What controls partial melting in migmatites?

Abstract The layers of six stromatic migmatites from Northern, Western, and Central Europe display small but systematic chemical and mineralogical differences. At least five of these migmatites do not show any signs of largescale metamorphic differentiation, metasomatism, or segregation of melts. It is concluded, therefore, that the compositional layering observed in most of the investigated migmatites is due to compositional differences inherited from the parent rocks. Almost isochemical partial melting seems to be the most probable process transforming layered paragneisses, metavolcanics, or schists into migmatites.
The formation of neosomes is believed to be caused by higher amounts of partial melts formed due to higher amounts of water moving into these layers. The neosomes have less biotite and more K-feldspar, if K-feldspar is present at all, than the adjacent mesosomes. These differences are small but systematic and seem to control the access of different amounts of water to the various rock portions. Petrographical observations, chemical data, and theoretical considerations indicate a close relationship between rock composition, rock deformation, transport of water, partial melting, and formation of layered migmatites.     

Genesis of Peraluminous Granites I. Experimental Investigation of Melt Compositions at 3 and 5 kb and Various H2O Activities

Melting experiments have been performed on a peraluminous quartzo-feldspathicgneiss from Northern Portugal. This gneiss occurs as xenolithsin the Tourem anatectic complex and is the most probable sourcerock for the surrounding anatectic granites. On the basis offield, petrological, and geochemical data, it can be shown thatanatexis took place in the stability field of cordierite + biotiteand that the evolution of magmas is the result of processesinvolving segregation of partial melt and separation of restiteminerals. Experiments were performed at 700, 750, and 800 ?C, 3 and 5kb, and various H₂O activities (aH₂O) to clarify the influenceof aH₂O and melt fraction on the composition of the generatedmelts. Biotite and cordierite are the two main ferromagnesianphases observed in the run products. Cordierite is formed byincongment melting of biotite. For relatively low melt fractions (< 30–35 wt. %),the partial melts coexisting with quartz, alkali feldspar, andplagioclase have a minimum or near-minimum melt composition.The melts become richer in potassium with decreasing aH₂O. Thisresult using a natural rock as starting material is in goodagreement with results achieved in the synthetic Qz-Ab-Or system.In the stability field of biotite and cordierite, the influenceof aH₂O on melt composition is at least as important as theeffect of changing P and/or T. For higher melt fractions the composition of the melt is stronglycontrolled by the disappearance of alkali feldspar and the meltsbecome richer in An and poorer in Or with increasing degreeof melting. The wide range of compositions (especially for K₂O, which variesfrom 3.5 to 5.4%) observed in the experimental peraluminousmelts demonstrates that a wide variety of granitoid magmas maybe produced from the same source rocks. The application of theexperimental results to the Tourem anatectic complex shows thatmelting occurred at H2O-undersaturated conditions (4–5wt. % H₂O in the melts, corresponding to aH₂O of {small tilde}0.5at 5 kb). Experimental melts similar in composition to the mostleucocratic granite of the Tourem anatectic complex (consideredto approximate the composition of the generated melt) were obtainedaround 800 ?C, suggesting that this temperature was attainedduring the peak of anatexis.     

Lodgement till and flow till: a discussion

A comment by J. Kriiger on I. Marcussen's paper 'Distinguishing between lodgement till and flow till in Weichselian deposits' (Boreas 4 , pp. 113–123) is followed by a reply from I. Marcussen.     

LINEAR MODELLING OF GLACIER LENGTH FLUCTUATIONS

In this contribution a linear first‐order differential equation is used to model glacier length fluctuations. This equation has two parameters describing the physical characteristics of a glacier: the climate sensitivity, expressing how the equilibrium glacier length depends on the climatic state, and the response time, indicating how fast a glacier approaches a new equilibrium state after a stepwise change in the climatic forcing. A prerequisite for the application of a linear model to a particular glacier is that length fluctuations over the period of interest are significantly smaller than the average length. The linear model is used to define and illustrate some concepts relevant to the study of glacier fluctuations. It is shown that glaciers are never in equilibrium with climate, and that a constant time lag between forcing and response cannot be expected. Next the linear glacier model is applied to real glaciers, showing how information on response times and a reconstruction of the climatic forcing can be extracted from length records. In the first application, two adjacent glaciers in the Oetztal Alps (Austria) are considered: Hintereisferner and Kesselwandferner. By optimizing the response times with a control method, reconstructed equilibrium‐line histories for these glaciers are almost identical. The corresponding response times are 31 years for Hintereisferner, and only 2.1 years for Kesselwandferner. In the second application, four glacier length records from the Oberengadin (Switzerland) are used to reconstruct equilibrium‐line histories. These appear to be mutually consistent, and the mean rise of the equilibrium line over the period 1894–2007 appears to be 1.4 m yr^?1. An equilibrium‐line history derived from data of a nearby climate station yields about the same trend over this period, but shows significant differences on the decadal time scale.     

The Soom Shale: a unique Ordovician fossil horizon in South Africa

The recent discovery of fossils with preserved soft tissues in the Cedarberg Mountains of the Cape Province has opened a new window on Ordovician life.