The excavations and research undertaken from 1993 onwards in the territory of the Alto Piano di Asiago are part of the “Progetto Altipiani-Resorse”. At Bostel di Rotzo, a site known for the presence of an Iron Age village, the investigations looked at two adjacent structures with sunken floors and shared perimeter wall but different uses. One structure (C2) had a single occupation phase and was a dwelling. The second structure (C) was characterised by two phases, the earliest residential, the later productive with the installation of several clay ovens, perhaps for firing ceramic objects. The structures seem to date to a period between the second half of the 2nd century B.C. and the beginning of the 1st century B.C. A series of trenches dug to the west of these structures revealed traces of substantial terracing, which formed the basis for the settlement construction on the Bostel plateau. Another trench, east of the plateau itself revealed the presence of two walls, in two phases, belonging to the double perimeter wall of the village and built using the technique typical of fortified villages.

The research project (Progetto Altipiani–Risorse) began in 1993 with the aim of analysing, evaluating, monitoring and valorising an area of eco-cultural resources spread over the Altipiano di Asiago and the neighbouring Altipiani Trentini (relating to investigations ranging from prehistoric archaeology, ethno-archaeology, archaeology of the War, actualistic archaeology). As regards “Public Archaeology” (Public Archaeology, Community Archaeology, CRM) and of “Archaeology for development” in 1999 the project created an archaeological park and numerous didactic-formative and experimental archaeology activities. During 2009 limited excavations were undertaken in the locality of Bostel a Castelletto di Rotzo (VI) due to lack of funding. Some areas excavated in previous years were cleaned (area D, E and house A, 1969 excavations) and structure C2 was excavated. Excavation of structure C2 Following the opening of the trench and general cleaning, the investigation of the “storeroom/space below stairs” to the south-east continued and was completed. This structure had been partially excavated in 2007 and 2008. Sampling was undertaken on soil from occupation layers for phosphate analysis. During the 2009 campaign a plan was drawn up regarding the work to be carried out in subsequent seasons in order to complete the research on the two structures found, C1 and C2, and guarantee an adequate preservation. Investigation of the room south-east of structure C2 In the space between post hole USS -5107, 5106 and the dolium-hoard USS -5101, 5102, 5103, 5111 previous campaigns (cf. 2008 campaign) had identified an anomaly that was another post hole: US -5116and its fill, US 5117. The excavation showed that this post hole was probably a support for the upper parts of the structure. Given its particular position (at the centre of the room) it must have been one of the supports for a staircase leading to the upper floor of the house. There were no finds in the fill except for a small limestone (not local) slab lying flat in the bottom of the hole. This object may be interpreted as the base on which the post rested with the aim of providing some protection for the timber from humidity (cf. “flat” stones in house ’69 and C1).

In July a check (ground truth) was made of the traces recorded from aerial photographs across the entire site of Bostel and its surroundings. The excavation concentrated on structure C1 relating to a workshop, inside which a number of pottery kilns were identified. Adjacent to this structure, along its northern perimeter wall, traces of a channel on a north-south alignment were investigated, which brought the water needed for pottery production into the building.

In July 2012, research was undertaken on the ground to check the traces observed on aereal photographs across the site of Bostel and surroundings, while the excavation concentrated on structure C1. This related to a workshop inside which were a number of pottery kilns. The main aim was to complete the investigation of one of the kilns and establish its function, structure, firing techniques used, and phases of use. The workshop structure in the proto-historic village of Bostel, denominated C1 is, without doubt, exceptional within the ambit of alpine settlements dating to the II Iron Age period. The presence of four kilns suggests a craft production supplying the settlement itself, rather than just one family. It is presumed the kilns were used for firing pottery vessels, as suggested by the large quantity of vases (jars, carenated drinking cups etc) completely or in large part reconstructable, found close to the mouths of the kilns themselves. The excavated kiln, situated south-east of C1, was a simple structure not comparable to other vertical kilns with perforated firing floor, given the absence of the latter.

The research in 2013 in the territory of the Comune of Rotzo was limite to air photographic and ground survey, with 1) experimentation of the innovative procedures of Image Enhancing and Object, Pattern, and Scenery Recognition. 2) ground truthing of the anomalies. In particular, the data from the following sensors were analyzed with appropriate algorithms. • LIDAR: Provincia Autonoma di Trento • RADAR: image-intensity from Intermap Technologies Ltd. of Alberta-Canada (Learjet 36: radar with synthetic aperture (STAR). • Aereophotographic time series: 1918, 1943, 1986, Orthophoto 2000, 2006, 2008; Close Infrared 2008 • Cartographic time series: the Austrian cadastral survey of 1850-60 (1:2800); IGM 25.000 (series pre- and post- the First World War); 1918 Map at a scale of 1: 20.000 of the French Army; CTN at 1:10.000 (Provincia Autonoma di Trento- Cadastral Urbanistic and Landscape Services; and the Biblioteca Bertoliana-Vicenza; Regione Veneto - Unita` di Progetto Sistema Informativo Territoriale e Cartografia). • -eCognition and OBIA. Survey procedures. The software eCognition (Definiens 2009), among the approaches to the so-called OBIA (Object-Based Image Analysis) introduces a working method derived from artificial intelligence and artificial vision, while Object/Pattern/Scenery Recognition, extracts information from images in a (semi) automatic fashion, using a hierarchical classification of images that contrasts methods based on pixels. Among the results of the remote sensing should be highlighted the comparative analysis of the results produced by the NIR, near Infrared images (in particular the index of slope based vegetation growth). The LIDAR images produced information about slope, hillshade, skyview factor, principal components analysis, smooth/difference, pit-detection and more), RADAR (various filters, including an adaptive filter.) and from the algorythms of eCognition (aimed, for example, at the discovery of subcircular, linear and square objects). The meta-results emerging from these initial photo-interpretations concern first of all the high level of ‘spatial redundance (the indentification of the same objects using the NIR, and LIDAR, with a high level of intesection in terms of space and group) confirmed in large part by eCognition. Considering that these are techniques totally independent of the target, all together they produce a comforting confirmation of the finds. This is gratifying, but also to some extent expected, given the past work on the surface of the mountain. ( see De Guio Frizzo 2010; De Guio 2012; De Guio et Alii 2013). The most surprising and exciting results come, in particular, from the case study of Millegrobe (Vezzena). This comes from the processing of the interpherometric Radar, which, with a sequence of dedicated filters, has revealed whole new classes of objects of an extraordinary clarity. These relate to various archeological, ethnographic (paleo) environmental and wartime landscapes. Above all there is a remarkable network of coherent, frame-based linear features. Analogous results can be also be seen in the case study of Rotzo, still underway, where a preliminary analysis of the image-intensity of Intermap is showing very high resolution, particularly in the discovery of linear infrastructural features (particularly terracing and connective structures of pre- and protohistoric periods). This also sheds light on the site of Bosel, as well as its suburbs.

In August 2014, Padova University carried out excavations, directed by Prof. Armando De Guio to continue the investigation and enhancement of the site of Bostel a Castelletto di Rotzo, on the Asiago plateau (VI). The occupation levels identified during previous campaigns were exposed and cleaned, and the protective netting covering the structures was replaced, while awaiting further excavation and the creation of a grid in areas of interest. The excavations continued with the research of previous seasons, with particular attention to gaining an understanding of the occupation sequence of the so-called structures D and E. Two new trenches were opened in the south-eastern part of the site. An extensive survey, also involving the use of a metal detector, was carried out in the west field adjacent to the archaeological area.

This season, the University of Padova’s excavations continued work begun in 2014, concentrating in particular on defining the occupation sequence in sector D. This is the only area being excavated where a micro-stratigraphy is present with an _in_ _situ_ succession of the two occupation phases: Final Bronze Age (with the presence of elements from the Luco culture) and the late Iron Age (sunken house with adjacent threshing floor-terracing), separated by a prolonged gap in occupation. The excavations have not reached the interior floor surface of the sunken structure and therefore it is not yet possible to propose the structure’s function. However, another layer of collapse was identified below US 1109, probably relating to the post-abandonment phase, as the stones in the collapse are contained in an incoherent and very dark matrix with abundant organic material (US 1110 and 1111). The floor surfaces excavated in preceding campaigns in sectors C1, C2 and E1, were cleaned. The area was photographed from the ground and by drone (Phantom 3 with GoPro camera) for the reconstruction of the virtual three dimensional models using Structure from Motion (SfM), starting with a set of about 500 photographs taken at 360° inside, outside and above the sectors. The software used for processing this data (Agisoft PhotoScan Professional) is based on a technique for computer visualisation associated with digital photogrammetry. This makes it possible to create digital elevation models (DEM) for the study of the terrains macro and micro morphologies, which are than implemented with a photographic texture to obtain explorable digital reproductions. Surface, geophysical surveys, together with geo-referencing and GPS, were carried out in the fields around the site of Bostel and in the locality of Castel. The investigations around the site did not identify any significant spreads of material suggesting the presence of archaeological structures coeval with the village; however, several sporadic finds of some importance were collected, such as small lead ingots and two fragments of aes rude. Remote sensing activities continued (from satellite and plane, also with the aid of radar, and from the historical time series of aerial photographs and new ones, in particular the 1915-1914; 1943-1944; 1986; 2000; 2004; 20108; 2010; 2012; 2014 series), in addition to image processing and photo-interpretation, with various checks on the ground (ground-truth) of various types of anomalies (mainly grass, crop; soil and shadow-marks).

The excavations undertaken between 2014 and 2016 concentrated on sector D, where excavations also took place in the 1990s (largely unpublished). It is a particularly complex area, badly disturbed by cuts and dumps relating to agricultural work on the plateau, however it is of fundamental importance as it puts the occupation phases of the Late Bronze Age in relation with the better-known ones of the later Iron Age. The most interesting new evidence emerged from the removal of the fill relating to what were interpreted as two possible walls belonging to a sunken dwelling of the ‘Retic’ type. Further investigation revealed the presence of a fill containing a collapse of stones several tens of centimetres in size and two more walls attributable to a small additional room or perhaps an entrance, given its stepped formation. So far, the interior floor surface has not been reached and therefore the function of the excavated room could not be ascertained. Situated in the north-western corner of the excavation area, in close association with the room/steps there were also two structures forming a double alignment of parallel stones filled with dumped material, small clasts and boulders/stones in collapse. Although only a short section of the alignment was exposed, the most likely interpretation is that it was a structure constituted by a first ancient terrace wall, in relation to which the sunken dwelling was built, and by a later wall to reinforce the terrace, positioned opposite the earlier one to function as a containing wall. This preliminary interpretation seems supported by the use of more squared and regular larger clasts in the lower wall, built directly on top of what is interpreted as a gravel floor surface contemporary with the dwelling’s final occupation phase. Excavations also took place in the southern part of the sector, which is of particular interest for exploring the relationship between the better-known later Iron Age phase and the layers attributable to the Late Bronze Age. An alignment of small carefully selected stone slabs, perhaps forming a floor surface, was exposed in a cut into the Final Bronze Age deposit. The elements were not completely uncovered, but the material is similar to the small calcareous slabs used in the construction of the little ovens found in sector C1, although from the point of view of shape and dimensions the structures are different. The layer above the slabs was characterised by a dark organic matrix containing numerous small pieces of charcoal and anthropological material. The discovery of a fragment of the foot from a firedog decorated with small incised concentric circles and a fragment of grey pottery was determinant for dating the structure to the later Iron Age.

The 2017 campaign continued the excavation of sector D and checked several anomalies identified by geophysical survey and drone images on the ground, which led to the opening of two new areas – sectors G and F. The excavation along the east side of sector D aimed to investigate the eastward extension of the dwelling associated with the entrance identified in 2015-2016. The south wall was followed for c. 6 m exposing part of the primary collapse of the perimeter wall. However, it was not possible to reach the structure’s eastern edge before the end of the excavation campaign. The opening of three new quadrants along the western side of the excavation area revealed a new row of stones belonging to another terrace wall, slightly out of axis with the others. The excavation continued of the structure built of small slabs of _biancone_ stone identified during previous campaigns, but still not investigated in depth as it continued into the west section of the excavation area. Following the extension of the area, a quadrangular structure was uncovered. Aligned north-west/south-east, it was built of small squared slabs, of which at least three/four courses were present. Towards the south, the structure had been removed, probably as the result of a modern trench; therefore, the complete perimeter could not be identified.

This season, work took place in two main sectors of the proto-historic settlement of Bostel di Rotzo. The excavations of sector D continued, the only sector where both occupation phases of the plateau are present. The plateau was probably sporadically occupied from the 11th century B.C. onwards and permanently settled from the 6th to the 1st century B.C. The second area of investigation was sector G, identified in 2017 during a campaign of Remote Sensing but only investigated in 2018. In sector D, an additional room was discovered to the north-west of the dwelling present in this sector. The obliteration sequence for this room was very different to that seen in the interior of the dwelling suggesting that it had already gone out of use in antiquity. The north edge of the structure in the area adjacent to the new room was identified. In this zone the walls were preserved to a height of at least seven stone courses although at present the ancient floor surface has not been reached. Along the south-western side of the sector the excavation continued of a structure built of small overlying slabs of biancone stone, which is difficult to interpret. In fact, the south-western part of the structure was partially compromised. Two large lead ingots were discovered up against the slabs, probably to be associated with a phase when the structure went out of use. The excavations in sector G were extended and deepened, revealing a structure with two rooms separated by a substantial internal wall. The primary collapse of the walls and a very charcoally matrix were present inside the north room. In contrast, the south room contained few charcoal traces and the obliteration was more homogeneous and uniform, suggesting a rebuild in antiquity that saw the first room go out of use. Among the most interesting finds was an imitation Massalian silver drachma that makes it possible to date the structure to around the 2nd century B.C.

Durante gli scavi nel settore D è stato raggiunto il piano di calpestio del “vano” accessorio della struttura abitativa seminterrata. Al contrario dei crolli primari nell’ambiente principale della capanna, qui le pietre non presentano un orientamento preferenziale e manca totalmente la matrice carboniosa. Questi elementi hanno portato a ipotizzare che si tratti di un tamponamento o di un rinforzo alla struttura muraria esterna, che era preposta al contenimento del pendio. Lungo il lato orientale, invece, non è stato ancora possibile identificare con assoluta certezza la muratura perimetrale dell’abitazione indagata. Attualmente, è stata esposta la testa di due pietre in possibile allineamento che si appoggiano a un masso di grosse dimensioni. A livello stratigrafico, questo allineamento funge da spartiacque tra gli strati di crollo tipici dell’interno struttura (a ovest) e un’unità stratigrafica molto più compatta (a est) che sembra definire un esterno. L’assenza, almeno in questa prima fase esplorativa, di materiale drenante verso l’esterno del muro risulta però un’anomalia rispetto alle indagini finora condotte che avevano sempre messo in luce pietrisco sciolto di diverse dimensioni a foderatura dei tagli. Il secondo complesso indagato, situato nell’angolo SW, è una struttura quadrangolare in lastrine di biancone. Le indagini hanno rivelato che a seguito dello scasso della struttura in lastrine sono state impostate cinque pietre in allineamento comparabili con il materiale da costruzione impiegato nel villaggio. A sud di questo allineamento, la stratificazione risulta sconvolta da un intervento di epoca moderna. Per quanto riguarda il vano C in settore G, l’orientamento dei clasti, manche il colore e la tessitura della matrice mostrano chiaramente che doveva essere stato defunzionalizzato già in antico; per questa ragione, esso è stato interpretato come un possibile ingresso di fase 1, abbandonato a seguito di una ristrutturazione dell’abitazione. Il settore C1 ha rivelato nel quadrante NW i resti di un tavolato ligneo combusto impostato su una serie di 9 pietre piatte incassate nel piano basale. Al di sopra del tavolato sono stati rinvenuti i resti lignei di una porta ancora dotata di maniglia in ferro e una serie di falangi e metapodiali animali. Le indagini nel quadrante NE, invece, suggeriscono la presenza di una struttura pirotecnologica, il cui scavo non è stato ancora completato.