Excavations in 1985-86 by local archaeologists at the site known as ‘Temple of Juno” (Sezze, LT) revealed the presence, below some 1.5m of alluvium, of a layer containing finds dating from the Middle Bronze Age to the Early Iron Age (Alessandri 2009, with references). The Groningen Institute of Archaeology (GIA) investigated the site and its surrounding through manual corings and field surveys in the 1990s under the direction of Attema, establishing that the protohistoric finds layer was associated with a sandy stream bed. A layer of organic clay, possibly part of the fill of this stream bed, was dated to 3340 BP (Attema and Delvigne 2000). A more thorough geo-archaeological investigation was initiated in 2007-08, with a larger number of corings to establish a more detailed stratigraphy on-site, as well as longer geological sections across the site (Bakker 2009). Finally, in May 2009 an exploratory trench was excavated through the site by GIA in collaboration with the Soprintendenza to enable the description and ecological sampling of the full stratigraphy, and the stratigraphical and chronological correlation of the archaeological material with the geological development of the site. A trench 25m in length and 3m wide was machine excavated in five levels down to a depth of circa 2m. Standard GIS operating and documentation procedures were used. The stratigraphy was described with the help of two specialists (Sevink and Vos), protohistoric and Archaic-to-Roman pottery was described by Anastasia and Tol, dating and ecological samples are currently being processed. The excavation established that a) there are two ploughsoils present, one of which is younger than the Roman Republican period; b) at least one late 2nd c. BC votive deposition was associated with the Juno temple; c) five deep and wide post-holes (some with padstones) indicate the presence of substantial late Archaic-early Republican structures; d) at about 1.5m depth a Late Iron Age-Archaic surface containing a fire pit and charcoal, pottery and bone fragments was present; e) between 1.7 and 2m depth, a BM1-2 layer containing stones, burnt clay, charcoal, pottery and bone fragments and emmer wheat grains was present. Radiocarbon dating of charcoal from the latter layer confirmed the archaeological date: 3495 +- 45 BP. The stream bed mentioned before was found to have partially eroded this layer and, in view of the find of an almost complete dolium a cordone (similar to one found in the 1980s) in association with a piece of Alnus dated to 2925 +- 20 BP, was active in the Early Iron Age (Latial phase 2a). The paleogeographical reconstruction starts with lacustrine clays and alluvia (probably from a stream running immediately to the south of the site) underlying the Bronze Age layer. Bronze Age occupation was established on this alluvial sediment, probably near the mouth of the stream since the finds layer slopes both to the north-east and to the south-east. The stream then shifted north and, during the Iron Age, both eroded and covered the Bronze Age layer, probably with seasonal high discharges as indicated by the sediment. The deposition of a complete Early Iron Age vessel in the stream bed indicates continued or renewed (ritual?) use of the site. The stream then shifted further north and continued to deposit some 70cm of alluvial sediment before the evidence of the post-holes shown significant (ritual, in view of the previously attested finds of architectural terracottas) building activity in the Archaic and subsequent Republican period. A further 60cm of alluvium accumulated before the start of the Imperial period, for which no further evidence of occupation was found.