Proposals to use scans of the fragments and to employ a sorting program to reunite them “at the push of a button” proved futile, since neither the software nor hardware essential for the project was available – and even if appropriate programs could have been developed and equipment to implement them obtained, the actual scanning and sorting would, in all likelihood, have taken years, with ultimate success not assured. The archaeologists decided against attempting theoretical solutions and creation of ‘virtual’ reconstructions but rather to trust their own professional and cognitive skills. Over 27,000 fragments needed to be examined, sorted, and identified. Pre-war black-and-white photographs of the sculptures – taken during the excavations and later when they were exhibited in the museum – provided the basis for making the final joins. All of the sculptures had been carved from monoliths. This made it comparatively easy to distinguished one sculpture from another optically, simply by noticing specific features of the stone, so that almost all fragments deriving from the core of the sculptures could be attributed.
Following preliminary work lasting only three months, restoration could begin in February 2002, in tandem with the continuing work of identification and attribution of fragments. Sculptures provisionally assembled in the sorting hall, were taken apart once more, the corresponding fragments clearly marked, and the surfaces of the breaks carefully prepared for gluing with epoxy resin. The bigger the individual pieces were, the more difficult it was to reassemble the sculptures which could be as much as 2.20 m. tall and weigh tons.