During the Tertiary Period, the northeastern Apennines of Italy functioned as a quiet depositional basin for fine-grained, volcanic ashfalls from distant explosive volcanism accompanying the Alpine/Apennine orogenesis. These ashes have been dated by several authors with different radioisotopic methods in the past decade (K/Ar, 40Ar/39Ar, and Rb/Sr on biotite separates; 40Ar/39Ar single-crystal laser fusion on plagioclase and sanidine; and U/Pb on zircon) and provide the means for a precise, accurate age calibration of the Tertiary magnetostratigraphic and biostratigraphic timescales. However, the provenance of these tephras has remained uncertain. To resolve this problem, we have carried out detailed grain size analysis of the so-called ''Livello Raffaello,'' a feldspar-bearing bentonite recognized in numerous outcrops throughout the region at the very base of the Aquitanian to Burdigalian Bisciaro formation. The Raffaello, stratigraphically located in the upper part of the planktonic foraminiferal Zone N4 and calcareous nannofossil Zone NN1, upper Chron 6A-r (middle Aquitanian), yielded an isochron age of 21.2¿0.5 Ma (2&sgr;) from 27 40Ar/39Ar dates by laser fusion on plagioclase.

We measured the mean grain size M&PHgr; of the >63-&mgr;m felsic fraction (i.e., feldspar and quartz) of the Raffaello in 11 representative outcrops in the region and compared the grain size against distance from two possible volcanic sources; the Venetian province (northern provenance) and the Sardinian province (western provenance). While the Venetian source shows no distinct trend of the M&PHgr; distribution, the Sardinian plot exhibits a very clear linear grain size decrease with distance. After having tentatively established a Sardinian source (no other volcanoes of Aquitanian age are known west of the Apennine basin), we revised the M&PHgr; distribution through palinspastic restoration, including the post-Aquitanian anticlockwise rotation of the Sardinian microplate and consequent orogenic shortening of the northeastern Apennine fold-and-thrust belt. For this we have used the ''thick-skinned'' model of Lavecchia et al. (1984), which envisions a shortening for the northeastern Apennines of 3 to 30 km, and the ''thin-skinned'' model of (Bally et al., 1986), which assumes a variable shortening of 50 to 250 km. We have also modeled the distribution of the M&PHgr; of felsic crystals for modern analogue stratospheric westerly winds blowing in this region from fall to spring at speeds of 10 to 30 m/s and altitudes ranging from 25 km to 40 km, which account for distal (>400 km) transport of fine tephra (Cornell et al., 1983).

In addition, we have calculated the M&PHgr; of felsic crystals from six representative sites of the Quaternary Y-5 ash, which was produced 38,000 years ago by an explosive volcanic event in the Neapolitan area and was distributed downwind throughout the eastern Mediterranean basin. Our results indicate that restoration by the thick-skinned model yields a crystal size distribution gradient incompatible with that predicted by the computed model and overall M&PHgr; values much finer than those observed in the Y-5 sample at a comparable distance from the source. On the other hand, restoration using the thin-skinned structural setting yields good agreement with that of the computed model and with the observed M&PHgr; distribution of the Y-5 ash.