites (and associated intrusive rocks), of the absarokite (basic) – shoshonite (intermediate) – banakite (intermediate-acid) series. The term shoshonitic was subsequently adopted to define the most K-rich of the five broad types of subduction-zone primary magmas (Peccerillo & Taylor 1976) leading to its use, often erroneously, with a strong geodynamic connotation. Fuelling this, a specific link between the association and tectonic context was suggested by Morrison (1980) who proposed that such rocks were subduction-related, either late-stage, far-fromtrench, or direction-flip arc products. Latterly, the meaning of shoshonitic has further expanded to include other K-rich rocks not obviously linked to a subduction environment, such as the plutonic appinite and vaugnerite (biotite-, hornblende-, plagiocase-bearing rocks, Le Maitre et al. 2002) series, and the minette, vogesite, kersantite and spessartite ‘calc-alkaline’ lamprophyres (Rock 1987). The term shoshonitic, used in a more relaxed way still, is currently applied to most K-rich rocks, even including granites (e.g. De Lima & Nardi 1998; Liegeois et al. 1998; Yao-Hui et al. 2002; etc.). A case of classifying a rock series as shoshonitic, simply because it plots within the shoshonitic field of the K2O–SiO2  2009 The Royal Society of Edinburgh. doi:10.1017/S1755691009008032

diagram, and, either explicitly or implicitly, aiming to use this characterisation to demonstrate a subduction zone setting, and even to a definite stage of the evolution of a magmatic arc-system, is well illustrated in the Variscides of Western Europe. In this region, K- and Mg-rich ultramafic to intermediate rocks, temporally and spatially closely associated with lamprophyres, are often the only mantle-derived products associated with huge volumes of crustal granites. Most of these rocks belong either to the vaugnerite (durbachite) or the appinite series, both of which are considered to be shoshonitic by Bowes & Kosler (1993). Interpreting the geodynamic setting based on this consideration alone can lead to conflict with other lines of evidence. For example, the Central Iberian Zone of Spain and Portugal comprises numerous granite– granodiorite batholiths scattered across a 600km-wide band centred on the orogenic axis (Bea et al. 1999). The granitoids show no perceptible across-orogen polarity in either age or granite typology (Bea et al. 2003) and, irrespective of their geographic position, are composed of peraluminous K-rich crustal granites which mostly formed within the interval 31015 Ma. Furthermore, the area lacks high-pressure metamorphism and contains neither ophiolites nor any other subduction-related rocks of Variscan age. The mantle-derived rocks consist of volumetrically insignificant, metric- to hectometric-sized bodies of vaugnerites and appinites which are coeval with the granites either as enclaves or as small intrusions (Dias et al. 2002; Montero et al. 2004). From their position in the K2O–SiO2 diagram, these rocks are shoshonitic (vaugnerites) or high-K (appinites) (Bea 2004). Their N-MORB normalised trace-element patterns are similar to

160

J. H. SCARROW ET AL.

Figure 1 (A) K2O–SiO2 plot of shoshonites (Sh), vaugnerites (Vg) and potassic lamprophyres (K-La). (B) The same but with regression lines with 95% confidence interval (prediction) instead of individual points. The boundaries between the shoshonitic, here renamed ultra-high-K, high-K and medium-K fields, are as recommended by Tatsumi & Eggins (1995). Note how the potassic lamprophyres are, by far, the most K2O rich. Vaugnerites and shoshonites have similar K2O average values, but they change with increasing SiO2. Vaugnerites are the richest in K2O for SiO2 <57 wt. % but this tendency is attenuated, even reversed, at higher SiO2 contents. Data sources discussed in text.

arc-magmas, showing an enrichment in the most incompatible elements with deep negative anomalies of Nb-Ta, Ti and Zr. Consequently, they have repeatedly been interpreted as evidence of Variscan subduction beneath the Central Iberian Zone (e.g. Castro et al. 2003; Lo´pez-Moro & Lo´pez-Plaza 2004) despite this idea finding no support from either geophysical, geological, or structural data, nor from any other petrological feature of the area (e.g., Bea 2004; Scarrow et al. 2009; Molina et al. 2009). This paper is aimed at understanding whether it is possible to define a set of chemical features specific to rocks which fall in the shoshonitic field of the K2O–SiO2 diagram, used for the subdivision of the sub-alkaline rocks of the total alkalis versus SiO2 (TAS diagram, Le Bas et al. 1986), and that have a subduction-related source, be it active arc or recycled arc products. If this is possible, then rocks of unknown origin having the same chemical features as the ‘arc’ shoshonites can be deemed to have had a subduction-related source, whereas those that do not, despite being ‘ultra’-high-K, can be inferred to have had a different origin. To address this issue, the chemical composition (major and trace elements, Sr and Nd isotopes) of some 600+ selected specimens from three main igneous rock associations that plot in the shoshonitic field of the K2O–SiO2 diagram was compared: shoshonite series

absarokites–shoshonites–banakites (henceforth referred to as shoshonites s.l.); vaugnerites; and the potassic ‘calc-alkaline’ lamprophyres: vogesites and minettes. The IUGS Subcomission on Igneous Rock systematics (Le Maitre et al. 2002) defined two groups of lamprophyres: calc-alkaline and alkaline. The former are feldspathoid-free and richer in SiO2, Al2O3, and K2O, and poorer in Fe2O3, MgO, CaO and TiO2. Throughout the current work, the first group, of interest here, are referred to as potassic lamprophyres to avoid any potentially erroneous petrogenetic association with the term calc-alkaline. Appinites were not included because, at least in the West European Variscides, most of them plot in the high-K field, a typology for which Roberts & Clemens (1993) have already argued there is no requirement to be related to subduction processes. The compositional similarities and differences between the three aforementioned associations were studied statistically and the paper discusses whether vaugnerites and potassic lamprophyres may, like the shoshonites, also have a subduction-related source. A set of discriminant functions was also computed based on major elements and different combinations of frequently determined trace elements, which seem far more useful than binary or multi-element plots to identify true arc-related shoshonites.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

161

Table 1 Summary statistics of the chemical composition of the three groups. Shoshonites n

Mean

S.D.

Min

Vaugnerites Max

n

Mean

S.D.

Min

260 53·20 3·15 47·20 62·94 180 55·88 4·16 SiO2 TiO2 260 0·88 0·23 0·26 1·93 180 1·05 0·31 260 17·12 1·74 10·78 20·63 180 16·16 2·11 Al2O3 FeO tot 260 7·52 1·27 3·08 12·24 180 6·08 1·37 MgO 260 4·48 2·32 0·72 16·44 180 5·62 2·82 MnO 260 0·16 0·03 0·06 0·28 180 0·10 0·03 CaO 260 7·90 2·03 2·30 13·14 180 5·40 1·62 260 3·11 0·79 0·48 6·08 180 2·98 0·71 Na2O K 2O 260 3·41 0·99 1·70 6·69 180 4·00 1·23 260 0·50 0·15 0·15 1·09 180 0·53 0·25 P 2O 5 Li 4 12 1 10 13 67 88 44 Rb 217 96 42 24 313 133 178 63 Cs 50 1·68 2·05 0·24 9·43 88 14·07 5·87 Be 0 36 5·5 1·6 Sr 217 916 437 210 3060 133 532 247 Ba 182 1173 487 405 3030 133 1576 783 Sc 132 20 8 4 51 42 19 7 V 129 212 47 83 301 49 127 34 Cr 145 88 139 2 846 124 233 230 Co 114 28 11 10 78 121 23 13 Ni 189 39 62 2 450 120 74 73 Cu 33 123 94 22 420 94 29 44 Zn 33 96 56 33 368 109 82 39 Ga 14 21 1 18 23 52 21 10 Y 192 28 7 7 48 88 25 8 Nb 196 19 13 1 101 80 14 6 Ta 74 0·5 0·5 0·0 1·7 59 1·4 0·8 Zr 211 182 72 24 424 103 258 144 Hf 79 3·5 1·8 0·9 9·6 63 7·4 4·3 Pb 67 13 8 2 44 89 32 9 U 70 2·3 1·5 0·3 6·0 84 5·0 1·6 Th 94 7·6 5·3 0·8 19·0 105 24·7 9·5 La 141 50·3 36·8 5·6 206·0 86 70·7 33·8 Ce 176 92·4 52·9 11·4 322·0 86 135·3 64·4 Pr 16 10·7 3·1 6·1 15·4 50 17·6 6·2 Nd 116 37·1 16·0 8·9 91·0 69 59·8 22·1 Sm 84 6·4 2·5 2·4 13·0 73 10·5 3·4 Eu 79 1·76 0·54 0·80 3·25 73 2·02 0·79 Gd 20 6·84 1·25 4·11 8·93 55 7·43 1·84 Tb 77 0·84 0·25 0·39 1·37 71 1·05 0·32 Dy 16 5·74 1·05 3·84 7·62 63 4·84 1·15 Ho 30 1·20 0·24 0·80 1·79 50 0·88 0·23 Er 16 3·19 0·60 2·28 4·20 55 2·29 0·57 Tm 0 53 0·34 0·14 Yb 79 2·40 0·76 1·10 4·28 73 2·09 0·56 Lu 76 0·39 0·13 0·16 0·68 61 0·30 0·08 87 Sr/86Sri 55 0·7040 0·0005 0·7029 0·7061 24 0·7074 0·0004 28 4·1 0·9 2·1 5·7 14
4·0 0·7
(Nd)i

1. Samples This work used 260 samples of shoshonites, 180 samples of vaugnerites, and 196 samples of potassic lamprophyres, mainly vogesites and minettes. The shoshonites and lamprophyres come from all over the world and have ages from Archean (the lamprophyres) or Mesozoic (the shoshonites) to the present, but the vaugnerites are restricted to the West European Variscides and Caledonides (Supplementary Material). The data were collected from the literature, except for 60 samples

Potassic lamprophyres Max

n

45·10 62·99 196 0·32 2·28 196 8·53 20·09 196 1·52 11·01 196 0·92 19·90 196 0·00 0·20 196 0·37 9·78 196 0·55 5·65 196 1·70 8·50 196 0·04 1·90 196 28 264 9 37 473 159 1·60 38·60 45 3·0 8·5 2 10 1401 160 250 6410 166 6 42 102 69 216 125 21 1405 148 3 74 121 12 459 148 2 405 109 39 449 116 5 82 72 7 46 147 4 43 137 0·1 4·6 59 10 1199 158 0·8 28·8 69 6 53 77 2·9 10·8 117 1·5 48·7 131 10·0 163·2 140 22·0 296·0 137 7·1 31·0 44 8·0 114·0 122 5·1 22·1 103 0·75 5·10 101 4·36 11·80 58 0·37 2·50 80 2·43 7·69 70 0·42 1·51 48 1·18 3·89 53 0·18 1·10 47 0·26 3·67 100 0·01 0·49 100 0·7068 0·7086 26
5·6
2·9 27

Mean

S.D.

Min

Max

51·52 4·47 43·00 62·45 1·11 0·51 0·39 3·20 12·38 1·83 5·61 18·36 6·74 1·78 2·05 12·83 8·23 2·78 2·64 22·38 0·13 0·04 0·03 0·35 6·98 2·22 1·21 14·80 1·97 0·90 0·04 4·82 5·13 1·50 1·76 9·03 0·92 0·53 0·10 2·78 41 10 30 57 206 135 10 871 7·87 9·45 0·27 55·50 14·0 9·9 7·0 21·0 1267 1038 210 5433 2674 2212 209 14150 21 7 9 53 154 56 64 337 445 225 7 1140 33 16 0 128 209 133 14 700 54 42 2 170 97 85 13 851 17 4 6 25 22 9 7 55 20 16 3 76 1·1 1·2 0·2 5·0 358 214 51 925 9·4 6·2 1·6 25·0 36 28 6 158 5·8 4·9 1·0 23·6 22·9 24·0 1·0 220·0 82·7 82·8 19·8 753·0 157·7 114·0 43·0 550·0 15·8 10·4 5·8 48·7 82·1 54·3 7·4 244·0 14·7 8·4 4·7 48·8 4·47 5·89 1·10 39·00 8·70 4·06 1·39 21·00 0·99 0·39 0·54 2·32 4·64 1·52 2·57 12·00 0·80 0·28 0·43 2·00 2·03 0·55 1·13 4·30 0·27 0·09 0·14 0·71 1·68 0·43 0·80 3·00 0·26 0·19 0·11 2·00 0·7091 0·0035 0·7054 0·7206
10·4 4·9
19·4
1·0

of Iberian vaugnerites analysed in the authors’ laboratory. Samples were assigned to a given category according to the authors’ classification in the original publication. This can be summarised as follows: shoshonites, for the most part, were classified compositionally from their plotting position in the K2O vs SiO2 diagram of Pecerillo & Taylor (1976); vaugnerites, on the other hand, were classified exclusively petrographically, being phlogopite/biotite-, amphibole-, and alkali feldspar-bearing gabbros; potassic lamprophyre classification was mainly petrographic based on the criteria

162

J. H. SCARROW ET AL.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

163

Figure 3 FeOtotal–MgO plot. Note how vaugnerites and, especially, potassic lamprophyres are more magnesian, for a given FeO content, than shoshonites.

(abundant alkali feldspar, alkali-rich pyroxene and amphibole and globular, panidiomorphic textures) of Rock (1991). To be included, each analysis had to fulfil the following criteria: (1) the published totals had to be equal to the sum of the reported oxides to exclude typing errors; (2) K2O had to be higher than 1·5 wt. %; and (3) SiO2 had to be in the range 43–64 wt. % to exclude oddities and highly fractionated specimens. The data sources, summarised in three tables of Supplementary Material (Table 1E, 2E and 3E) are as follows: Shoshonites (Supplementary Table 1E) (mostly found in the GEOROC database, http://georoc.mpch-mainz.gwdg.de): Barberi et al. 1974; Beard & Johnson 1993, 1997; Beccaluva et al. 1985; Bonaccorso et al. 1996; Calanchi et al. 1983; Carmichael et al. 1996; Carr 1998; Conrey et al. 2001; Cox & Bristow 1984; De Astis et al. 2000; Del Moro et al. 1998; Dickinson et al. 1968; Ellam et al. 1988, 1989; Erlikh 1966; Esperanca et al. 1992; Feeley et al. 2002; Francalanci 1989; Francalanci et al. 1989; Gauthier & Condomines 1999; Gibson et al. 1991, 1992; Hogg 1972; Hornig-Kjarsgaard et al. 1993; Jakes & White 1969; Jacques 1976; Keller 1974; Kepezhinskas et al. 1997; Kesson & Smith 1972; Lacroix 1928; Mackenzie 1976; Mackenzie & Chappell 1972; Rogers & Setterfield 1994; Schellekens 1988; Skipp & McGrew 1977; Sun & Stern 2001; Tsvetkov et al. 1985; Xu 1988. Vaugnerites (Supplementary Table 2E): Bowes & Kosler 1993; Buda & Dobosi 2004; Debon et al. 1998; Ferre & Leake 2001; Galan et al. 1997; Gallastegui 1993; Janousek et al. 1995, 2000;

Klotzli et al. 2004; Michon 1987; Rossi & Cocherie 1991; Sabatier 1978, 1980, 1991; Washington 1917; Wenzel et al. 2000. Potassic lamprophyres (Supplementary Table 3E): Allen & Balk 1954; Ashley et al. 1994; Bor-ming et al. 1979; Buhlmann et al. 2000; Canning et al. 1996; Carlier et al. 1997; Carmichael et al. 1996; Greenough et al. 1993; Hoch et al. 2001; Huang et al. 2002; Jahn et al. 1979; Janousek et al. 1995, 2000; Kapp 1960; Kemp & Billingsley 1921; Knopf 1936; Loughlin 1919; Michaels 1969; Nemec 1973; Peterson et al. 2002; Pirsson 1905; Prelevic et al. 2004; Righter & Rosas-Elguera 2001; Rittman 1940; Schmidtt et al. 1974; Searle et al. 1992; Sheppard & Taylor 1992; Tingey et al. 1991; Wagner & Velde 1993; Wenzel et al. 2000; Wierzcholowski 2003; Williams 1936; Witkind 1973; Wyman & Kerrich 1993.

2. Comparative geochemistry Shoshonites, vaugnerites and potassic lamprophyres bear undeniable chemical similarities. Most of them plot in the shoshonitic field of the K2O–SiO2 diagram (Fig. 1a). The averages of the three groups (Table 1) yield sub-parallel N-MORB normalised trace-element patterns (Fig. 2a, b, c) with a considerable enrichment in the most incompatible elements, deep negative Nb and Ti anomalies, strong positive Pb and K anomalies and LREE>HREE (Fig. 2d, e, f), a set of

Figure 2 (A, B, C) N-MORB normalised multi-element plot of the average of the shoshonites (Sh), vaugnerites (Vg) and potassic lamprophyres (K-La). Data sources discussed in text (see also Supplementary Material Tables 1E, 2E and 3E). Grey lines mark 1 standard deviation for each element, arrows indicate values out of range. The grey field in (B) and (C) is the shoshonite field from (A). All have features considered characteristic of subduction magmas such as the enrichment in the most incompatible elements, negative anomalies of Nb and Ti, and positive anomalies of Pb. Both vaugnerites and potassic lamprophyres are notably richer than shoshonites in the most incompatible elements. Note also the strong enrichment in Cs (and Li, see Table 1) and the negative anomalies in Sr and Ba of the vaugnerites. (D, E, F) Chondrite-normalised REE patterns of the averages of the three groups. Grey lines mark 1 standard deviation for each element. The grey field in (E) and (F) is the shoshonite field from (D). Note how vaugnerites and potassic lamprophyres are enriched in LREE but slightly depleted in HREE with respect to shoshonites. Data sources discussed in text. (G) Average potassic lamprophyre and vaugnerite normalised to average shoshonite.

164

J. H. SCARROW ET AL.

features that are characteristic of subduction-related magmas. Figure 2g summarises the differences in trace elements between the shoshonites and vaugnerites and potassic lamprophyres. The similarities between the three groups, however, finish there. For a given level of evolution, the vaugnerites and potassic lamprophyres are far more magnesian than the shoshonites (Fig. 3, Table 1), so that the average Mg number (Mg#=100 molar Mg/Mg+Fe) decreases from Mg# 716 for the potassic lamprophyres to Mg# 628 for the vaugnerites and Mg# 519 for the shoshonites. Notwithstanding a broad overlap of Na2O contents in the three groups, it is notable that for a given SiO2 content, the first extend to higher sodium contents and the last to lower sodium contents (Fig. 4a). This distribution is also observed for V (not shown). Conversely, albeit again with considerable overlap between the rock types, relative to the shoshonites the vaugnerites and, even more so, the potassic lamprophyres, scatter to higher concentrations of P2O5 (Fig. 4b), Zr (Fig. 4c), Th (Fig. 4d), Cs (Fig. 4e), Cr (Fig. 4f) and Ni (Fig. 4g). Especially relevant are the differences in radiogenic heat production between the three rock types, calculated from the abundances of K, Th and U (Fig. 5). Shoshonites yielded an average of 1·3 w m
3 s
1 with a mode of 0·7, vaugnerites and potassic lamprophyres are notably more radioactive and yield identical averages of 3·6 w m
3 s
1 though with different distribution patterns, nearly symmetrical in the vaugnerites but strongly tailed to high values in the potassic lamprophyres, the mode of which is 1·7 w m
3 s
1. The meaning of these figures becomes clear when the following points are considered (i) the average heat production of the subduction-related rocks of the Urals is 0·7 w m
3 s
1, akin to the shoshonites; (ii) the giant agpaitic massif of Khibina, Kola, which derived from a very enriched mantle source, produces about 1·7 w m
3 s
1, similar to the mode of the lamprophyres; and (iii) the crust-derived Spanish S-type leucogranites, locally related to uranium deposits, yielded 3·1 w m
3 s
1, somewhat less than the average of vaugnerites (authors’ unpublished data). In the opinion of the present authors the enrichment in heat-producing elements shown by the potassic lamprophyres and vaugnerites is most easily explained by the involvement of U–Th–K-rich continental crust, something which the shoshonites do not require. Supporting this hypothesis, is the distribution of Cs (Fig. 4e), an element extremely partitioned into crustal micas, which reaches concentrations more characteristic of evolved peraluminous granites in the vaugnerites than in either of the other rock types. Moreover, the three rock associations have elevated, sometimes very high, concentrations of Ba and Sr which reach a maximum in some lamprophyres (Table 1). It should be emphasised, nonetheless, that in an N-MORB normalised trace element plot, the average vaugnerite has small but perceptible Ba and Sr negative anomalies relative to adjacent elements (Fig. 2c); this would be an unusual feature for subduction-related magmas. Remarkably, whereas a

Figure 4 (A) Na2O–SiO2 plot. Shoshonites extend to somewhat higher values of Na2O than the vaugnerites and both these are richer in this element than the lamprophyres. (B) P2O5–SiO2 plot. Some of the potassic lamprophyres are far richer in P2O5 than shoshonites or vaugnerites. (C) Zr-SiO2 plot. Note how some of the potassic lamprophyres have higher values of Zr. (D) Th-SiO2 plot. The potassic lamprophyres and the vaugnerites extend to higher Th values than the shoshonites. (E) Cs–SiO2 plot. Vaugnerites extend to higher values of Cs than the shoshonites and most lamprophyres, so reflecting their greater crustal contribution. See text for details. (F) Cr–SiO2 plot. (G) Ni–SiO2 plot. Vaugnerites and lamprophyres extend to higher values of Cr and Ni than the shoshonites.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

165

Figure 5 Histograms of the radiogenic heat production of the three group associations. Heat production was calculated from the element concentrations of K, U and Th, according to the Schmucker (1969, p. 197) expression. Vertically dashed lines labelled as S, A, and SG represent the averages of subduction-related rocks of the Urals (S), the giant alkaline agpaitic massif of Khibina, Kola (A), and the U deposit-related S-type leucogranites of Spain (SG) (authors’ unpublished data). The high heat production of vaugnerites and some potassic lamprophyres are indicative of a marked crustal component. See text for details.

significant fraction of shoshonites have Ba/Sr<1, most vaugnerites and lamprophyres have Ba/SrZ1. The element ratios showing the sharpest differences between the three rock associations, however, are as follows: La/Yb changes from z14 in shoshonites to z38 in vaugnerites and z56 in lamprophyres (Fig. 6a); Th/U is z3·2, close to the asthenospheric mantle ratio, in the shoshonites. Although some of the vaugnerites and potassic lamprophyres also have Th/U z3, they are both generally enriched in Th relative to U, more typical of crustal rocks, with Th/U up to 12–13 (Fig. 6b), values characteristic of U-depleted lower crustal granulites (e.g. Bea & Montero 1999). In addition, most shoshonites and lamprophyres have Nb/Ta clustering around 17·5 (Fig. 6c), whereas most vaugnerites have much lower ratios of around 12 or less, values that Green (1995) considered characteristic of primitive mantle and post-Archean continental crustal materials respectively. Even more important are the differences in Sr and Nd isotopes (Table 1, Fig. 7). Shoshonites have a depleted, relative to bulk Earth, isotope signature with 87Sr/86Sri 0·70400·0005 and
(Nd)i 4·10·9. When shoshonites erupt through continental crust, such as in the Aeolian Arc (Del Moro et al. 1998) or in the Absaroka Volcanics (Feeley 2003), 87Sr/86Sri may increase up to z0·7050 and
(Nd)i may decrease to
1, but rarely to more enriched, relative to bulk Earth, values. Vaugnerites, on the other hand, have more enriched Sr and Nd isotope compositions, with 87Sr/86Sri 0·70740·0004 and
(Nd)i z
4·00·7, and show little, if any, overlap with shoshonites. To the present authors’ knowledge, only the Variscan vaugnerites from the Meissen Massif, of the northern Bohemian Massif (Wenzel et al. 2000), have isotopic values that approach the shoshonites. The isotope compositions of the potassic lamprophyres are still more enriched, with averages of 87Sr/86Sri 0·70910·0035 and
(Nd)i
10·44·9 (Fig. 7), although the distribution is not unimodal: 87Sr/86Sri

shows two maxima at 0·707 and 0·711, and
(Nd)i shows another two maxima at
16 and
5. Up to this point it has been shown that the chemical compositions of shoshonites, vaugnerites and potassic lamprophyres, while overlapping somewhat, are different. To check whether these differences are statistically significant, a series of paired t-tests were performed, parting from the hypothesis that elements and isotope ratios have the same mean within the three rock associations. The results obtained are shown in Table 2. Comparing shoshonites and vaugnerites, only P2O5, Sc, Ga, Gd and Dy have a probability higher than 5% of having the same mean. Likewise, between shoshonites and potassic lamprophyres only Sc, Zn, Nb and Tb have a probability higher than 5% of having the same mean. Between vaugnerites and potassic lamprophyres there are more similarities; the equality of the means of TiO2, Rb, Sc, Zn, Y, Ta, Pb, U, Th, Pr, Gd, Tb, Dy, Ho and Lu cannot be rejected at the 0·05 significance level. If it is assumed that this dataset is representative of the three rock associations, these results confirm the existence of fundamental chemical differences between shoshonites, on the one hand, and vaugnerites and potassic lamprophyres, on the other hand. Consequently, if these premises are accepted, it must be concluded that there is no reason to consider these latter two rock associations as shoshonitic. It would be better, therefore, to remove the term shoshonitic from the K2O–SiO2 diagram and instead have a field marked very- or ultra-high-K (Fig. 1a).

3. Identification of shoshonites from their chemical composition Once it is recognised that neither vaugnerites nor potassic lamprophyres are compositionally comparable to shoshonites, i.e., they are not true shoshonites, the question is whether it is

166

J. H. SCARROW ET AL.

Figure 6 (A) Yb–La plot. Note how La/Yb increase from shoshonites to vaugnerites and potassic lamprophyres. (B) Th–U plot. All shoshonites plot along the line Th/U 3·2, close to the mantle ratio, but some potassic lamprophyres and most vaugnerites are significantly enriched in Th, with Th/U up to 12–13, values only found in metapelitic granulites (e.g. Bea & Montero 1999). (C) Ta–Nb plot. Most shoshonites and potassic lamprophyres have Nb/Ta clustering around 17·5 whereas most vaugnerites have Nb/Ta around 12 or less, values considered characteristic of mantle and crustal materials respectively (Green 1995).

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

167

Figure 7 (A) Histograms of the initial 87Sr/86Sr. Note that most shoshonites are below 0·705, whereas most vaugnerites and all lamprophyres are above this value. (B) Histograms of
(Nd)i. Most shoshonites have positive values, whereas most vaugnerites and the lamprophyres have negative
(Nd)i. Note the population of lamprophyres with
(Nd)i <
10. See text for details.

possible to identify a rock series as shoshonitic solely from the chemical composition of the specimens that form it. Since neither element (or element-ratio) scatter-plots nor multielement spider-plots were found to be selective enough, the present authors turned to multivariate statistical techniques, specifically to discriminant function analysis which seemed best suited for their purposes. Discriminant functions are calculated for a given combination of variables (in this case analysed elements or isotope ratios) and are the linear combination of these variables that best separate two or more previously defined groups, here being shoshonites, vaugnerites and potassic lamprophyres. Depending on the chosen combination of variables, it is possible to calculate different discriminant functions for the same data set. The quality of discrimination depends on the representativeness of the initial data set and how well it is partitioned into the groups to be discriminated; this depends on the availability of data and the criteria of the researcher and cannot be statistically quantified. Nevertheless, how well the previously defined groups are separated by the discriminant functions can be evaluated by means of the Mahalanobis distance, which compares the distances between group means, with a larger value indicating a better discrimination. In addition, the authors checked that the Mahalanobis distance was greater than zero, that is the groups did not coincide, using an F test which determines the equality of variances (average squared

deviation of all possible observations from the mean) of the groups based on the theoretical distribution of values to be expected by randomly sampling a normal population (Fdistribution). Table 3 presents the results of the multivariate analysis on the shoshonites, vaugnerites and potassic lamprophyres showing the discriminant functions found to be most useful. A first analysis used the major elements except MnO (because when expressed as an oxide percentage this element suffers from high relative errors and may therefore introduce noise into the discriminant function) (260 shoshonites, 180 vaugnerites, 196 potassic lamprophyres); and a second analysis used the major elements plus trace elements Rb, Cs and Th (47 shoshonites, 79 vaugnerites, 43 potassic lamprophyres). The significant improvement in goodness of fit upon inclusion of the trace elements is, in fact, somewhat deceptive, being, in part, a result of the reduction in the number of samples in the data set (only a selection of the rocks were analysed for those trace elements) and, it is suggested, a relative increase in the quality of the data. This was demonstrated by repeating the analysis using the major elements alone for those samples for which Rb, Cs and Th were analysed. The percentage of rocks classified by rock type was practically indistinguishable from the major elements plus selected trace elements discriminant functions for the same data set. The clear division of the rock types and the improvement in separation, as indicated by the larger

168

J. H. SCARROW ET AL.

Table 2 Paired t-test with unequal variance. Shoshonites/Vaugnerites t TiO2 Al2O3 FeO tot MgO MnO CaO Na2O K 2O P2 O 5 Li Rb Cs Sr Ba Sc V Cr Co Ni Cu Zn Ga Y Nb Ta Zr Hf Pb U Th La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu 86 Sr/86Sri
(Nd)i

Shoshonites/K-lamprophyres

Vaugnerites/K-lamprophyres

t

prob

d.f.

t

prob

d.f.

t

prob

d.f.


64·118 52·628 101·964
49·078 187·018 137·614 33·445
61·268
17·990
141·313
135·256
179·719 120·326
48·073 0·9312 125·400
67·235 25·200
39·281 59·341 27·781
0·1114 51·726 45·659
74·382
38·008
66·562
80·296
27·013
57·009
12·924
28·561
54·618
45·478
45·396
0·1354
18·533
27·660 29·164 59·795 53·207 27·052 41·074
269·811 133·760

0·0000 0·0000 0·0000 0·0004 0·0000 0·0000 0·0099 0·0000 0·0730 0·0000 0·0000 0·0000 0·0000 0·0000 0·3547 0·0000 0·0000 0·0106 0·0000 0·0000 0·0068 0·9116 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0694 0·0000 0·0092 0·0000 0·0000 0·0054 0·0000 0·0000 0·0000

347 383 402 382 424 468 442 360 312 67 221 118 377 259 75 122 241 251 302 66 78 81 178 284 87 237 79 161 165 179 260 227 51 170 181 170 53 160 24 55 22 176 161 77 53


54·732 291·750 51·346
148·124 84·623 33·919 136·985
129·381
94·783
82·833
81·379
42·992
43·082
71·946
15·278 80·956
149·135 34·492
144·733 33·965 0·4938 73·765 70·413
0·7091
37·150
79·736 76·789
43·220
13·297
28·694
34·893
53·384
30·892
70·200
71·181
40·390
29·248
13·866 35·383 62·194 71·198 86·637 45·060
69·410 121·476

0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0001 0·0000 0·0000 0·1280 0·0000 0·0000 0·0007 0·0000 0·0005 0·6222 0·0000 0·0000 0·4789 0·0004 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0033 0·0000 0·0000 0·0000 0·0047 0·1679 0·0012 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000

217 395 293 324 282 353 323 272 197 9 154 48 169 160 228 254 211 171 174 72 149 78 227 245 72 201 78 68 176 199 162 158 48 153 137 98 68 132 33 68 24 160 155 25 37


14·004 195·907
30·635
89·041
56·007
81·569 105·137
68·341
80·093 74·423
0·6601 40·260
90·624
49·451
20·739
39·392
70·223
53·380
116·817
48·827
15·757 37·894 11·437
40·543 14·721
46·166
22·241
13·580 15·094 0·9875
28·199
33·702 0·1118
42·338
45·575
40·058
18·469 0·9203 12·066 13·941 30·079 60·353 18·441
23·866 66·275

0·1625 0·0000 0·0058 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·0000 0·5099 0·0002 0·0000 0·0000 0·0415 0·0001 0·0000 0·0000 0·0000 0·0000 0·1178 0·0003 0·2540 0·0001 0·1441 0·0000 0·0280 0·1794 0·1328 0·3251 0·0054 0·0009 0·9114 0·0000 0·0000 0·0001 0·0689 0·3590 0·2300 0·1675 0·0033 0·0000 0·0674 0·0247 0·0000

276 368 338 367 311 329 332 346 248 55 206 62 152 185 75 142 267 178 190 184 118 100 225 188 103 239 122 62 193 142 156 177 53 161 130 98 71 139 116 74 101 140 133 26 28

Notes: Prob is the probability of the means being equal; d.f. represents Satterthwaite’s degrees of freedom. Silica is not included because samples were selected within the interval SiO2 43–64 wt. %. Italic values are those for which equality of means cannot be rejected at the 0·05 significance level.

Mahalanobis distance, resulting from analysis of the smaller dataset, in particular when the trace elements are included (Table 3), is clearly illustrated in Figure 8 which shows the distribution of the three groups in discriminant function space. Generally, in discriminant function analysis, the first function is the most important discrimination tool with later functions providing additional, most often significant, differentiation. This is reflected in function 1 having an elevated canonical correlation (a measure of the usefulness of a function

in determining the differences between the groups) and eigen value (a measure of the relative discriminating power of the function) compared with function 2 (Table 3). In the present case function 1 broadly discriminates between the shoshonites and the vaugnerites plus the potassic lamprophyres, with function 2 dividing the latter group (Fig. 8). The relative importance of each element in distinguishing between the different rock types is indicated by the standardised canonical discriminant function coefficients (Table 3). The shoshonite versus vaugnerite plus potassic lamprophyre distinction is

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

169

Table 3 Standardised coefficients for the canonical discriminant function, the canonical correlation, the eigenvalues and the percentage total variance accounted for. To apply a discriminant function to an unknown sample, sum the products of the concentration of the elements, in the appropriate units, by the corresponding coefficient. Number of samples in the major element only data set: 260 shoshonites; 180 vaugnerites; 43 potassic lamprophyres. Abbreviations: Sh=shoshonites; Vg=vaugnerites; K-La=K-rich lamprophyres. Major elements

SiO2 TiO2 Al2O3 FeO tot MgO MnO CaO Na2O K2O P2O 5 Rb Cs Th Canonical correlation Eigenvalue Variance accounted (%)

Major elements and trace elements Major elements smaller dataset

fn1

fn2

fn1

fn2

fn1

fn2


0·8858 0·2128
0·8065
0·9117
0·081
0·7689
0·2726
0·1989 0·2595


0·2036
0·5975
0·6399 0·5087
0·9325 0·7966 0·0989 0·4849 0·1464 0·2164 0·5865 0·3461


0·9529 0·4974
0·4985
1·1382 0·6856
1·1391
0·0046
0·6546 0·4259 0·0625 0·2439 0·4916

0·9141
0·1436 1·3235 0·2503 1·1838 0·4226 0·1863 0·0828 0·0577


1·3792 0·6146
1·0456
1·3587
0·0418
1·5614
0·3885
0·8495 0·1483

0·3579 0·3031 0·9466
0·6462 1·1286
0·6312
0·4183
0·4864
0·0850

0·8421 2·43762 78·5

0·6323 0·66608 21·5

0·9364 7·11650 68·0

0·8773 3·34160 32·0

0·9133 5·03050 68·4

0·8363 2·32730 31·6

Mahalanobis squared distance between groups Major elements

Vg K-La

Major elements and trace elements

Major elements smaller dataset

Sh

Vg

Sh

Vg

Sh

Vg

6·06 13·80

7·45

34·97 46·69

20·37

18·65 37·09

18·31

Percentage of samples classified as rock type True lithology

Sh Vg K-La

Major elements

Major elements and trace elements

Sh

Vg

K-La

Sh

91·9 2·8 7·1

7·3 87·8 8·2

0·8 9·4 84·7

100 0 0

principally controlled by differences in SiO2, FeO, CaO and K2O (function 1), with Al2O3, MgO and Th playing a more important role in separating the vaugnerites and potassic lamprophyres (function 2). The larger data set using major elements alone to discriminate shoshonites from the other two groups yielded w8% of false positives (i.e., specimens classified as shoshonites when they are not) and 8% of false negatives for shoshonites (i.e., shoshonites classified as non-shoshonites) (Table 3). Whereas in the smaller data set, discrimination improved the fit such that the percentage of false positives and negatives for shoshonites both decreased to 0%. Similar decreases in the number of false positives and negatives between the larger and smaller data sets were also observed for the vaugnerites and potassic lamprophyres (Table 3). As stated before, discriminant functions should be used cautiously, for their validity strongly depends on the representativeness of the data set used to calculate them, something

Percentage classified (%) Vg K-La 0 98·7 2·3

0 1·3 97·7

Major elements smaller dataset

Sh

Vg

K-La

100 0 2·3

0 98·7 2·3

0 1·3 95·4

very difficult to estimate in geological collections. In the present case, however, given the number of specimens and their varied age and provenance, it seems reasonable to assume that the data set is adequately representative. It is concluded, therefore, that for high quality data, discriminant function analysis using major elements alone (although if available we recommend including trace elements Th and Cs), will distinguish between shoshonites, vaugnerites and potassic lamprophyres. So, if application of the discriminant functions in Table 3 to a rock series that plots in the upper region of the K2O–SiO2 diagram results in the rejection of most samples as shoshonites, then the present authors believe there is no reason for qualifying the series as shoshonitic. If, on the other hand, most samples are not rejected and they have no other chemical and isotopic feature radically different from shoshonites, the ground seems firm enough to classify the series as truly shoshonitic and to claim for it a subduction-related source.

170

J. H. SCARROW ET AL.

Figure 8 Discriminant function 1–Discriminant function 2 plot. The shoshonite versus vaugnerite plus potassic lamprophyre distinction is principally controlled by differences in SiO2, Al2O3, FeO, CaO and K2O (function 1), with Al2O3, MgO and Th playing a more important role in separating the vaugnerites and potassic lamprophyres (function 2). See text for details.

4. Do vaugnerites and potassic lamprophyres represent: crust-contaminated shoshonite magmas or enriched lithospheric mantle melts? From the data presented in the previous sections, it seems clear that vaugnerites and potassic lamprophyres are chemically and isotopically different from subduction-related shoshonites, from which they can be efficiently separated on a compositional basis. This fact, however, still does not preclude a subduction source for the two rock types if it is accepted that they could have been derived from shoshonite magmas variously contaminated with crustal materials. The elevated Li, Cs, Th, U, high Th/U, and low Nb/Ta of most vaugnerites and some lamprophyres lends support to the idea of crustal contamination; to assess whether the initial magma was shoshonitic, we turn to Sr and Nd isotopes. The solid curve in Figure 9a represents the evolution of a magma with the average composition of shoshonites (1300 ppm Sr, 87Sr/86Sr 0·703882, 28 ppm Nd and 143Nd/144Nd 0·512884) contaminated with increasing percentages of a crustal material (CM) with 300 ppm Sr, 87Sr/86Sr 0·715, 26 ppm Nd and 143Nd/144Nd 0·5122. This crustal composition is especially favourable to causing perceptible contamination effects because it combines high Sr and Nd contents (higher than the average composition of greywackes, Wedepohl 1995), elevated 87Sr/86Sr (similar to the present-day average composition of suspended load of worldwide rivers, Goldstein & Jacobsen 1988) and low 143Nd/144Nd. Looking at Figure 9a, it is evident that production of a composition similar to more isotopically primitive vaugnerites requires the assimilation of 30% to 50% of crustal material, with the percentage rising up to w70% to produce the most common vaugnerites or potassic lamprophyres. These values are totally unrealistic, and incompatible with both the thermodynamics of assimilation and the relatively unfractionated major and trace element composition of vaugnerites and lamprophyres, so it must be concluded these two rock types do not represent contaminated shoshonitic magmas. Nevertheless, as indicated before, crustal contamination of some type seems likely in the generation of vaugnerites and probably some lamprophyres, but the magma that underwent contamination must have had an initial isotope

composition distinct from most shoshonite magmas (e.g. Prelevic et al. 2004). The key to understanding the isotope composition of vaugnerite and lamprophyre magmas is to consider that they derive from metasomatised mantle with anomalous Rb/Sr and Sm/Nd enriched long before the melting event. This idea was proposed for a group of potassic lamprophyres with extremely low
(Nd)t (<
10) and moderate 87Sr/86Sri (0·705 to 0·707) (Tingey et al. 1991; Buhlmann et al. 2000), but it seems applicable to many such rocks. Figure 9b shows the isotopic evolution with time of three different metasomes with identical initial 87Sr/86Sr 0·703882 and 143Nd/144Nd 0·512884, representing a wide spectrum of enriched mantle sources. The first has Rb/Sr 0·13 and Sm/Nd 0·19, similar to the Puy Beaunit mantle xenoliths, French Massif Central (Fe´me`nias et al. 2004); the second has Rb/Sr 0·05 and Sm/Nd 0·15 similar to the mantle xenoliths found in the giant agpaitic massifs of Kola, Lovozero and Khibina (Arzamastsev et al. 2001 and authors’ unpublished data); and the third has Rb/Sr 0.31 and Sm/Nd 0.215 similar to the phlogopite peridotites of Finero, Ivrea Zone (authors’ unpublished data). For simplicity, it is assumed that metasomatism changed just the element ratios but kept the isotope ratios constant. The calculations show that the isotope composition of vaugnerites may be easily attained by magmas derived from 200 to 800 M.y. old metasomes with Rb/Sr w0·1 to 0·13 and Sm/Nd w0·17 to 0·20, especially if they underwent some additional crustal contamination. The isotope composition of most potassic lamprophyres, on the other hand, requires older metasomes with a longer residence, between 1000 and 1500 M.y. To discuss the petrogenesis of vaugnerites and potassic lamprophyres in detail is beyond the scope of this paper. For the present purposes, it is sufficient to highlight that they need a long-term enriched metasomatic source.

5. Geodynamic considerations Many of the compositional features of the shoshonites point towards them having a subduction-related source, be it active arc or recycled arc products.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES

171

Figure 9
(Nd)i–87Sr/86Sri plot. (A) The solid grey curve represents the mixing line between a magma with the average composition of shoshonites (1300 ppm Sr, 87Sr/86Sr 0·73882, 28 ppm Nd and 143Nd/144Nd 0·512884) and increasing percentages of a crustal material (CM) with 300 ppm Sr, 87Sr/86Sr 0·715, 26 ppm Nd and 143Nd/144Nd 0·5122. Note how the isotopic composition of the less ‘crustal’ vaugnerites requires assimilation of about 30% of crustal material by a shoshonitic magma; this percentage reaches 65–70% for the most common vaugnerites. (B) Dashed grey lines represent the temporal evolution of three distinct metasomes with identical initial isotope composition and different Rb/Sr and Sm/Nd ratios. It follows that the isotope composition of vaugnerites may be easily attained by magmas derived from 200 to 800 M.y. old metasomes with Rb/Sr w0·1 to 0·13 and Sm/Nd w0·17 to 0·20, especially if they underwent some crustal contamination. The isotope composition of most potassic lamprophyres, on the other hand, requires metasomes with longer residence, between 1000 and 1500 M.y.

By contrast, the compositional characteristics of the vaugnerites and lamprophyres are interpreted to be the result of metasomatism of their mantle source. The cause of such metasomatism resulting in the isotopic characteristics seen in the vaugnerites and lamprophyres could be the dehydration of a subducting slab but, if so, it must have occurred long before

the mantle melting event that generated these rocks. It is not surprising, therefore, that the most favourable scenarios for these rocks are cratonic areas or Hercynotype (Pitcher 1979) orogens such as the Variscides of western Europe, in which old metasomes can later be involved in melting events as a result of mantle dynamics (e.g., Tingey et al. 1991; Bea et al. 1999). As

172

J. H. SCARROW ET AL.

mentioned before, most vaugnerites also require a significant percentage of crustal contamination as indicated by their elevated heat production, high Th/U, low Nb/Ta (Fig. 5), and elevated Cs contents (Table 1, Fig. 4e). Considering the repeated association between vaugnerites and crustal-melt granites (e.g. Sabatier 1991, and references therein), the likeliest scenario is of magmas derived from a mantle metasome intruding into a crustal anatectic zone, where extensive mixing with already molten materials is likely to occur. Conversely, the role of crustal contamination in potassic lamprophyres is not essential. Lamprophyres with 87Sr/86Sri <0·707 and
(Nd)i <
10, characteristic of non-orogenic areas, are likely to represent old-metasome melts with little crustal contamination, as reflected by their mantle-like Th/U ratio, low Cs (<5 ppm) and high Rb/Cs (>40). The lamprophyres found in orogenic areas, on the other hand, frequently show evidence of crustal contamination and their mineralogy and chemical and isotopic composition overlap with the vaugnerites with which they are often spatially and temporarily associated (e.g. Sabatier 1991; Wenzel et al. 2000; Buda & Dobosi 2004). These lamprophyres likely represent lesscontaminated portions of mantle-magma which gave rise to the vaugnerites.

6. Conclusions + The term shoshonitic should not be applied to a rock series, such as the appinite–vaugnerite series or the potassic lamprophyres, solely because it plots in the shoshonitic region of the K2O–SiO2 diagram. + The identification of shoshonites cannot be carried out solely by means of the K2O–SiO2 and the N-MORB normalised multi-element plots. These do not discriminate true subduction-source shoshonites from other K-rich rocks. The discrimination of shoshonites from other rocks that plot in the shoshonitic field of the K2O–SiO2 diagram can be carried out by means of dichotomous discriminant functions such as those shown in Table 3. + Vaugnerites generally show a considerable degree of crustal contamination. The contaminated magma, however, was not shoshonitic. Its isotope composition is consistent with derivation from a metasomatic layer of the lithospheric mantle (metasome) enriched at least 200 M.y. before melting. The source of metasomatism may be the dehydration of a subducting slab but, if so, subduction must have occurred in an earlier orogenic cycle. + Potassic lamprophyres, on the other hand, require older metasomes (1000 to 1500 M.y.) and may also have undergone variable degrees of crustal contamination, though rarely to the same extent as vaugnerites. In orogenic areas, the potassic lamprophyres that appear temporally and spatially closely connected to vaugnerites might represent the less-contaminated portions of the magma that, upon contamination, gave rise to vaugnerites.

7. Acknowledgements The authors are indebted to the creators and curators of the GEOROC database at Mainz, Germany, which greatly facilitated the finding of data. We are grateful to J. P. Lie´gois and J. R. Stern for detailed reviews which helped us to considerably improve the manuscript. C. H. Donaldson is thanked for his editorial handling. This work has been supported by the Spanish CICYT projects CGL2005–05863/BTE and CGL2008–02864, and the Andalusian grant RNM1595.

8. Supplementary Material Tables 1E, 2E and 3E are published as Supplementary Material with the on-line version of this paper. This is hosted by the Cambridge Journals Online service and can be viewed at http://journals.cambridge.org/tre

9. References Allen, J. E. & Balk, R. 1954. Mineral resources of Fort Defiance and Tohatchi quadrangles, Arizona and New Mexico. New Mexico Bureau of Mines and Mining Bulletin 36, 192. Arzamastsev, A. A., Bea, F., Glaznev, L. V., Arzamastseva, L. V. & Montero, P. 2001. Kola alkaline province in the Paleozoic: evaluation of primary mantle magma composition and magma generation conditions. Russian Journal of Earth Sciences 3, 151–77. Ashley, P. M., Cook, N. D. J., Hill, R. L. & Kent, A. J. R. 1994. Shoshonitic Lamprophyre Dykes and Their Relation to Mesothermal Au–Sb Veins at Hillgrove, New South Wales, Australia. Lithos 323, 249–72. Barberi, F., Innocent, F., Ferrara, G., Keller, J. & Villari, L. 1974. Evolution of Eolian Arc Volcanism (Southern Tyrrhenian Sea). Earth and Planetary Science Letters 21, 269–76. Bea, F. 2004. La naturaleza del magmatismo de la Zona Centro Ibe´rica: consideraciones generales y ensayo de correlacio´n. In Vera, J. A. (ed.) Geologı´a de Espan˜a, 128–33. Madrid: SGE-IGME. Bea, F., Montero, P. & Molina, J. F. 1999. Mafic precursors, peraluminous granitoids, and late lamprophyres in the Avila batholith: A model for the generation of Variscan batholiths in Iberia. Journal of Geology 107, 399–419. Bea, F., Montero, P. & Zinger, T. 2003. The Nature and Origin of the Granite Source Layer of Central Iberia: Evidence from Trace Element, Sr and Nd Isotopes, and Zircon Age Patterns. Journal of Geology 111, 579–95. Bea, F. & Montero, P. 1999. Behavior of accessory phases and redistribution of Zr, REE, Y, Th, and U during metamorphism and partial melting of metapelites in the lower crust: An example from the Kinzigite Formation of Ivrea-Verbano, NW Italy. Geochimica et Cosmochimica Acta 63, 1133–53. Beard, B. L. & Johnson, C. M. 1993. Hf Isotope Composition of Late Cenozoic Basaltic Rocks from Northwestern Colorado, USA – New Constraints on Mantle Enrichment Processes. Earth and Planetary Science Letters 119, 495–509. Beard, B. L. & Johnson, C. M. 1997. Hafnium isotope evidence for the origin of Cenozoic basaltic lavas from the southwestern United States. Journal of Geophysical Research-Solid Earth 102 (B9), 20149–78. Beccaluva, L., Gabbianelli, G., Lucchini, F., Rossi, P. L. & Savelli, C. 1985. Petrology and K/Ar Ages of Volcanics Dredged from the Eolian Seamounts – Implications for Geodynamic Evolution of the Southern Tyrrhenian Basin. Earth and Planetary Science Letters 74, 187–208. Bonaccorso, A., Cardaci, C., Coltelli, M., Del Carlo, P., Falsaperla, S., Panucci, S., Pompilio, M. & Villari, L. 1996. Annual Report of the World Volcanic Eruptions in 1993. Bulletin of Volcanic Eruptions 33, 7–13. Bor-ming, J., Sun, S. S. & Nesbitt R. W. 1979. REE distribution and petrogenesis of the Spanish Peaks Igneous Complex, Colorado. Contributions to Mineralogy and Petrology 70, 281–98. Bowes, D. R. & Kosler, J. 1993. Geochemical Comparison of the Subvolcanic Appinite Suite of the British Caledonides and the Durbachite Suite of the Central-European Hercynides – Evidence for Associated Shoshonitic and Granitic Magmatism. Mineralogy and Petrology 48, 47–63. Buda, G. & Dobosi, G. 2004. Lamprophyre-derived high-K mafic enclaves in Variscan Granitoids from the Mecsek Mts. (South Hungary). Neues Jahrbuch fu¨r Mineralogie-Abhandlungen 180, 115–47. Buhlmann, A. L., Cavell, P., Burwash, R. A., Creaser, R. A. & Luth, R. W. 2000. Minette bodies and cognate mica-clinopyroxenite xenoliths from the Milk River area, southern Alberta: records of a complex history of the northernmost part of the Archean Wyoming craton. Canadian Journal of Earth Sciences 37, 1629–50. Calanchi, N., Lucchini, P. & Rossi, P. L. 1983. Relationship between the K-rich series and the Indonesian volcanic arc in light of data on the Muria and Lasem volcanoes, Java. Mineralogica et Petrographica Acta 27, 15–34.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES Canning, J. C., Henney, P. J., Morrison, M. A. & Gaskarth, J. W. 1996. Geochemistry of late Caledonian minettes from Northern Britain: Implications for the Caledonian sub-continental lithospheric mantle. Mineralogical Magazine 398, 221–36. Carlier, G., Lorand, J. P., Audebaud, E. & Kienast, J. R. 1997. Petrology of an unusual orthopyroxene-bearing minette suite from southeastern Peru, Eastern Andean Cordillera: Al-rich lamproites contaminated by peraluminous granites. Journal of Volcanology and Geothermal Research 75, 59–87. Carmichael, I. S. E., Lange, R. A. & Luhr, J. F. 1996. Quaternary minettes and associated volcanic rocks of Mascota, western Mexico: a consequence of plate extension above a subduction modified mantle wedge. Contributions to Mineralogy and Petrology 124, 302–33. Carr, P. F. 1998. Subduction-related late Permian shoshonites of the Sydney basin, Australia. Mineralogy and Petrology 63, 49–71. Castro, A. Corretge´, G., de la Rosa, J. D. & Fernande´z, C. 2003. The appinite–migmatite complex of Sanabria, NW Iberian massif, Spain. Journal of Petrology 44, 1309–44. Conrey, R. M., Hooper, P. R., Larson, P. B., Chesley, J. & Ruiz, J. 2001. Trace element and isotopic evidence for two types of crustal melting beneath a High Cascade volcanic center, Mt. Jefferson, Oregon. Contributions to Mineralogy and Petrology 141, 710–32. Cox, K. G. & Bristow, J. W. 1984. The Sabine River basalt formation of the Lebombo monocline and south-east Zimbabwe. Special Publication of the Geological Society of South Africa 13, 125–47. De Astis, G., Peccerillo, A., Kempton, P. D., Lavolpe, L. & Wu, T. W. 2000. Transition from Calc-Alkaline to Potassium-Rich Magmatism in Subduction Environments – Geochemical and Sr, Nd, Pb Isotopic Constraints from the Island of Vulcano (Aeolian Arc). Contributions to Mineralogy and Petrology 139, 684–703. De Lima, E. F. & Nardi, L. V. S. 1998. The Lavras do Sul Shoshonitic Association: Implications for the origin and evolution of Neoproterozoic shoshonitic magmatism in southernmost Brazil. Journal of South American Earth Sciences 11, 67–77. Del Moro, A., Gioncada, A., Pinarelli, L., Sbrana, A. & Joron, J. L. 1998. Sr, Nd, and Pb isotope evidence for open system evolution at Vulcano, Aeolian Arc, Italy. Lithos 43, 81–106. Debon, F., Guerrot, C., Menot, R. P., Vivier, G. & Cocherie, A. 1998. Late Variscan granites of the Belledonne massif (French western Alps): an early Vise´an magnesian plutonism. Schweizerische Mineralogische und Petrographische Mitteilungen 78, 67–85. Dias, G., Simoes, P. P., Ferreira, N. & Leterrier, J. 2002. Mantle and crustal sources in the genesis of late-Hercynian granitoids (NW Portugal): Geochemical and Sr-Nd isotopic constraints. Gondwana Research 5, 287–305. Dickinson, W. R., Rickard, M. J., Coulson, F. I. E., Smith, J. G. & Lawrence, R. L. 1968. Late Cenozoic shoshonitic lavas in north-western Viti Levu, Fiji. Nature 219, 148. Ellam, R. M., Menzies, M. A., Hawkesworth C. J., Leeman W. P., Rosi, M. & Serri, G., 1988. The Transition from Calc-Alkaline to Potassic Orogenic Magmatism in the Aeolian-Islands, Southern Italy. Bulletin of Volcanology 50, 386–98. Ellam, R. M., Hawkesworth, C. J., Menzies, M. A. & Rogers, N. W. 1989. The Volcanism of Southern Italy – Role of Subduction and the Relationship between Potassic and Sodic Alkaline Magmatism. Journal of Geophysical Research-Solid Earth and Planets 94 (B4), 4589–601. Erlikh, E. N. 1966. Petrochemistry of Cainozoic volcanic province of Kurile and Kamchatka. U.S.S.R. Academy of Sciences Siberian Branch, Institute of Volcanology, Moscow, 289. Esperanca, S., Crisci, G. M., Derosa, R. & Mazzuoli, R. 1992. The Role of the Crust in the Magmatic Evolution of the Island of Lipari (Aeolian Islands, Italy). Contributions to Mineralogy and Petrology 112, 450–62. Feeley, T. C. 2003. Origin and Tectonic Implications of Across-Strike Geochemical Variations in the Eocene Absaroka Volcanic Province, United States. Journal of Geology 111, 329–46. Feeley, T. C., Cosca, M. A. & Lindsay, C. R. 2002. Petrogenesis and Implications of Calc-Alkaline Cryptic Hybrid Magmas from Washburn Volcano, Absaroka Volcanic Province, USA. Journal of Petrology 45, 663–705. Fe´me`nias, O., Coussaert, N., Berger, J., Mercier, J. C. & Demaiffe, D. 2004. Metasomatism and melting history of a Variscan lithospheric mantle domain: evidence from the Puy Beaunit xenoliths (French Massif Central). Contributions to Mineralogy and Petrology 148, 13–28. Ferre, E. & Leake, B. E. 2001. Geodynamic significance of early orogenic high-K crustal and mantle melts: example of the Corsica Batholith. Lithos 59, 47–67.

173

Francalanci, L. 1989. Trace-Element Partition-Coefficients for Minerals in Shoshonitic and Calc-Alkaline Rocks from Stromboli-Island (Aeolian Arc). Neues Jahrbuch Fu¨r MineralogieAbhandlungen 160, 229–47. Francalanci, L., Manetti, P. & Peccerillo, A. 1989. Volcanological and Magmatological Evolution of Stromboli Volcano (Aeolian Islands) – the Roles of Fractional Crystallization, Magma Mixing, Crustal Contamination and Source Heterogeneity. Bulletin of Volcanology 51, 355–78. Galan, G., Corretge, L. G. & Laurent, O. 1997. Low-potassium vaugnerites from Gueret (Massif Central, France). Mafic magma evolution influenced by contemporaneous granitoids. Mineralogy and Petrology 59, 165–87. Gallastegui, G. 1993. Petrologı´a del macizo granodiorı´tico de BayoVigo (provincia de Pontevedra, Espan˜a). PhD Thesis, University of Oviedo, Spain. Gauthier, P. J. & Condomines, M. 1999. Pb-210–Ra-226 radioactive disequilibria in recent lavas and radon degassing: inferences on the magma chamber dynamics at Stromboli and Merapi volcanoes. Earth and Planetary Science Letters 172, 111–26. Gibson, S. A., Thompson, R. N., Leat, P. T., Morrison, M. A., Hendry, G. L. & Dickin, A. P. 1991. The Flat Tops Volcanic Field 1: Lower Miocene Open-System, Multisource Magmatism at Flander, Trappers Lake. Journal of Geophysical Research 96 (B8), 13609–627. Gibson, S. A., Thompson, R. N., Leat, P. T., Dickin, A. P., Morrison, M. A., Hendry, G. L. & Mitchell, J. G. 1992. Asthenospherederived magmatism in the Rio Grande Rift, Western USA; implications for continental break-up. Geological Society, Special Publication 68, 61–89. London: The Geological Society. Goldstein, S. J. & Jacobsen, S. B. 1988. Nd and Sr isotopic systematics of river water suspended material: Implications for crustal evolution. Earth and Planetary Science Letters 87, 249–65. Green, T. H. 1995. Significance of Nb/Ta as an indicator of geochemical processes in the crust-mantle system. Chemical Geology 120, 347–59. Greenough, J. D., Owen, J. V. & Ruffman, A. 1993. Noble-Metal Concentrations in Shoshonitic Lamprophyres – Analysis of the Weekend Dykes, Eastern Shore, Nova-Scotia, Canada. Journal of Petrology 34, 1247–69. Hoch, M., Rehkamper, M. & Tobschall, H. J. 2001. Sr, Nd, Pb and O isotopes of minettes from Schirmacher Oasis, East Antarctica: A case of mantle metasomatism involving subducted continental material. Journal of Petrology 42, 1387–1400. Hogg, N. C. 1972. Shoshonitic lavas in west-central Utah. Brigham Young University Geological Studies 19, 133–84. Hornig-Kjarsgaard, I., Keller, J., Koberski, U., Stadlbauer, E., Francalanci. L. & Lenhart, R. 1993. Geology, stratigraphy and volcanological evolution of the Island of Stromboli, Aeolian Arc, Italy. Acta Vulcanologica 3, 21–68. Huang, Z., Chongqiang, L., Yang, H., Xu, C., Han, R., Xiao, Y., Zhang, B. & Li, W. 2002. The geochemistry of lamprophyres in the Laowangzhai gold deposits, Yunnan Province, China: Implications for its characteristics of source region. Geochemical Journal 96, 91–112. Iddings, J. P. 1895. Absarokite-shoshonite-banakite series. Journal of Geology 3, 935–59. Jacques, A. L. 1976. High-K2O island-arc volcanic rocks from the Finisterre and Adelbert Ranges, northern Papua New Guinea. Geological Society of America Bulletin 87, 861–7. Jahn, B., Sun, S. S. & Nesbitt, R. W. 1979. REE Distribution and Petrogenesis of the Spanish Peaks Igneous Complex, Colorado. Contributions to Mineralogy and Petrology 70, 281–98. Jakes, P. & White, A. J. R. 1969. Structure of the Melanesian Arcs and correlation with distribution of magma types. Tectonophysics 8, 223–36. Janousek, V., Rogers, G. & Bowes, D. R. 1995. Sr-Nd isotopic constraints on the petrogenesis of the Central Bohemian Pluton, Czech Republic. Geologische Rundschau 84, 520–34. Janousek, V., Bowes, D. R., Rogers, G., Farrow, C. M. & Jelinek, E. 2000. Modelling diverse processes in the petrogenesis of a composite batholith: the Central Bohemian Pluton, Central European Hercynides. Journal of Petrology 41, 511–43. Joplin, G. A. 1968. The shoshonite association – a review. Journal of the Geological Society of Australia 15, 275–94. Kapp, H. 1960. Zur petrologie der subvulkane Szwischen Mesters Vig und Antarctic Havn (Ost-Gronland). Medd Gronland 153, 203. Keller, J. 1974. Petrology of Some Volcanic Rock Series of Aeolian Arc, Southern Tyrrhenian Sea – Calc-Alkaline and Shoshonitic Associations. Contributions to Mineralogy and Petrology 46, 29–47.

174

J. H. SCARROW ET AL.

Kemp, J. F. & Billingsley, P. 1921. Sweet Grass Hills. Geological Society of America Bulletin 32, 437–78. Kepezhinskas, P., McDermott, F., Defant, M. J., Hochstaedter, A., Drummond, M. S., Hawkesworth, C. J., Koloskov, A., Maury, R. C. & Bellon, H. 1997. Trace element and Sr-Nd-Pb isotopic constraints on a three-component model of Kamchatka arc petrogenesis. Geochimica et Cosmochimica Acta 61, 577–600. Kesson, S. E. & Smith, I. E. 1972. TiO2 content and the shoshonite and alkaline associations. Nature Physical Science 236, 110–11. Klotzli, U. S., Buda, G. & Skiold, T. 2004. Zircon typology, geochronology and whole rock Sr-Nd isotope systematics of the Mecsek Mountain granitoids in the Tisia Terrane (Hungary). Mineralogy and Petrology 81, 113–34. Knopf, A. 1936 Igneous geology of the Spanish Peaks Region, Colorado. Geological Society of America Bulletin 47, 1729–84. Lacroix, A. 1928. La constitution lithologique des iles volcaniques de la Polynesie Australe. Academy of Sciences, Paris, Memoir 59, 1–80. Le Bas, M. J., Le Maitre, R. W., Streckeisen, A & Zanettin, B. 1986. A chemical classification of volcanic rocks based on the total alkali–silica diagram. Journal of Petrology 27, 745–50. Le Maitre, R. W., Streckeisen, A., Zanettin, B., Le Bas, M. J., Bonin B. & Bateman P. (eds) 2002. Igneous Rocks: A Classification and Glossary of Terms: Recommendations of the International Union of Geological Sciences Subcommission on the Systematics of Igneous Rocks. Cambridge: Cambridge University Press. Liegeois, J. P., Navez, J., Hertogen, J. & Black, R. 1998. Contrasting Origin of Postcollisional High-K Calc-Alkaline and Shoshonitic Versus Alkaline and Peralkaline Granitoids – The Use of Sliding Normalization. Lithos 45, 1–28. Lo´pez-Moro, F. J. & Lo´pez-Plaza, M. 2004. Monzonitic series from the Variscan Tormes Dome (Central Iberian Zone): petrogenetic evolution from monzogabbro to granite magmas. Lithos 72, 19–44. Loughlin, G. F. 1919. Two lamprophyre dikes near Santaquin and Mount Nebo, Utah. United States Geological Survey Professional Paper 120-E, 101–9. Mackenzie, D. E. 1976. Nature and origin of late Cainozoic volcanoes in western Papua New Guinea. In Johnson, R. W. (ed.) Volcanism in Australasia, 221–38. Amsterdam: Elsevier. Mackenzie, D. E. & Chappell, B. W. 1972. Shoshonitic and Calc-Alkaline Lavas from Highlands of Papua New Guinea. Contributions to Mineralogy and Petrology 35, 50–62. Michaels, J. 1969. Studies of the volcanic petrology of the Navejo-Hopi area, Arizona. PhD thesis, University of California, Berkeley, 114 pp. Michon, G. 1987. Vaugnerites of Eastern French Massif-Central – a Multivariate Statistical Approach on Major Elements. Bulletin De La Societe Geologique de France 3, 591–600. Molina, J. F., Scarrow, J. H., Montero, P. & Bea, F. 2009. High-Ti amphibole as a petrogenetic indicator of magma chemistry: evidence for mildly alkalic-hybrid melts during evolution of Variscan basic-ultrabasic magmatism of Central Iberia. Contributions to Mineralogy and Petrology 158, 69–98. Montero, P., Bea, F. & Zinger, T. F. 2004. Edad 207Pb/206Pb en cristal u´nico de circo´n de las rocas ma´ficas y ultrama´ficas del sector de Gredos, Batolito de Avila (Iberia Central). Revista de la Sociedad Geologica Espan˜ola 17, 157–65. Morrison, G. W. 1980. Characteristics and tectonic setting of the shoshonite rock association. Lithos 13, 97–108. Nemec, D. 1973. Differentiation series of minettes in the Central Bohemian Pluton. Journal of Geology 81, 632–64. Peccerillo, A. & Taylor, S. R. 1976. Geochemistry of the Eocene calc-alkaline volcanic rocks from the Kastamonu area, northern Turkey. Contributions to Mineralogy and Petrology 58, 63–81. Peterson, T. D., Van Breemen, O., Sandeman, H. & Cousens, B. 2002. Proterozoic (1.85–1.75 Ga) igneous suites of the Western Churchill Province: granitoid and ultrapotassic magmatism in a reworked Archean hinterland. Precambrian Research 119, 73–100. Pirsson, L. V. 1905. Petrography and geology of the igneous rocks of the Highwood Mountains. United States Geological Survey Bulletin 237, 208. Pitcher, W. S. 1979. The nature, ascent and emplacement of granitic magmas. Journal of the Geological Society, London 136, 627–62. Prelevic, D., Foley, S. F., Cvetkovic, V. & Romer, R. L. 2004. Origin of minette by mixing of lamproite and dacite magmas in Veliki Majdan, Serbia. Journal of Petrology 45, 759–92. Righter, K. & Rosas-Elguera, J. 2001. Alkaline Lavas in the Volcanic front of the Western Mexican Volcanic Belt: Geology and Petrology of the Ayutla and Tapalpa Volcanic Fields. Journal of Petrology 42, 2333–61.

Rittman, A. 1940. Studien an eruptivegesteinen aus Ost-Gronland. Medd Gronland 115, 156. Roberts, M. P. & Clemens, J. D. 1993. Origin of high-potassium, calc-alkaline I-type granitoids. Geology 21, 825–8. Rock, N. M. S. 1987. The nature and origin of lamprophyres: an overview. Alkaline Igneous Rocks. In Fitton, J. G. & Upton, B. G. J. (eds) Alkaline Igneous Rocks. Geological Society Special Publication 30, 191–26. London: The Geological Society. Rogers, N. W. & Setterfield, T. N., 1994. Potassium and incompatibleelement enrichment in shoshonitic lavas from the Tavua volcano, Fiji. Chemical Geology 118, 43–62. Rossi, P. & Cocherie, A. 1991. Genesis of a Variscan Batholith – Field, Petrological and Mineralogical Evidence from the Corsica Sardinia Batholith. Tectonophysics 195, 319–46. Sabatier, H. 1978. New Data on Vaugnerites of French Massif Central. Comptes Rendus Hebdomadaires Des Seances De L Acade´mie Des Sciences Serie D 286, 9–12. Sabatier, H. 1980. Vaugnerites and Granites – a Particular Association of Acid and Basic Plutonic Rocks. Bulletin de Mineralogie 103, 507–22. Sabatier, H. 1991. Vaugnerites: Special lamprophyre-derived mafic enclaves in some Hercynian granites from Western and Central Europe. In Didier, J. & Barbarin, B. (eds) Enclaves and Granite Petrology, 63–81. Amsterdam: Elsevier. Scarrow, J. H., Molina, J. F., Bea, F. & Montero, P. 2009. Withinplate calc-alkaline rocks: Insights from alkaline mafic magmaperaluminous crustal melt hybrid appinites of the Central Iberian Variscan continental collision. Lithos 110, 50–64. Schellekens J. H. 1988. Geochemical evolution and tectonic history of Puerto Rico. Geological Society of America, Special Paper 322, 35–66. Schmidtt, H. H., Swann, G. A. & Smith, D. 1974. The Buell Park kimberlite pipe, Northeastern Arizona. Geological Society of America, Rocky Mountain Section Guidebook 27, 672–98. Schmucker, U. 1969. Geophysical Aspects of Structure and Composition of the Earth. In Wedepohl, K. H. (ed.) Handbook of Geochemistry, 134–226. Berlin: Springer-Verlag. Searle, M. P., Crawford, M. B. & Rex, A. J. 1992. Field Relations, Geochemistry, Origin and Emplacement of the Baltoro Granite, Central Karakoram. Transactions of the Royal Society of Edinburgh: Earth Sciences 83, 519–38. Sheppard, S. & Taylor, W. R. 1992. Barium-Rich and LREE-Rich, Olivine-Mica-Lamprophyres with Affinities to Lamproites, Mt Bundey, Northern-Territory, Australia. Lithos 28, 303–25. Skipp, B. & McGrew, L. W. 1977. The Maudlow and Sedan Formations of the Upper Cretaceous Livingston Group on the west edge of the Crazy Mountains Basin, Montana. United States Geological Survey Bulletin 1422-B, 1–68. Sun, C. H. & Stern, R. J. 2001. Genesis of Mariana shoshonites: Contribution of the subduction component. Journal of Geophysical Research 106 (B1), 589–608. Tatsumi, Y. & Eggins, S. M. 1995. Subduction Zone Magmatism. Frontiers in Earth Sciences. Cambridge, Massachusetts: Blackwell Science. Tingey, D. G., Christiansen, E. H., Best, M. G., Ruiz, J. & Lux, D. R. 1991. Tertiary Minette and Melanephelinite Dikes, Wasatch Plateau, Utah: Records of Mantle Heterogeneities and Changing Tectonics. Journal of Geophysical Research 96 (B8), 13529–44. Tsvetkov, A. A., Sukhanov, M. K. & Govorov, G. I. 1985. Regular evolution of magmatism in Recent and Paleogene arcs (for instance Kurile islands and northern Caucasus). In Bogatikov, O. A. (ed.) Oceanic magmatism – evolution, geological correlation, 185–218. Moscow: Nauka. Wagner, C. & Velde, D. 1993. Paleozoic Olivine-Bearing Lamprophyre from the Couy (Cher, France) Borehole – Mineral-Composition and Alteration Phenomena. European Journal of Mineralogy 5, 85–96. Washington H. S. 1917. Chemical analysis of igneous rocks. United States Geological Survey. United States Geological Survey Professional Paper 99, 473. Wedepohl, K. H. 1995. The composition of the continental crust. Geochimica et Cosmochimica Acta 59, 1217–32. Wenzel, T., Oberhansli, R. & Mezger, K. 2000. K-rich plutonic rocks and lamprophyres from the Meissen Massif (northern Bohemian Massif): Geochemical evidence for variably enriched lithospheric mantle sources. Neues Jahrbuch fu¨r Mineralogie-Abhandlungen 175, 249–93. Wierzcholowski, B. 2003. Potassium-rich dyke rocks of Rogo´wek (Sudetes, Poland). Archiwum Mineralogiczne 54, 77–97. Williams, H. 1936. Pliocene volcanoes of the Navajo-Hopi country. Geological Society of America Bulletin 47, 111–72.

‘ULTRA’-HIGH-K ROCKS: SIMILARITIES AND DIFFERENCES Witkind, I. J. 1973. Igneous rocks and related mineral deposits of the Barker Quadrangle, Little Belt Mountains, Montana. United States Geological Survey Professional Paper 752, 58. Wyman, D. A. & Kerrich, R. 1993. Archean Shoshonitic Lamprophyres of the Abitibi Subprovince, Canada – Petrogenesis, Age, and Tectonic Setting. Journal of Petrology 34, 1067–109. Xu, H. 1988. Petrology and Geochemistry of the Alkali Rocks from Dogo, Oki Islands, Shimane Prefecture, Southwestern Japan. Scientific Report 3–17, 1–106. Japan: Tohoku University.

175

Yao-Hui, J., Shao-Yong, J., Hong-Fei, L., Xun-Rou, Z., Xing-Jian, R. & Wan-Zhi, Y. 2002. Petrology and geochemistry of shoshonitic plutons from the western Kunlun orogenic belt, Xinjiang, northwestern China: implications for granitoid geneses. Lithos 63, 165–87.

MS received 25 March 2008. Accepted for publication 26 February 2009.