6.4.2 Trace elements:
Our trace element data for the Gumburanjun could only be compared with those of Herren (1993) for the larger Zanskar area, as Ferrara et al. (1992) did not analyse these elements on the Gumburanjun intrusion.
The Zanskar leucogranites quite systematically show high contents in large ion lithophile elements K, Rb, Ba, and low contents in High-field strength elements (Nb, Hf and Zr), which strongly supports a micaceous crustal source for the magmas.
The discriminant Rb versus Nb+Y diagram (fig. 6.11), not surprisingly shows that our analyses for the Gumburanjun leucogranite lie within the collisional field, as generally do the other Himalayan leucogranites (Inger and Harris, 1993).
The evolution in the content of the granites in trace elements Rb, Sr and Ba, as well as in the major elements K2O and Na2O, provides indication on the melting process of the source region (Inger and Harris, 1993; Guillot, 1993). Pelitic and quartzo-feldspathic rocks do produce melts essentially through fluid saturated incongruent melting of muscovite (Ms + Pl + Qtz + H2O => Melt), fluid absent melting of muscovite (Ms + Pl + Qtz => Als + Kfs + Melt ± Bt) or through fluid absent melting of biotite after exhaustion of muscovite (Bt + Pl + Als + Qtz => Grt + Kfs + melt). These reactions have implications for the volume of melt produced and the mineralogy of the restites, which in turn control the trace-element chemistry of the resultant magma. Hence, depending on P-T conditions and aH2O, one of these three reactions causes melts to be produced, modifying the orthose/albite ratio, and directly influencing the contents in Ba as well as the ratios Rb/Sr and K2O/Na2O in the source and the magma. In vapour-absent conditions, the amount of K-feldspar in the restites will increase, but with higher temperatures, K-feldspar will be assimilated in the melt. This will be marked by an increase in the Rb/Sr ratio and a diminution in the Ba content of the granite.
Harris and Inger (1992), Inger and Harris (1993) and Harris et al. (1995) discussed the methodology of modelling melt reactions in pelitic protoliths, and explored the geochemical consequences of muscovite and biotite melting reactions in metapelites. They showed that the three main melting reactions have a distinctive restite mineralogy that controls the trace-element distribution between source and magma and that the Rb, Sr, Ba abundance in most Himalayan leucogranites are compatible with small degree (F~12%) partial melt of the mica schists of the source regions, under fluid-absent conditions through the muscovite dehydration reaction. These authors modelled melting vectors to illustrate the relationship of source to magma produced by the three potential melting reactions (fig. 6.12). These vectors show that a concomitant increase in Rb/Sr with a decrease in both Ba and Sr is the result of the vapour-absent breakdown of muscovite. Our analyses coupled with those of Herren (1987) and Ferrara et al. (1992), clearly show such a trend and the Zanskar leucogranites can thus be interpreted as resulting from muscovite dehydration reaction. Interestingly, our samples systematically show a lower Rb/Sr ratio than those of Herren or Ferrara et al.. Such low ratios where also found by Inger and Harris (1993) for the biotite leucogranites of the Langtang. For these authors, such low Rb/Sr ratios (< 2) indicate either that more hydrous conditions occurred during melting, or that the granite is derived from a source with a higher feldspar content. More hydrous condition do however not signify that external fluids where incorporated into the system. The crystallization of melts derived from the breakdown of muscovite can liberate enough fluids to account for the production of subsequent magmas under hydrous conditions. Wet magma is not synonymous with wet source!.
It should be noted that the proportion of Rb and Sr is readily modified by fractional crystallisation and therefore these elements can only be used when the granite composition is representative of a primary melt composition. The intrasuite covariations between Rb, Sr and Ba (fig. 6.12) indicate for the Zanskar leucogranite a control by alkali feldspar. Alkali feldspar may coexist with the primary melt as a peritectic phase or alternatively might be extracted from the melt by fractional crystallisation. The study of leucogranite thin section clearly indicates that quartz overgrows mica and plagioclase and is therefore the last major phase to crystallise (Harris et al. 1993). Hence, variations in Rb, Sr and Ba are determined by peritectic alkali feldspar and not by fractional crystallization.