Hi,

there are numerous examples worldwide of metaluminous granitic melts that evolved towards peraluminous compositions by simple fractional-crystallization composition, provided that the degree of fractionation was sufficiently high. Fractionation of hornblende is not mandatory to straddle the meta/pera line.

Michael, have you try peritectic phase entraitment?: taking your peraluminous granite compositions and adding possible peritectic phases from the melting event (Cpx or/and Opx, Plg, etc.) and then check out how ASI evolves. See Clemens and Stevens 2012 (Lithos 134-135).

Hi, Michael:

You have to find evidence to support your idea. Cumulates are an effective way to raise ASI of the residual liquid. According to your description, the rock types of your studied area are not so simple. Are they adequately covered and represented. You need to check porphyries and aplites in addition to the main varieties of this orthogneiss. The porphyries could trap some cumulate or crystal-mush. The phenocrysts themselves are early-crystallized mineral phases that can be used to check the feasibility of your idea. Is any Al-depleted phases are present in the phenocryst or glomerocryst？ If so, it makes a good beginning.

Attached, please find a paper about MME as cumulate. But I am not sure that it could be of much help because hornblende is an critical fractionated phase which, according to your description, is absent from your samples.

First of all, thank you all for contributing to my problem! I really appreciate that.

@Dr. Förster: i am only aware of fractionation of hornblende and clinpyroxene with ASI

Hi, Michael:

If the MME and cumulate are really absent, you must consider alternative models. Is there any possibility that the two orthogness suites, melanocratic and leucocratic, are independent rock suites that were derived from different igneous suites, say, one gabbroic (ASI1). These rocks were subsequently metamorphosed or remelted to form metaluminous and peraluminous I-type granites. You idea relies strongly on geochemical data and modeld but seems lacking adequate petrographic support. You must try to do more petrographic observations, electron microprobe analysis together with adequate field work. These works are especially important because these rocks have been metamorphosed.

An alternative explanation could be alkali loss during orthogneissification processes. Have you considered it? Most orthogneisses are peraluminous, while they could come from initially metaluminous, even peralkaline, protoliths.

@Wei Liu: It is unambiguous that both orthogneisses were derived from the same source. This is shown by identical Lu-Hf isotopic signatures of zircon in both gneisses. Also the trend in incompatible vs. compatible elements shows a clear relationship which is best to be explained by fractional crystallization.

@Bernard Bonin: This is an interesting hypothesis I havent considered yet. An alkali loss might be indicated by K/Rb< 100 of both orthogneiss varieties, which I have attributed to the hydrothermal mineralization processes so far. I will think about that!!

What do your biotites look like? Are they igneous or metamorphic? Any traces of residual amphibole? Is it possible the amphibole in the orthogneiss has been completely replaced by biotite during metamorphism? If, as you state, magmatic phenocrysts are very rare, metamorphism must have been extensive.

That is true Roel Verberne and biotites have re-equilibrated during metamorphism. Some large biotites have grown syn- to posttectonically and TiO2 concentrations are low in all biotites. Additionally poikiloblastic garnet was formed only in the metaluminous orthogneisses. The metamorphic assemblage is biotite, muscovite, chlorite, feldspar, garnet, zoisite and titanite. According to other authors, the orthogneisses have experienced amphibolite-facies during Variscan and upper greenschist-facies conditions during Alpine orogeny. Therefore amphibole may have been consumed to form zoisite and chlorite. This is a good point  which deserves definitvely more attentionin my discussion, thank you Roel  

Is there petrographic evidence for metamorphic replacement of amphibole by biotite, zoisite or chlorite?  I would look for pseudomorphs and relict remnants of amphibole.

Zircon Lu-Hf isotopes cannot warrant a genetic link of leucocratic from melanocratic rock via fractional crystallization. Derivation of the two suites from separate protoliths, e.g., one gabbroic to troctolitic and another anorthitic to troctolitic, is also compatible with similar zircon Lu-Hf isotopes because the gabbroic and anorthitic protoliths were ultimately derived from, say, an arc mantle. Incompatible vs compatible element variations are argument for a fractionation link. One also need petrographic evidence. Any mineral phase as early-crystallized in the melanocratic rock and any mineral phase as late-crystallized in the leucocratic? Does a mineral phase from the two suite show evolutionary link that can be explained by crystal fractionation.

Allanite, biotite, titanite and plagioclase have early crystallized in the melanocratic variety. Biotite is nearly absent in the leucocratic gneiss and the others are rare. Biotite has a strong influence on the geochemical composition of the rocks. Addtionally, zircon are two-phase in both rocks with systematically increasing trace element concentrations indicating crystallization from a more evolving melt. Both types are clearly magmatic in origin and have identical 176Hf/177Hf which precludes any input of radiogenic Hf from an external source.

According to my data I dont see that both granites were derived from different sources. They additionally share the same U-Pb ages and they occur always together with sharp tectonic boundaries in the field.

One sample of the melanocratic type might show minute relics of amphibole in a quartz-rich matrix but I am not really sure. Next week I will take a look with the EMP on it. Pseudomorphs havent been found.

I suspect that the second "suite" on only slightly elevated in ASI, that it simply represents fractionation from the I-types...maybe amphibole cumulates at depth?

Michael, is there a trend in your biotite compositions from low Al (annite-phlogopite) toward higher Al (siderophyllite-eastonite)? If any of the biotite is primary, you should see this trend. Also you should see a decrease in the Mg# and an increase in the Mn#. It's possible for a metaluminous melt to crystallize biotite as the first AFM mineral. During fractional crystallization the next mineral to start crystallizing could be garnet. You might look at my perhaps dated article on the subject of the metaluminous-peraluminous fractionation possibilities. The paper is "AFM liquidus projections for granitic magmas, with special reference to hornblende, biotite and garnet," 1981, Can. Mineral., 19, 103-110. At the very least, the article might provide some additional food for thought. Cheers.