• The GPR maps and profiles showed a number of interesting buried features, which were targeted for excavation.  An assessment of the accuracy of the GPR data was then made by correlating profiles and maps to the excavated features. 
• During GPR data collection in Grid 1, a very distinctive feature was visible on the radar system computer screen that appeared to be a platform with a distinct edge or wall, bounded on the west by an area with little radar reflection.  It appeared that this was an architectural feature on the edge of an open area, perhaps the garden.  Excavation of the feature began immediately, even before the data were processed.  Figure 33 below shows the uncovering of the west wall of this feature in exactly the location predicted by GPR.
  Figure 33
  

• The processed data of one of the profiles that was collected directly over this feature showed a distinct paved surface, with one step down to west and a larger drop into an area where there was no significant architecture (Figure 2).  The interior of the platform showed a number of reflections, as would be produced from many rocks in a rubble fill.

Figure 2

• Figure 34 illustrates the interpretation of the profile in Figure 2, which crossed the feature from east to west.

Figure 34

• Trench 5 was then placed to fully uncover this feature (Figure 35).  The platform and step seen in the GPR data were revealed.
  Figure 35
  

• This platform appeared to be somehow related to the water distribution systems, discovered in earlier excavations to the south (Figure 9).  It had a small water basin on its southern wall, which may have been connected to the water distribution system found in earlier excavations just to the south.  

Figure 9

• The platforms are visible in the amplitude slice-maps in the southwestern corner of the grid, in this slice from 25-50 cm (Figure 18).  They are visible in the map as red features, indicating high amplitude reflections.

• In the slice from 50-75 cm (Figure 19) the platforms can be seen as red anomalies in the bottom right side of the grid.  The blue color to the west indicates open areas with no architecture, which could be the garden areas adjacent to the platforms.

Figure 19

• A second feature that appeared to be another platform was also visible in Grid 1 to the north.  It appeared in reflection profiles (Figure 36) to be a series of walls, or perhaps a partially collapsed platform, similar to the one excavated just to the south.  This feature is also bounded on the west by the same open area.

Figure 36

• Figure 37 is our interpretation of this profile (after the main feature was excavated).  We found that what appeared to be three walls on the right end of the profile were actually a platform, similar to the one just to the south.

Figure 37

• This architectural feature was excavated in Trench 2 (Figure 38).  A platform was found, much like that in Trench 5, but a portion of its top had been robbed of stone, making it appear in GPR profiles to look like a series of walls.
  Figure 38
  

• A detailed profile crossing this feature (Figure 39) clearly shows the east and west walls of the platform, with the hole in the middle where stones were removed, perhaps to be used in construction elsewhere at the site.

Figure 39

• In Grid 2 (Figure 24), where our detailed radar profiles were collected, there were two targets for excavation: the large buried building, which we called the North Building, and the distinct strata outside the building.  The strata were postulated to be the remains of gardens in the open areas between the buildings. 

Figure 24

• Trench 6 was placed to encounter the southeast corner of the building and Trench 8 for the northwest corner (Figure 40).

• Profile 54 (File 54 ) crosses the east wall of the structure at an oblique angle (Figure 40).  The north and east walls are visible, as well as two distinct strata outside the structure, which we hypothesized might be the remains of garden soils (Figure 25).  A deeper reflection north of the structure showed up on the renderings as a linear feature (Figure 24), which we hypothesized might be an ancient water pipe.

Figure 25

• Trench 6 was placed to encounter Corner A (Figure 27).

• Corner A of the North Building was uncovered in the exact location mapped by GPR.  We were somewhat surprised, however, to find that this corner had an extending wall to the south.  A closer look at the slice-map (Figure 27), however, showed this extension, but as it was right at the edge of Grid 2 it was not at first recognizable.   To the east of Corner A (an area which had not yet been uncovered when the photo in Figure 41 was taken) two organic-rich buried soil horizons were discovered, which were predicted in the two-dimensional profiles (see Figure 25).

Figure 41

• The northwest corner of the North Building was chosen for excavation in Trench 8, where corner B was mapped (Figure 27).  

Figure 27

• Corner B of the structure was encountered in the correct location according to the GPR maps (Figure 42).  The trench was also deepened outside the structure to test other distinct strata visible in the GPR profiles.  These also turned out to be buried garden soils horizons.  
  Figure 42
  

• Although we did not recognize the hole in the southern platform (Figure 38) as such immediately, it made perfect sense after we uncovered the feature.  Our initial interpretation was that there were three walls next to each other (which didn't make much sense), because there were three distinct reflection hyperbolas (Figure 39).  We now understand that the hyperbolas were created by the highly reflective corners and top of the platform. 

Figure 39

• A closer look at the profiles adjacent to File 43 (Figure 39), would have shown this platform to be intact to the north and south.  It was the hole in the platform (Figure 38) that fooled us into believing there were a series of walls in this area.

Figure 38

• We also did not recognize the wall extension to the south of Corner A in the North Building, because it was at the edge of the grid (Figure 27).  In the future we will make our grids much larger than may seem absolutely necessary so that all features associated with the buried architecture will be visible.

• Almost all the other architectural and stratigraphic features visible in the GPR profiles and mapped were found in the locations predicted.  Many features, such as the possible deep pipe (thought to be a Roman water line) to the north of the North Building, await excavation.