J. G. Holland

Geochemistry of the Lesser Antilles volcanic island arc

New analyses of 1518 rocks for major and certain trace elements are used to examine chemical variations between the 15 larger volcanic islands of the Lesser Antilles island arc. The depth to the top of the subduction zone dipping westward at about 40° lies about 100km below all the volcanoes of the arc. Most of the sampled eruptions are post-Miocene (5-1 m.y.) although south of Martinique, the Oligocene-Miocene and the younger arc are superimposed. There is a chemical variation along the arc axis, from alkalic (southern) through calc-alkalic (central) to tholeiitic (northern) volcanic suites. Three islands are examined in detail as type examples of this variation, i.e. Grenada (south), Dominica (centre), and St. Kitts (north). The Grenada suite includes basanites, alkalic basalts, and subalkalic basalts, andesites and dacites. The subalkalic basalts, andesites and dacites each fall into three chemical groupings along the axis of the arc, distinguished especially by K, Zr, Ni and Cr abundances. The whole Lesser Antilles assemblage is characterised by low K abundances and low K/Rb ratios, compared with other island arcs. The magmas are believed to have evolved through processes of partial melting and crystal fractionation. Partial melting of garnet Iherzolite at about 100km depth in a relatively ‘fertile’ zone of upper mantle in the southern sector, above the subducted slab of basaltic ocean crust, could have produced the undersaturated alkalic magmas. In the central and northern sectors, where the crustal structures are more complex, partial melting may have occurred within more ‘barren’ upper mantle, to produce tholeiitic and calc-alkalic magmas depleted in certain trace elements. In either case, water was probably added to the melted zone from the subducted and hydrated oceanic crust, since the whole arc assemblage was erupted explosively and the rocks are rich in A12O3, plagioclase is very calcic, and amphibole is an important phase. The second process was crystal fractionation at low pressure, as evidenced by the abundance of cumulate xenoliths. Separating phases for the southern volcanoes were olivine, calcic augite and Cr-spinel, followed by hornblende, anorthite and Ti-magnetite at lower temperatures. There is little evidence of the higher-temperature fractionation controls for the central and northern volcanoes.

Yttrium geochemistry applied to petrogenesis utilizing calcium-yttrium relationships in minerals and rocks

Abstract A survey of Y data from all sources shows that Y behaves systematically in igneous, metamorphic and sedimentary rock series, due to its incorporation in a predictable and uniform manner in Ca minerals. Compared with average calc-alkali basalt melts, plagioclase, kaersutite, augitic clinopyroxene and calcite have low Y for their CaO contents; whereas hornblende, garnet, orthopyroxene, apatite, sphene, zircon and most K, Na rich minerals other than plagioclase have high Y for their Ca contents. In sedimentary processes, Ca/Y becomes lower in shales and sandstones, but higher in limestones than their source. In metamorphic processes Y appears to be inert. In all igneous series for which Y data has been assembled Ca/Y falls as Ca falls. These series can be classified into three categories: (i) a standard calc-alkali trend, used as a reference (ii) J-type trends which become progressively impoverished in Y as Ca falls when compared with the standard trend, and (iii) L-trends which become progressively enriched in Y as Ca falls when compared with the standard trend. Despite little knowledge of partition coefficients, the J-type trends appear to have a significant component of hornblende control in their fractionation processes, whereas plagioclase and augitic clinopyroxene dominate the L-type trend. Alternative names for these series might therefore be the standard series, the hornblendic and the pyroxenic series respectively. Modern ocean floor basalts appear to be significantly richer in Y than modern calc-alkali basalts, and these in turn may be richer in Y than their Archaean counterparts.

Major element chemical composition of shields and the continental crust

Abstract Chemical data from the Canadian Shield and the Precambrian of Scotland indicates that there is a progressive enrichment towards the surface in potassium and possibly in titanium, but observed variations in titanium do not form a simple pattern. The averages for the surface of the shield areas are more siliceous and sodic, and poorer in potassium, titanium and the ferro-magnesian elements than are most published continental crustal averages. Assuming that granulite facies rocks form a substantial part of the shields, the total shield analysis becomes richer in iron, magnesium and calcium, and poorer in silicon and potassium than published estimates of the composition of the shields, and because the shields form the dominant fraction of the continental crust, substantial differences in the average for the continental crust emerge compared with estimates based on igneous rocks. In particular, a total continental crust rich in granulite facies rocks is likely to be richer in silicon and sodium, poorer in iron, magnesium and potassium and very much poorer in titanium than one based on igneous compositions.

A geochronological study of the Lewisian from Loch Laxford to Durness, Sutherland, N. W. Scotland

Potassium-argon biotite and hornblende ages from near Laxford Bridge, Sutherland, range up to 1770 m.y.: higher apparent ages of 2060 and 1840 m.y. occur in the biotite pods along the southern margin of the Laxford assemblage. A rubidium-strontium isochron from whole-rock specimens from the gneisses of the Laxford assemblage gives 1850 ± 50 m.y., initial 87Sr/86Sr = 0.7061 ± 0.0005 (T 1 2 = 4.85. 1010y). It is suggested that 1850 ± 50 m.y. is a minimum age for the climax of the Laxfordian metamorphic episode, and that this metamorphism had effectively terminated by 1750 m.y. the highest K-Ar age of a mineral from the type area. Younger apparent ages of minerals from and close to the type area spread down to 1575 ± 50 m.y., associated with late effects of the metamorphism, continued pegmatite-formation, and a general slow-cooling. Closure of radiogenic systems occurred later at Durness than at Laxford Bridge. Comparison with the sparse data from nearby shield areas suggests a closer correlation of the Lewisian with comparable rocks in Greenland than with Fennoscandia.

Structural regimes and metamorphic facies

Abstract This paper is an attempt to bring together the physical and chemical aspects of metamorphism by relating rheology to metamorphic facies in the orogenic environment. Five regimes are considered and defined within an idealised orogenic crust. The first regime extends from below the level of diagenesis to the point at which metamorphic reactions commence. Experimental studies suggest that at this level deformation by creep has a logarithmic relationship with time, but it is known that other factors such as fluid-pressure will profoundly influence actual strains. The onset of metamorphic reactions defines the upper level of the second regime, which is characterised by the formation of micas and amphiboles at the expense of pre-existing sedimentary minerals. These reactions are all associated with the release of volatiles, which will bring about a marked increase in the creep-rate at this level. These reactions are accompanied by new structural styles, similar folds, widespread schistosities and recumbent isoclinal folds on all scales. The third regime occupies the levels approximately represented by the epidote-amphibolite and amphibolite facies, characterised by a restricted mineralogy and similar Theologies for all the common rock types. Deformation will be primarily by grain boundary or intra-lattice diffusion following the β-creep law (ϵ = βtn; n Regime four is equated with the higher pressure-temperature amphibolite facies and is distinguished from regime three by two groups of reactions, the first, the dehydration of the remaining common hydrous minerals, and the second, partial melting with the release of a granitic fluid. Both reactions effect an increase in the creep-rate, accompanied by the formation of gneissose structures, extensive flow folding and possible diapiric movements: rates of strain will vary widely with bulk chemistry of the system. Regime five corresponds to the granulite facies, extending into the upper mantle. Deformation is by prolonged laminar flow, causing simple structural styles, with all rock types possessing similar rheology.

The petrography and chemistry of the igneous complex of the Malvern Hills, England

The nineteenth-century researches into the Malvernian led to the view that the complex consisted of metamorphic rocks, but the basis for this conclusion was not founded on a broad knowledge of the characters of igneous and metamorphic rocks. Field evidence relating to their origin leads to few definite conclusions. The 9 km2 complex is well exposed in numerous large quarries, but each quarry contains a wide variety of types brought together by shearing under moderate temperature conditions and later brittle faulting. The general shearing increases in intensity from north to south, the rocks in the two southernmost blocks being mainly chlorite-schist. In the north, however, some magnesium-rich ultramafic homblendites occur, occupying small areas among diorites, tonalites and granite. The diorites and tonalites, characterised by hydrolysed andesine or oligoclase, may be classed into olivine-normative mafic types and quartz-normative leucocratic types, together forming about 70 per cent of the complex. Shearing under moderate temperatures and high water vapour pressures has produced a wide variety of schistose rocks, some showing a weak lamination, erroneously interpreted earlier as a gneissose banding. Elsewhere in the Malverns, sheared pegmatites and granite veins in diorite produce similar banded rocks. The diorites and tonalites are also magnesium-rich, being generally high in potash as well. Small areas of a lineated and weakly foliated two-mica microcline-granite of near ‘minimum-melting’ type are present, as well as granite-pegmatites and potash and soda pegmatites which cut the ultramafics, diorites and tonalites, but predate the general hydrolysis and shearing. Microdiorite dykes cut the complex after some of the shearing and hydrolysis. Faulted against the plutonic rocks is a small area of pillow lavas and felsic tuffs, the Warren House volcanics, the felsic rocks being sodic dacite and quartz-keratophyre. The volcanics are cut by dykes which are mildly alkaline in composition; these dykes are not affected by the strong shearing of the plutonic rocks. During the general shearing and hydrolysis hornblende reacted to give chlorite and epidote, biotite was chloritised, and plagioclase became altered to muscovite. An early record of sillimanite is regarded as due to misidentification of zoisite; garnetiferous granites are present as well as one example of what may be a garnet-mica-schist of sedimentary origin.

Geochemistry of the western part of the Moinian assemblage

Synopsis Comparison of chemical analyses of the principal units of the Morar Division suggest that the inter-relationship between psammites and pelitic rocks is dominated by mechanical rather than chemical processes. This conclusion is drawn from similar ratios of Fe2O3/MgO, K/Rb, K/Ba and Ca/Y, and dissimilar ratios of Ti, Nb and Zr to Fe. Rocks of the Glenfinnan Division differ from comparable members of the Morar Division in having higher Nb/P2O2 and Nb/Y for any given Sr or Fe2O3/Ni ratio, and higher Sr/Y for any given SiO2 level. These distinctions may reflect consistent differences in the ratios of quartz and feldspar, ferromagnesians, and accessory minerals. Samples from the Loch Eil Division tend to have higher Zr/P2O5, Y/P2O5, CaO/P2O5, Rb/Sr and lower Ca/Y and K/Rb than their counterparts in the Morar and Glenfinnan Divisions. The Morar Basal Pelite has an immature chemistry showing affinity with the Lewisian basement upon which it rests. It can be distinguished from other Moinian pelitic units by its high Sr/Y. Certain individual psammitic horizons can be distinguished by their contents of P, Ti, Nb and Zr, and sometimes by Cr/Fe. The Morar Division has a chemistry consistent with deposition in a basin close to its supposed Lewisian source. In contrast, the Loch Eil Division may have been partly derived from a (post-Lewisian) sedimentary source, and may rest unconformably upon the underlying Moinian rocks.

Time, space and intensity relationships of the Precambrian and lower Palaeozoic metamorphisms of the Scottish Highlands

Summary Isotopic evidence suggests that a complex sequence of events, collectively known to us as the Morarian event, affected the Morar and Glenfinnan divisions of the Moine from 1000 to 750 Ma. A regional thermal anticline, with a minimum age of 750 Ma, extends north from Ardgour to northern Ross-shire: it probably extends southwards under the eastern Dalradian. This thermal anticline is cut across and shifted westwards south of Monar by the Sgurr Beag slide, on which up to 25 km thrust movement is estimated. The Loch Eil Division lacks old isotopic ages; it appears to be structurally and metamorphically simpler than the Morar and Glenfinnan divisions and it also differs from them geochemically. The simplest explanation of these facts is that Loch Eil Division rests unconformably on the metamorphosed Glenfinnan Division. The Central Highland Group (including the Monadhliath Schists) is presumed to occupy the same tectono-stratigraphical niche as the Loch Eil Division. Grampian events (510-480 Ma) produced three thermal highs of sillimanite grade: one in the Monadhliaths, the second in the NE Grampians, and the third affecting E Sutherland. Throughout most of the area under consideration, Grampian age isograd surfaces appear to dip at low angles.

RbSr and zircon study of ~2800 Ma Lewisian silicic gneisses from the Torridon Inlier of NW Scotland: Dyke intrusion and an open system

The Torridon Inlier of the Lewisian of Scotland contains ~30 sq. km of granodioritic gneisses. A coordinated study of this inlier, including detailed mapping, geochemical analyses and geochronological work, has indicated that Rb-Sr whole rock analyses yield “ages” that are functions of the local structural setting and of the chemistry and mineralogy of the rock, as well as of the overall geological history of the region. This Archean region has been invaded by a swarm of northwest striking dykes of Inverian (2400 to 2240 Ma) age, accompanied by new foliations and major structures. However, several structurally distinct areas, termed “pips”, 1–2 km2 in extent, are free of Inverian foliation and largely free of Inverian dykes. The 38 samples analysed for Rb and Sr scatter widely about a 2.66 Ma isochron; but give meaningful results only when subdivided according to their structural setting and lithology. A subset of silicic gneiss samples from the area most strongly overprinted by Inverian structures yields an age of 2240 ± 70 Ma with an initial 87Sr86Sr of 0.7098 ± 18. A subset of samples from “pips” further north yields an age of2790± 100 Ma (initial 87Sr86Sr = 0.7020± 10), identical to the 2780± 70 Ma U-Pb age obtained from zircon fractions separated from two of the samples. Samples from this northern district, located outside the “pips” in areas showing weak Inverian foliation, or within 50 m of Inverian dykes, or both, have grossly disturbed Rb-Sr systematics. The apparent ages of the felsic gneisses have been increased, while those of their mafic enclaves have been decreased (to as low as 1380 Ma). Field and geochemical evidence relate the increase directly to Rb loss and/or radiogenic Sr gain in the immediate vicinity of major dykes, which themselves have gained Rb and K by comparison with dykes elsewhere in the Lewisian. In terms of Rb-Sr work on complex terranes, these results must caution against direct interpretation of such data, especially in the absence of detailed field knowledge.

Petrology and geochemistry of the archean rocks of the malton gneiss complex, British Columbia

Abstract Hornblendic and felsic groups of Archean age and each of broadly similar geochemical composition across the Malton Gneiss Complex southeast of Valemount, Bristish Columbia, show detailed mineralogical and chemical variations from one geographical district to another. The complex appears to be composed of meta-igneous rocks which were originally developed as petrologically discrete units on a roughly 10-km scale, but which all belonged to one similar petrographic province. Overall they are characterized by rather high immobile trace and alkali element abundances. Their nearest geochemical equivalent appear to be found in the grey gneiss complexes of the North Atlantic Craton rather than in the granite-greenstone complexes of the Canadian Shield, but no other Archean complex possesses their overall characteristics. Some similarities in A-F-M and Q-Ab-Or content of these gneisses and those of the Laramie Range, Wyoming, are noted.