Virtual Reality has become nowadays not simply a way to graphically represent archaeological hypotheses concerning space archaeology or architecture, but also a tool within the process of interpretation of remains. As it is a suitable technique to visualise 3D environments, it helps to perceive correctly volume, perspective and proportions, all features that can be hardly represented in traditional 2D archaeological drawings. The process of building a VR model will be a continuous testing of major and minor hypotheses, and formulating new questions as new needs show up. These questions lead to new interpretations, more fieldwork in search of new data and search of parallels.

This conception of VR has allowed us to propose a solution for the evolution of two close buildings in the NE area of the forum of the Roman city of Pollentia (Mallorca, Spain). Pollentia was founded after the Roman conquest of the Balearic islands (123 BC), probably between 70 and 60 BC. During the last 10 years archaeological research in the site has uncovered an area of the Forum where several walls overlap, and some of them even remained isolated at some moment, without any connection to others. Altogether makes difficult their architectural interpretation.

The building of a VR model has allowed an approach to the evolution of this area. A building (“Building A”), belonging to one of the first phases of the city in the Republic, stood northwards. In the Early Empire, a new building with a fine opus sectile pavement (“Building B”) was raised up southwards, overlapping the southern part of the Building A. Some of the walls of Building A remained destroyed to their foundations and isolated within the boundaries of Building B, while others were completely removed to clear the area for the new building. The evolution of this spot is difficult to follow until the 4^th cent. AD, when the Building B was altered, probably related to a renewal of the city after a great fire that affected the city around 270/280 AD. The abandonment is also confusing as it seems to be abandoned slowly somewhere in the Late Antique period and affected by an Early Medieval necropolis.

The interpretation of the evolution of the area E of the forum, through VR offers a feasible solution to the problem of interpreting poorly preserved structures, and it may be a good help to plan new archaeological campaigns.