The Neogene-Quaternary magmatism of the northern Apenninic arc took place in four phases separated in space and time which become progressively younger from west to east: Phase I, approximately 14 Ma; Phase II, 7.3-6.0 Ma; Phase III, 5.1-2.2 Ma; Phase IV, 1.3-0.1 Ma. This magmatism is the result of the activation of three physically separate sources: (1) the Adriatic continental crust, extracted from the mantle in the late Proterozoic; (2) a strongly refractory, recently K-enriched harzburgitic mantle located in the mechanical boundary layer (MBL) of the lithosphere; and (3) a recently metasomatized, cpx-rich mantle, compositionally variable from lherzolite to wehrlite-clinopyroxenite, interpreted as an ephemerally K-enriched asthenosphere. The Adriatic continental crust is the dominant source of the acid plutonic and volcanic rocks of the Tuscan region. The acid magmatism is mostly found inside an ellipsoidal area (about 150 x 300 km) centred on Giglio Island, here defined as the Tuscan Crustal Dome. Within this area, mantle-derived magmas unaffected by important crustal contamination processes and mixing with crustal anatectic melts have so far not been found. Pure crustal magmas are rare but are represented, for example by some of the San Vincenzo and Roccastrada rhyolites. Virtually all the Tuscan acid centres show evidence of mixing with potassic mantle-derived magmas. Major and trace elements, as well as Sr-87/Sr-86 and Nd-143/Nd-144 data, on primitive rocks (Mg# > 65) reveal two groups of mantle-derived magmas. These define two distinct mantle enrichment trends, both essentially due to the additions of K-rich components which metasomatized separate, compositionally diverse, upper mantle sectors. In both cases the most remarkable mineralogical effect of these enrichment processes is the production of variable amount of phlogopite through reaction between fluids and/or melts with the mantle. The rocks of group I (ol-hy and Q-normative, lamproites, ultrapotassic high-Mg latites, ultrapotassic shoshonites and shoshonites: saturated trend) are considered to be derived by partial melting at low pressure (< 50 km) of strongly (lamproites) to moderately depleted phlogopite harzburgitic sources produced by reaction of residual peridotites with a K-Si-rich, Ca-Sr-poor melt with high ratios of Sr-87/Sr-86 (> 0.717), Ce/Sr (> 0.3) and K2O/Na2O (> 6-7), and low ratios of Nd-143/Nd-144 (approximately 0.5121-0.5120) and Ba/La (< 20) ratios; it is proposed that this component was formed by partial melting of subducted carbonate-free material of the upper crustal reservoir (e.g., non-restitic felsic granulites). This material is very common in the central Mediterranean region either as granitoid plutons/terrigenous sediments or as metasedimentary, non-restitic lower crust. The primitive rocks of group II are critically undersaturated, mostly leucitites, tephritic leucitites, leucite basanites, melilitites (undersaturated trend). Experimental petrology suggests that these rocks were formed by partial melting of a variably enriched phlogopite, clinopyroxene-rich mantle at higher pressure than group I primitive magmas. Trace-element modelling indicates that three components were involved in the genesis of group II mantle source: (a) a typical MORB-OIB-like mantle; (b) a component with very high Sr, Ca and Sr/Ce values and very low silica and sodium content, probably carried by a carbonatite melt somehow related to subducted marine carbonates; and (c) a recently added K-rich, Ca-Sr-poor crustal component, relatively well constrained to high Sr-87/Sr-86 (> 0.712) and K2O/Na2O (> 8-9) values, and low Nd-143/Nd-144 (< 0.51205), Ba/La (< 20) and Ce/Sr (> 0.10) ratios. These constraints do not allow to exclude a complete identity between the K-rich components which metasomatized the mantle sources of the saturated and undersaturated trend magmas. The geochemical and isotopic features of the components that metasomatized the mantle sources of the northern Apenninic arc magmatism can be explained by a geodynamic process which causes a large amount of crustal materials to be incorporated within the upper mantle. We propose that the delamination and subduction of the Adriatic continental lithosphere related to the still ongoing northern Apennine continental collision provide a viable mechanism to explain the genesis and eastward discontinuous migration of the magmatism in central Italy. The subduction of delaminated lithospheric mantle with lower crustal slivers would have exposed uppermost mantle (Adriatic MBL) and crustal units previously imbricated in the Apennine chain to the heating advected by the upwelling of a recently and ephemerally K-enriched asthenospheric mantle wedge and by the underplating of magmas derived from it. We consider that the diapiric uprising of a hot, crustally contaminated asthenosphere occurs in the wake left above the sinking of the Adriatic delaminated/ subducting continental lithosphere. The delamination/subduction process of the Adriatic lithosphere has probably started in the Early-Middle Miocene, but earlier than 15-14 Ma ago, as indicated by the age and petrologic characteristics of the first magmatic episode (Sisco lamproite) of the northern Apennine orogenesis.