Search

Zircons from two Myanmar jadeitite samples were separated for texture, mineral inclusion, U–Pb dating and trace element composition analyses. Three types of zircons, with respect to U–Pb isotope system, were recognized. Type I zircons are inherited ones, yielding an igneous protolith age of 160 ± 1 Ma; Type II zircons are metasomatic/hydrothermal ones, giving a (minimum) jadeitite formation age of 77 ± 3 Ma; and Type III zircons are incompletely recrystallized ones, with non-coherent and geologically meaningless ages from 153 to 105 Ma. These Myanmar jadeitites would therefore have formed through whole-sale metasomatic replacement processes. Compared with Type I zircons, Type II zircons show typical metasomatic/hydrothermal geochemical signatures, with low Th/U ratio ( The Myanmar jadeitites, based on the present study, might have formed during the Late Cretaceous subduction before the beginning of India–Asia continental collision at Paleocene. Previously proposed Late Jurassic ages for Myanmar jadeitites are suggested as results rooted on data retrieved from incompletely recrystallized inherited zircons.

Zircon separates from one jadeitite sample (JJ) from Tone, Japan and one from Guatemala (GJ) were studied for mineral inclusions, age dating, trace-element determination and Hf isotope analysis. These zircons can be categorized into two types. Type I (igneous) zircons are characterized by the presence of mineral inclusions, among others K-feldspar, which is not present in jadeitite matrix. They also show higher Th/U ratios, larger Ce anomalies and higher 176 Lu/ 177 Hf ratios. Type II (metasomatic/solution-precipitate) zircons contain omphacite/jadeite inclusions and exhibit lower Th/U ratios, smaller Ce anomalies and lower 176 Lu/ 177 Hf ratios. Both types of zircons display high eHf( t ) values, slightly lower than the depleted mantle evolution line. The JJ sample contains both type I and II zircons. SHRIMP and geochemical data indicate that this jadeitite sample was formed through the mechanism of whole-sale metasomatic replacement at ~80 Ma from an igneous protolith of juvenile origin with an age of 136 ± 2 Ma. The GJ sample contains only type II zircons and may have formed through a mechanism of, or close to, vein precipitation at 98 ± 2 Ma. The two samples therefore testify that both mechanisms may have been in operation during jadeitite formation. Based on Hf isotope composition of type I zircons and the back-calculated REE pattern of the presumed protolith, the geochemical characteristics of the protolith of the Tone jadeitite were shown to be similar to those of oceanic plagiogranites derived from partial melting of cumulate gabbros or subduction-zone adakitic granites originated from partial melting of subducted oceanic crust. The latter, however, is a more probable candidate because the former is known to be poor in K 2 O, which, in contrast, is a notable chemical component in Tone jadeitite. On the basis of the available data, it is also suggested that the protolith, the physicochemical conditions and the extent of jadeitization may all play a role in dictating the chemical variations of jadeitites.

Major-element, trace-element and O-isotope analyses were carried out for garnetiferous Ti–clinohumite rocks and eclogites from the Kumdy Kol area, as well as talc–kyanite–phengite–almandine schists (previously named as whiteschists) and eclogites from the Kulet area of the Kokchetav ultrahigh-pressure (UHP) terrane, to explore their possible origins. Both garnetiferous Ti–clinohumite rocks and talc–kyanite–phengite–almandine schists are possibly genetically related to eclogites, as they occur in close association with eclogites enclosed within gneiss/metapelite in these two areas. Most eclogites from the Kumdy Kol and the Kulet areas exhibit chemical characteristics similar to that of ocean floor basalt, although a few show arc affinity with slight Ta and Nb depletions. Garnetiferous Ti–clinohumite rocks from the Kumdy Kol region are characterized by very high Mg, low Si, Ca, Na, and K contents compared with eclogites. They also exhibit LREE depleted patterns with negative Eu anomalies. Their O-isotope compositions are less than + 5.6‰. On the other hand, the rocks show comparable Al, Ni, Cr, Co and HREE contents as eclogites. The Mg enrichment of these garnetiferous Ti–clinohumite rocks is suggested to have originated from their mafic protoliths that had been subjected to mafic rock–ultramafic rock–fluid interactions. Talc–kyanite–phengite–almandine schists from the Kulet area are typically low in Ca and Na, but high in Mg, K, Rb, Ba, Cs, Th, U and O-isotope compositions (+ 6 to + 10‰) relative to eclogites. This peculiar chemical composition can be accounted for if the basaltic protoliths had been subjected to sequential high-temperature (enrichment of Mg) and low-temperature (enrichment of K, Rb, Ba, Cs, Th, U and O-isotope compositions) seawater–basalt interactions. One talc–kyanite–phengite–almandine schist sample KLW 7 is distinctly high in Si and K, but low in Mg and Fe, which can be attributed to more severe low-temperature hydrothermal alterations. However, except for Mg, the enrichment of Rb, Ba, Cs, Th, U and O-isotope compositions in these schist samples could also be enhanced by, or otherwise be interpreted as a result of, fluid influx sourced from enclosing metapelite during subduction.

Zircons from one jadeitite sample, collected from serpentinite melange north of the Motagua fault, Guatemala, were separated for SHRIMP-RG U–Pb dating and trace element analyses. The sub- to anhedral crystal form, lack of typical magmatic oscillatory zoning, the presence of fluid and albite/quartz/jadeite inclusions, and the low Th/U ratios ( Both the whole-rock jadeitite and its zircons have very low REE contents, ∼ 1 ppm and 0.5–42 ppm, respectively. The whole-rock jadeitite shows a flat and slightly concave REE pattern and a positive Eu anomaly (Eu/Eu* = 1.24). Zircons are enriched in HREE and their REE patterns can be divided into two groups: one with negative Eu anomaly and one with positive Eu anomaly. The latter tends to have smaller positive Ce anomalies. The fluid from which jadeitite formed probably evolved over time, becoming more reducing and more dominated by plagioclase decomposition reactions. Alternatively, trace element compositions of zircons simply demonstrate complicated variations of fluid chemistry during jadeitite formation. A reducing fluid with high pH values capable of mobilizing Al, Na, Zr and Hf is inferred to be the media during jadeite/zircon formation. If jadeitite formation was related to serpentinization fluid, the ultramafics must contain olivine as the major primary phase subjected to serpentinization, yielding reducing fluids with high pH values. Two additional requirements are (1) the presence of feldspar/mica-bearing protolithic blocks survived from rodingitization in peridotites to provide Al and Na, and (2) the presence of fluid channels as well as pulses of in/out-flux fluid to trigger metasomatic reactions. These preconditions in subduction zones may contribute to the rare occurrences of jadeitite in global high-P metamorphic belts.

To investigate factors affecting the evolution of nitrogen system in rocks during progressive metamorphism, unmetamorphosed to greenschist-facies pelitic samples were collected from both pelitic and (pelitic members of) psammitic rock sequences along two transects of the Taiwan mountain belt. Among all samples, greenschist-facies metapelites from the Tananao metamorphic basement of this mountain belt exhibit the most variable and depleted nitrogen content and δ15N values. The observations are difficult to interpret and can only be accounted for by complicated tectonic/metamorphic history of the basement complex. On the other hand, irrespective of the different metamorphic grades, samples from the metamorphosed cover strata, which have been subjected to only one stage of metamorphism, show comparable nitrogen and N-isotope composition. The only exception is that greenschist-facies metapelites from the pelitic members of psammitic rock sequence (i.e., the G(Ps) group samples) of the cover strata are enriched in the 15N isotope by at least 1‰. Taking Al content as a normalizing factor, averaged N/Al and LOI/Al show systematic decrease through progressive metamorphism. Given the non-homogeneous nature of the sedimentary protoliths, the comparable N-isotope composition among the unmetamorphosed to zeolite-facies (i.e., S-Z) group, prehnite-pumpellyite-facies (i.e., PP) group and greenschist-facies (i.e., G) group samples from pelitic rock sequences can be explained by rock devolatilization with limited nitrogen depletion (

New Sr–Nd–Pb isotopic ratios and trace element data for volcanic mafic rocks outcropping along a E–W transect in southern Italy, from Mt. Vulture to Neapolitan volcanoes, are reported. The variation of LILE/HFSE, HFSE/HFSE and radiogenic isotopes along this transect indicates that all of these volcanoes contain both intra-plate and subduction-related signatures, with the former decreasing from Mt. Vulture to Campanian volcanoes. New data are also reported for the Paleocene alkaline rocks from Pietre Nere (Apulia foreland), which show isotopic ratios mostly overlapping the values for Mediterranean intra-plate volcanoes as well as the Eocene–Oligocene alkaline mafic lavas from the northern Adria plate. Pietre Nere provides evidence for an OIB mantle composition of FOZO-type, free of subduction influences, that is present beneath the Adria plate (Africa) before its collision with Europe. After this collision, and formation of the southern Apennines, westward inflow of mantle from the Adria plate to the Campanian area occurred, as a consequence of slab break off. Interaction of subduction components with inflowing Adria mantle generated hybrid sources beneath the Vulture–Campania area, which can explain the compositional features of both Mt. Vulture and the Campanian mafic rocks. Therefore, mafic magmas from these volcanoes represent variable degrees of mixing between different mantle components.

The initial volcanic phase of Cretaceous island arc strata in central Puerto Rico, at the eastern end of the extinct Greater Antilles Arc, comprises a 6-km thick pile of lava and volcanic breccia (Rio Majada Group). Preserved within the sequence is a conspicuous shift in absolute abundances of the more incompatible elements, including Th, Nb, and the light rare earth elements (LREE: La, Ce, Pr and Nd). The compositional shift is marked by a decrease in La/Sm from averages of 2.11 in the lowest third of the pile (Formation A) to 1.48 at the top (Formation C), and by a distinctive flattening of LREE segments of chondrite-normalized REE patterns. i87Sr/86Sr and ɛNd average about 0.7035 and 8.2, respectively, in early Formation A basalts. These ranges normally overlap samples from later Formations B and C. Isotope compositions of the latter group are more variable, however, and several samples are considerably more radiogenic than Formation A basalts, such that i87Sr/86Sr averages almost 0.7042 while ɛNd-values decrease to 7.5 in Formation B and C basalts. Theoretical models of non-modal melting processes in both amphibole peridotite and spinel lherzolite sources provide insight into the origin of depleted Th, Nb, and LREE abundances in Puerto Rican basalts. Low Nb concentrations less than normal mid-oceanic ridge basalts in Formation A basalts indicate the wedge was slightly depleted by low-volume decompression fusion due to induced convection in the back-arc region prior to entry of the source into the arc melting zone. However, depleted patterns in Formation C basalts cannot be generated by relatively greater degrees of decompression fusion in the back-arc, because addition of the La-enriched slab-derived component to more depleted source material invariably produces elevated rather than decreased La/Sm. Refluxing of Formation A harzburgitic residua is similarly precluded. In contrast, the observed patterns are readily reproduced by multistage melting models involving hybridized sources containing normal Formation A lherzolite source material blended with recycled, unrefluxed harzburgite residua. Successful models require hybrid sources containing large volumes of recycled harzburgite (up to 50%) during generation of Formation C basalts. Slightly elevated radiometric Sr and Nd isotopes in a few flows from Formation C are attributed to partial refluxing of the hybrid sources within the wedge.

Late Cretaceous granitic rocks constitute an essential part of the pre-Tertiary Tananao metamorphic basement complex of Taiwan. They are dominantly of granodiorite to quartz monzonite composition. Most granitic rocks are peraluminous (A/CNK > 1.0 and normative corundum > 1%) and display moderately fractionated LREE and relatively unfractionated HREE patterns with negative Eu anomalies. On a primitive mantle-normalized trace-element diagram, they show a significant Nb depletion which is typical of the calc-alkaline magmatism from the subduction-zone environment. They fall within the volcanic arc field on the discrimination diagram of Pearce, Harris and Tindle (1984). The lack of systematic inter-element relationships suggests that the role of fractional crystallization is not significant and that these granitic rocks were derived from heterogeneous protoliths. Geochemical data suggest Taiwan granitic rocks are contaminated I-type and I-type granites related to the subduction of the Paleo-Pacific plate beneath the eastern margin of the Eurasia plate during late Mesozoic time.

The volatile transport of precious metals in sealed silica-glass tubes at 1000°C has been investigated further by varying thermal gradient and using charges of FeNiPGEAu monosulfide ± S ± FeCl2 ± NaCl ± C. Addition of FeCl2 permits FCl2 to be increased to 10−4.8 at FS2 = 00.1. The transport of PGE-Au is controlled and reproducible when the thermal gradient along the glass tubes is low and essentially restricted to the cool end. Although appreciable transport of precious metals requires the presence of metal chlorides, in addition to a high FS2 and adequate permeability, transport does not appear to depend on FCl3, further supporting our earlier suggestion of complex speciation of PGE in the vapor phase. In combination with added S and NaCl, FeCl2 does not markedly enhance the transport of precious metals over the amounts associated with NaCl alone and, in the absence of NaCl, FeCl2 results in reduced transport of Au. Transported Pd correlates with Pt (r = 0.89) and Rh with Ir (r = 0.86), but Au is only weakly correlated with Pt (r = 0.67) and Pd (r = 0.43), as is Os with Ir (r = 0.57). The AuPt and AuPd correlations improve to r = 0.87 and r = 0.66, respectively, when experiments without NaCl are excluded. Transport of PGEAu varies with composition of the base-metal sulfide matrix: FeS results in fractionation of PGE-Au with greater transport of Pt, Pd, and Au, reproducing the positive PGE-Au pattern of the Fe-rich magmatic PGE-sulfide ores of mafic rocks, whereas NiSFeS results in little fractionation, reproducing the relatively flat PGEAu pattern of the Ni-rich sulfide ores of ultramafic rocks. Late-magmatic S- and Cl-bearing fluids could selectively mobilize base and precious metals in mafic/ultramafic complexes and concentrate them in permeable pathways with the characteristic geochemical signatures observed in nature.

The metamorphosed paleo-weathering horizon(s) developed on basaltic rocks in the Tananao Metamorphic Complex forms distinct AlFe-rich rocks and occurs as two parallel belts, demonstrating its potential as a stratigraphic correlation unit in this structurally complicated metamorphic terrain. These AlFe-rich rocks have been metamorphosed under amphibolite-/greenschist-facies conditions, leading to the formation of garnet-chloritoid rocks/chloritoid-rich rocks. The chemical changes along three profiles, from metabasite to garnet-chloritoid rock, show prominent K, Rb, Cs and Ba enrichment, yet, also, a large depletion of other elements if it is assumed that Ti, Zr, Nb, Ta and Hf remained relatively immobile. The REE's, however, behaved non-coherently. These chemical characteristics are attributed to pre-metamorphic processes and are shown to be similar to those observed with submarine weathering of basaltic rocks. In contrast, the chloritoid-rich rocks reveal distinct K, Rb, Cs and Ba depletion. It is suggested that this resulted from subaerial weathering processes. K enrichment is not an uncommon feature in Paleozoic/Precambrian paleosols. The present case study demonstrates that submarine weathering should be considered as a probable cause.

Granular xenoliths (ejecta) from pyroclastic deposits emplaced during the latest stages of activity of the Alban Hills volcano range from ultramafic to salic. Ultramafic types consist of various proportions of olivine, spinel, clinopyroxene and phlogopite. They show low SiO2, alkalies and incompatible element abundances and very high MgO. However, Cr, Co and Sc are anomalously low, at a few ppm level. Olivine is highly magnesian (up to Fo%=96) and has rather high CaO (1% Ca) and very low Ni (around a few tens ppm) contents. These characteristics indicate a genesis of ultramafic ejecta by thermal metamorphism of a siliceous dolomitic limestone, probably with input of chemical components from potassic magma. The other xenoliths have textures and compositional characteristics which indicate that they represent either intrusive equivalents of lavas or cumulates crystallized from variably evolved ultrapotassic magmas. One sample of the former group has major element composition resembling ultrapotassic rocks with kamafugitic affinity. Some cumulitic rocks have exceedingly high abundances of Th (81–84 ppm) and light rare-earth elements (LREE) (La+Ce=421–498 ppm) and extreme REE fractionation (La/Yb=288–1393), not justified by their modal mineralogy which is dominated by sanidine, leucite and nepheline. Finegrained phases are dispersed through the fractures and within the interstices of the main minerals. Semiquantitative EDS analyses show that Th and LREE occur at concentration levels of several tens of percent in these phases, indicating that their presence is responsible for the high concentration of incompatible trace elements in the whole rocks. The interstitial position of these phases and their association with fluorite support a secondary origin by deposition from fluorine-rich fluids separated from a highly evolved potassic liquid. The Nd isotopic ratios of the cjecta range from 0.51182 to 0.51217. 87Sr/86Sr ratios range from 0.70900 to 0.71036. With the exception of one sample, these values are lower than those of the outcropping lavas, which cluster around 0.7105±3. This indicates either the occurrence of several isotopically distinct potassic magmas or a variable interaction between magmas and wall rocks. However, this latter hypothesis requires selective assimilation of host rocks in order to explain isotopic and geochemical characteristics of lavas and xenoliths. The new data indicate that the evolutionary processes in the potassic magmas of the Alban Hills were much more complex than envisaged by previous studies. Interaction of magmas with wall rocks may be an important process during magmatic evolution. Element migration by gaseous transfer, often invoked but rarely constrained by sound data, is shown to have occurred during the latest stages of magmatic evolution. Such a process was able to produce selective enrichment of Th, U, LREE and, to a minor degree, Ta and Hf in the wall rocks of potassic magma chamber. Finally, the occurrence of xenoliths with kamafugitic composition points to the existence of this type of ultrapotassic magma at the Alban Hills.

The platinum-group elements (PGE) are appreciably volatile in dry systems in the presence of both Cl and S at 1000°C. The present experiments were conducted in sealed silica-glass tubes, with charges of either pure metals or PGE-bearing Fe-Ni-Cu sulfides and alloys; Na halide (principally NaCl); ± S, Te, As, Bi, C, and basalt powder. Metal transport was detected by visual inspection, XRD, SEM, and neutron activation analysis (NAA). Equivalent partial pressures were estimated from the transport of pure NaCl. For single and binary metal charges, transport occurs in the following sequences: Fe > Cu > Ni > Pd; and Cl > I ⋙ F. In a series of controlled experiments with polymetallic charges and analysis by NAA, the maximum amounts of PGE transported in two days at 1000°C are as follows: Os, 70 ng; Ir, 56 ng; Rh, 150 ng; Pt, 2550 ng; and Pd, 2500 ng on an absolute basis; and Os, 1100 ppm; Ir, 610 ppm; Rh, 860 ppm; Pt, 12,400 ppm; and Pd, 5900 ppm as a proportion of Fe sulfide transported. Platinum and Pd are the dominant PGE transported in experiments with high |s2 and the chondritenormalized amounts exhibit a positively fractionated PGE pattern. The greater volatility of Pt and Pd in S-Cl-bearing systems may point to a general explanation for the positive PGE patterns of mafic rocks and magmatic sulfide and PGE deposits.

The 1500 m thick sequence of Huronian continental volcanics at Thessalon, Ontario is subdivided into 4 volcanic cycles, each of which includes abundant early mafic end-members, central intermediate flows, and late rhyolite units. Major and trace element concentrations are dominated by extensive gabbroic fractionation trends that ultimately produced two types of felsic flows: (1) rhyolites with high light rare earth element (LREE) and relatively low large-ion lithophile element (LILE) concentrations (high-LREE, low-LILE rhyolites), and (2) following late separation of REE-rich accessory phases, rhyolites depleted in LREE (low-LREE, high-LILE rhyolites). Mafic end-members of individual volcanic cycle are progressively less siliceous and less enriched in LILE and LREE with height in the stratigraphic section. Ti/Zr ratios gradually rise from 35 in early mafic flows to stabilize at about 85 in late units, while average SiO2 contents decrease from 56 to about 50% and Mg# rises from about 48 to 52. ɛ-Nd values are consistently negative, indicating variable degrees of pre-fractionation crustal contamination of the end-member magmas during their uprise through the crust. Mixing models are consistent with up to 50% contamination by crustal material of tonalitic hornblende-gneiss composition. A progressive increase in ɛ-Nd, from about-5.0 to-0.5 upward in the volcanic succession, reflects a decreasing degree of crustal contamination due to development of insulating layers along margins of the feeder system. Detailed stratigraphic variations suggest that successive magmas batches were intercepted by a progressively fractionating, periodically replenished magma source, giving rise to open-system magmatism. Despite the prevalence of crustal assimilation in the Huronian lavas, (La/Sr)N ratios are too low in least contaminated end-members to be explained by contamination of tholeiitic magmas. The late basalts resemble instead modern island are basalts, and it is suggested that the subcontinental mantle source was enriched by subduction-related processes during crustal formation. Within individual volcanic cycles gabbroic fractionation trends systematically deviate from calculated factors toward compositions characteristic of hornblende-gneiss. Such relations suggest that further crustal contamination of the magmas occurred simultaneous with crystal fractionation. probably within undulating sills at upper crustal levels. Quantitative analysis suggests assimilation/fractional crystallization (A/FC) ratios of about 0.45. As a result of extensive two-stage contamination, rhyolites from the initial volcanic cycle incorporate a total of over 60% of crust.

The Mulock granite batholith is exposed over 530 km2 in the northern part of the Grenville Province, 25 km north of North Bay, Ontario. The batholith has a UPb zircon emplacement age of 1244+4/-3 Ma. Two regional tectonic events affected the batholith: (1) moderate-grade regional metamorphism; and (2) shearing and mylonitization related to the Grenville Front Tectonic Zone which slices across the northern third of the batholith. A suite of fine-to coarse-grained, pink, gneissic granites containing accessory fluorite and with colour indices between 1 and 27, dominate the batholith and are cut by abundant aplite dikes. Anorthosite dikes cut the granites and aplite in the central part of the batholith. The granites were originally alkali feldspar granite and syenogranite, and the coarser-grained varieties were originally hypersolvus. The suite is mainly metaluminous to marginally peraluminous; peralkaline varieties are rare. Most of the suite is subalkalic, but subordinate alkalic granites enriched in biotite and ferrohastingsite are early phases. The suite shows all the major and trace element compositions and elemental ratios diagnostic of A-type granite suites. Metamorphism altered the primary distribution of some of the trace elements, particulary Rb and F. The granites crystallized under relatively low pressures (

The pre-Tertiary Tananao Schist Complex of Taiwan forms the basement rocks on which younger rocks are deposited. The Complex is composed of two belts of different lithologic characteristics: the western Tailuko Belt of high T/P metamorphism and the eastern Yuli Belt of high P/T metamorphism (Yen, T.P. 1963. Proc. Geol. Soc. China 6 , 72–74). The juxtaposition of these two belts has been thought to have resulted from the Late Cretaceous subduction process. In the Tailuko Belt, abundant metabasites occur as a mappable unit, or as enclaves in Late Mesozoic granitic intrusions. Forty-three metabasite samples were collected from the Tailuko Belt and analysed for the major and trace element compositions. The results show that most metabasites are tholeiitic in nature. These metabasites can be divided into three groups based upon their geochemical characteristics: group 1 samples resemble oceanic island-arc tholeiites; group 2 samples have elemental abundances similar to those of T-type MORBs; and group 3 samples are similar to within-plate oceanic plateau-seamount basalts. Metabasite-marble-metachert association in the belt also suggests environment(s) above/near the carbonate compensation depth (CCD). All of these results fit rather well with a marginal/back-arc-basin tectonic setting. Based on the recent PbPb dating result from the associated marble, it is suggested that the island arc/seamounts were accreted to the Southeast Asian continental margin during Early Jurassic time, and the Tailuko Belt represents a subduction/collision complex which has undergone several periods of metamorphism in association with Late Mesozoic and present tectonic events.

Petrological and geochemical data are reported for basalts and silicic peralkaline rocks from the Quaternary Gedemsa volcano, northern Ethiopian rift, with the aim of discussing the petrogenesis of peralkaline magmas and the significance of the Daly Gap occurring at local and regional scales. Incompatible element vs incompatible element diagrams display smooth positive trends; the isotope ratios of the silicic rocks (Sr/Sr ˆ 0 70406---0 70719; Nd/Nd ˆ 0 51274---0 51279) encompass those of the mafic rocks. These data suggest a genetic link between rhyolites and basalts, but are not definitive in establishing whether silicic rocks are related to basalts through fractional crystallization or partial melting. Geochemical modelling of incompatible vs compatible elements excludes the possibility that peralkaline rhyolites are generated by melting of basaltic rocks, and indicates a derivation by fractional crystallization plus moderate assimilation of wall rocks (AFC) starting from trachytes; the latter have exceedingly low contents of compatible elements, which precludes a derivation by basalt melting. Continuous AFC from basalt to rhyolite, with small rates of crustal assimilation, best explains the geochemical data. This process generated a zoned magma chamber whose silicic upper part acted as a density filter for mafic magmas and was preferentially tapped; mafic magmas, ponding at the bottom, were erupted only during post-caldera stages, intensively mingled with silicic melts. The large number of caldera depressions found in the northern Ethiopian rift and their coincidence with zones of positive gravity anomalies suggest the occurrence of numerous magma chambers where evolutionary processes generated silicic peralkaline melts starting from mafic parental magmas. This suggests that the petrological and volcanological model proposed for Gedemsa may have regional significance, thus furnishing an explanation for the large-volume peralkaline ignimbrites in the Ethiopian rift.

a Institute of Earth Sciences, Academia Sinica, Taipei, 11529, Taiwan, ROC b Graduate School of Engineering and Resource Science, Akita University, Japan c Department of Applied Arts, Fu-jen Catholic University, Hsinchuang, Taiwan, ROC d Department of Earth Sciences, University of Western Ontario, London, Ontario, Canada e Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305, USA

Oxygen isotopic compositions of whole rocks and coexisting quartz–feldspar pairs have been determined for nine pre-, and syn- to late-kinematic granitoid plutons in the Grenville Province of Ontario. These new data demonstrate that granitoid rocks (Algonquin, Mulock) in migmatite terrain of the Ontario Gneiss Segment possess normal δ18O values (18O contents (δ18O > 9.0‰), in accord with previous results.In the Algonquin sodic suite, a gross covariance of δ18O with compositional indices is present, from 6.4‰, SiO2 = 50.5 wt. % (gabbro) to 8.7‰, SiO2 = 72 wt. % (trondhjemite), resulting from combined assimilation–fractional crystallization. Mafic members of the sodic suite are 18O enriched overall (5.8–7.9‰) relative to fresh tholeiites (5.7 + 0.3‰), implicating some 18O contamination of the protolith. The dispersion of δ18O values in the Algonquin potassic suite, from 4.3 to 9.3‰, is independent of composition and attributed to isotopic exchange with low-18O thermal waters during emplacement. Biotite–hornblende granite of the Mulock batholith is characterized by a limited oxygen isotope compositional range, where the average δ18O = 8.1 ± 0.5‰; δ18O correlates with SiO2 but not with the zonal distribution of Ba, Rb, and Sr abundances.The Union Lake quartz diorite (δ18O = 8.5 ± 0.1‰) and White Lake trondhjemite (δ18O = 7.3 ± 0.6‰) have oxygen isotope compositions comparable to those of other trondhjemitic suites in the CMB. A systematic enrichment of ~1.2‰ in the Union Lake pluton, together with enhanced Ca, Mg, Fe, and Sr, can be accounted for by assimilation of ~5% marbles and 10% amphibolites from the country rock. Uniformly high δ18O values of 11.5 ± 0.8‰ characterize the Elphin granite–syenite complex. The largest values (11.7–12.7‰) and lowest SiO2 (54–56 wt. %) are in the partially assimilated host gabbro–diorite complex, endorsing the presence of 18O-enriched source regions. The Cheddar biotite–hornblende granite, one of a population of intrusions within the alkalic belt of the western CMB, has a restricted isotopic span, where δ18O = 8.8 ± 0.9‰. An unusual concave rare-earth-element (REE) distribution may result from interaction with a heavy rare-earth -element (HREE) enriched volatile phase. The Coe Hill biotite granite (δ18O = 10.4 ± 0.4‰) is isotopically in compliance with other granites and syenites of the CMB. Covariance of δ18O and SiO2, in conjunction with smooth and continuous geochemical trends, is interpreted in terms of assimilation–fractional crystallization.Peralkaline granite of the Deloro pluton includes a hypersolvus phase with high, scattered δ18O values (9.1–11.8‰) and a subsolvus counterpart attributed to late influx of water that induced isotopic reequilibration toward a more constrained range (δ18O = 9.2–10.2‰). REE distributions of a calcic syenite phase are compatible with its evolution by fractional crystallization of a low-K tholeiitic magma, and the high-18O character (δ18O = 11.1–12.6‰) requires 18O enrichment of the protolith and (or) 18O contamination of the magma. Peralkaline rhyolitic volcanics, compositionally coherent with the Deloro pluton and possibly representing extrusive equivalents, possess significantly higher and more variable δ18O values, from 11.7 to 14.2‰; this is attributed to 18O enrichment during low-temperature exchange with thermal waters, superimposed on a primary high-18O magma. The Barber's Lake two-mica granite contains enhanced abundances of U (15 ppm) and Th (36 ppm) in conjunction with systematically elevated δ18O values (10.4 ± 0.5‰). Geochemical constraints are compatible with its evolution from a trondhjemitic magma, but the isotopically enriched nature requires extensive 18O contamination of the protolith and (or) magma. These nine granites variously retain "memory" of primary and (or) secondary features, including δ18O of the source region, covariance of isotopic and compositional parameters, and sporadically superimposed disturbance by exchange with thermal waters. During metamorphism, quartz and feldspar were systematically reset to high-temperature fractionations, but the extent of open-system exchange with rock reservoirs was limited.Despite some probable disturbance by metamorphism and the limited data available, O–Sr isotope systematics of the Grenville granitoids indicate that (1) high-18O granites from the Frontenac Axis were derived from in situ anatexis of Grenville Supergroup metasediments, (2) synkinematic granites were derived by mixing of a primary magma generated at a lower crustal (granulite facies) or upper mantle level with the fusion products generated by partial melting of the Archean–Early Proterozoic type metasediments, and (3) the tonalite–trondhjemite suite in this part of the Grenville Province was derived from a similar lower crustal or upper mantle primary magma by direct fractional crystallization.