*3.2. Archaeobotany*

The soil samples in this investigation were selectively collected during the 2017 and 2018 excavation campaigns.

For each archaeological layer—also defined by Harris [60] units of stratification (US)—, soil samples equal to about five litres were taken manually and randomly, while only in a particular case—that is, the layer of burnt soil relating to the early medieval furnace—was a total sampling of the layer carried out. The stratigraphic units were analyzed, dividing them by the three chronological phases, but this analysis subsequently focused only on the two medieval phases.

The samples were water floated at the Botanical Garden of Palermo, using two different sieves (5 and 1 mm) to separate the clayey matrix of the soil. Subsequently, the samples were screened using a binocular microscope at the Department of Biological Chemical and Pharmaceutical Sciences and Technologies (STEBICEF) of the University of Palermo.

For each stratigraphic unit, quantitative data (absolute number of fragments) and volumetric data (volume of each wood charcoal remain) for each species were measured. The comparison between the quantitative and volumetric numerical data resulted in a deviation of up to 16% per species. In most cases, for small quantities, this could be considered negligible, while in large quantities, the difference increased exponentially; for example, in the case of *Quercus ilex* in the second phase.

Only a significant fraction of the stratigraphic unit was observed by microscopic analysis rather than the whole record, considering that the variability in excess of 13–15 wood charcoals usually did not increase [41].

A total of 436 wood charcoals were observed for the three archaeological phases, and among them, 346 samples were related to the two medieval phases analyzed in this work. For the Period 1 (late 8th–9th centuries AD), 276 samples were analyzed (from earlier to later: Phase 1, US 43, 47, 49, 52; Phase 2, US 14, 40, 42, 50, 61; Phase 3, US 9, 12, 37; Phase 4, US 8, 36) and 70 samples (US 1, 2) for the Period 2 (10th–11th centuries AD). The chronology of these phases was based on stratigraphic sequence and radiocarbon analysis. Of these, 19% were not taxonomically identified (due to the small size of the fragments in some US or the state of preservation, where combustion had strongly deformed the xylem).

To identify the genera/species, a comparison was made with the reference atlases [61,62], with some scientific papers and with the two available online tools—InsideWood [63] and Microscopic Wood Anatomy [64]—as well as with some reference samples which were not available for all species.

Each sample was observed in the three transversal, tangential and radial sections at a maximum of 40x and measured to obtain a value for the total volumes by species, and therefore their incidence not only numerically but also regarding their actual size with reference to the total was measured. Whenever possible, the diameters of some small branches were measured.
