Anorogenic magmatism of the circum-Mediterranean area (Tyrrhenian Sea, Sardinia, Sicily Channel and Middle East) and of continental Europe (French Massif Central, Eifel, Bohemian Massif and Pannonian Basin) has been proposed to be related to the presence of one or more mantle plumes. Such conclusions based on geochemical data and seismic tomography are not fully justified because: 1) a given chemical and isotopic composition of a magma can be explained by different petrogenetic models; 2) a

more »

... n petrogenetic process can produce magmas with different chemical and isotopic composition; 3) tomographic studies do not furnish unique results (i.e., different models give different results); 4) the commonly adopted interpretation of seismic wave velocity anomalies exclusively in terms of temperature is not unique -velocities are dependent also on other parameters such as composition, melting, anisotropy and anelasticity. Tomography and geochemistry are powerful tools but must be used in an interdisciplinary way, in combination with geodynamics and structural geology. Alone they cannot provide conclusive evidence for or against the existence of mantle plumes. The existence of large and/or extensive thermal anomalies under Europe is here considered unnecessary, because other models, based on the existence of upper mantle heterogeneity, can explain the major, trace-element, and isotopic variability of the magmas. Volcanism in central Europe (the French Massif Central, Germany and the Bohemian Massif) is concentrated in Cenozoic rifted areas and is here interpreted as the result of passive asthenosphere upwelling driven by decompression. Similarly, anorogenic magmatism in Sardinia, the Tyrrhenian Sea and the Pannonian Basin is explained as the result lithospheric stretching in a back-arc geodynamic setting. The most important factors determining the locus and, in part, the 2 geochemical characteristics of magmatic activity are the Moho and the lithosphere/asthenosphere boundary depths. Where both are shallowed by tectonic processes (e.g., in rift zones or back-arc basins) passive upwelling of asthenospheric mantle can explain the magmatic activity. asthenosphere (with T p ~1280 °C) and mantle plume material can be as high as 300 °C (e.g., Richards et al., 1989; Griffiths and Campbell, 1990) , even if its ∆T p can be reduced to as low as 100 °C (e.g., McKenzie and Bicke, 1988; White and McKenzie, 1989) . The requirement of such a large ∆T p (i.e., the existence of extremely hot upwelling volumes of mantle) has been proposed to explain the huge volumes of magma emplaced in continental large igneous provinces such as the Deccan, Paranà-Etendeka and Siberian traps, Kerguelen and Ontong Java etc. (e.g., Saunders et al., 1992; Mahoney and Coffin, 1997) . These geochemical and geophysical models are based on the assumption that the source regions of large igneous provinces (volumes of magmas of the order of some million km 3 produced in relatively short time, ~1-2 Ma) are entirely peridotitic. However, during the last decade, new models have suggested the presence of lithologies (eclogites, pyroxenites, garnet granulites etc.) with solidus temperatures several hundred degrees lower than peridotic mantle (e.g., Cordery et al., 1997; Hirschmann, 2000; Yaxley, 2000; Kogiso et al., 2004) . At least in some cases, enhanced melt productivity could be the consequence of chemical anomalies (e.g., the presence of low temperature melting point assemblages) rather than thermal anomalies (as proposed in mantle plume models). As an example of questionable plumes, defined on the basis of geochemical and tomography analyses, we discuss some volcanic areas in the European and circum-Mediterranean area. The main conclusions of this review can be summarized: 1) geochemical data must be used with care because a) a given chemical and isotopic composition can be explained by different 4 models, and b) a single petrogenetic model can produce magmas with different chemical and isotopic composition; 2) tomographic studies do not furnish unique results and different studies give different results; 3) the presence of deeply rooted mantle plumes or the presence of anomalously hot upper mantle is unlikely and unnecessary to propose in the circum-Mediterranean area; 4) most Cenozoic magmatism in the circum-Mediterranean area can be considered related to passive upwelling of asthenospheric material as a consequence of lithospheric thinning in rift or back-arc areas.