**5. Discussion**

Our integrated application of LiDAR-derived imagery and multiple geophysical survey techniques at the Johnston Site is formulated to help untangle the complex palimpsest of pre-Contact American Indian use and construction of the landscape. Simultaneously, we implemented these techniques to interrogate the validity and preservation of built features as mapped in 1917, and in doing so, also tried to better understand the impact of modern American agricultural practices on this important site. Our results pushed us toward a more detailed understanding of the Johnston Site that no single method we employed could have offered on its own. This multi-staged approach allows us to revise the reported dimensions and forms of mounds at the Johnston Site and work toward new research questions for future work. From this perspective, in the face of increasing erosion around the edges of the Johnston site, we can use our results to provide new information toward e ffective site conservation.

#### *5.1. Toward a New Map of the Johnston Site*

One of the primary goals of our research was to better situate the Johnston Site within understandings of the broader Middle Woodland landscape along the SFFDR. This required a formal interrogation of the 1917 map based on the work of Buck and Myer. Because of heavy vegetation, our use of LiDAR-derived imagery o ffered the only way to reassess the built environment at the northern-most portion of the site. In this area, we have strong topographic evidence that Mound 1 is both authentic and similar in size to the dimensions reported in 1917 (see Tables 1 and 3). We identified Mound 1 as nearly 2 m smaller in the base diameter but nearly the same height. One explanation for these di fferences in size could be attributable to the erosion of the mound surface, but this should lead to a larger base diameter. Therefore, our measurements may just be more accurate than those made in 1917, or the surveyor Buck may have directed less attention to mapping Mound 1 because it was not as large as other mounds at the site (e.g., Mounds 4 and 5).


**Table 3.** Revised dimensions and shapes of mounds at the Johnston Site based on work presented here.

Di fferences we can identify with Mound 2 are more complex to describe. A small rise is present in the location where this mound was mapped in 1917; however, the rise and the base dimensions of this rise are much smaller than what was reported previously. The placement on the edge of the Johnston terrace may indicate that much of this mound was lost as the terrace retreated. Such active erosion is visible today [83]. Alternatively, this topographic feature may be a natural relict of the landscape and its proximity with Mound 1 may have led Buck and Myer to interpret it as a. Further research is needed to confirm the nature of this rise.

The parallel embankment walls that are described from Mounds 1 and 2 south to Mound 4 are also hard to explain. In the northern reaches of the site there is strong topographic evidence for the eastern wall, and good evidence for the western wall. However, as the walls extend south into areas where we have gradiometer and magnetic susceptibility coverage, the topographic evidence is minimal and there is no correlation to geophysical signatures where the walls should be. This leaves us to question whether the history of modern agricultural plowing at the site has destroyed topographic evidence for the embankments in most unforested portions of the site. If the embankments are present near Mounds 1 and 2, but are not clearly visible further south, the early map may have continued the walls for consistency. The area northeast of Mound 3 and associated with the natural spring and western opening of the embankments shows extremely high magnetic susceptibility values for the entire site. This lends some validity to there being an important activity area in this portion of the site, but o ffers little support for walls or an opening to any walls here.

The location of Mound 3 as described in 1917 exhibits a clear conical rise with a nearly identical height, but like Mound 1 the base dimensions are o ff by nearly 3 m. This implies an error in the 1917 base measurements for this mound since any erosion should have led to a potentially larger, rather than smaller, mound base area. The two structures on the 1917 map labeled 'A' and 'B' corelate with high magnetic susceptibility. In the case of 'A', a lightning strike is visible in the gradiometer data. This suggests that the surveyors may have seen reddish soil discoloration in these areas and mapped those patterns as archaeological features. This said, we cannot definitively assign such an interpretation to 'B' because we only know that it is associated with high magnetic susceptibility. Extending our gradiometer coverage north will o ffer more insights into whether the 'B' structure is also related to a lightning strike.

There is no topographic or geophysical evidence for any of the streets or walls that connect Mounds 4 and 5, or those associated with mounds further south. However, our work does o ffer new insights into Mound 5, specifically its history and shape. Our topographic imagery shows that the base of Mound 5 is 2–3 m smaller than previously recorded. In contrast, the surface area is larger by about 3–4 m. This likely relates to the impact of plowing around and over this small platform. Our gradiometer data o ffers clear evidence for a ditch enclosure present before the construction of Mound 5. Our test excavations confirmed this feature and show preliminary evidence that it was refilled with anthropogenic materials. If the deconstruction of the enclosure and the construction of Mound 5, a platform mound, marks an important shift in the use of this space, then we can begin to build new research questions for the Mound 5 locale. For instance, Middle Woodland enclosures have been interpreted as collective monuments that imply an internal exclusivity for those who enter and use their interior spaces. These monuments are built to enclose an area, interpreted as creating a perception of 'us and inside' versus 'them and outside', and have been considered a place where people from diverse geographic scales participated in specialized ritual events [35,92–96]. Alternatively, Middle Woodland and Late Woodland platform mounds, while also considered monuments emphasizing collective notions of society, are commonly interpreted as socially inclusive because of their association with the remains of feasting [41,97–99]. We reference Late Woodland (ca. AD 600–1000) platform mounds here also because we have no relative or absolute chronological information for Mound 5; we only know that it post-dates a Middle Woodland enclosure.

The current topographic data for Mound 4 shows that the base dimensions are less by 3 and 4 m, while the surface dimensions are greater by 3 to almost 5 m. This may be a combined result of erosion and deflation of the mound surface, leading to a larger surface platform, while plowing around the mound would move mound sediments around the field, causing the base to decrease in size rather than increase from the accumulation of eroded surface sediments. Magnetic susceptibility from the surface of Mound 4 shows a combination of high and low values. These di fferences may relate to variation in mound fill used to construct Mound 4 or it may represent features associated with the use of the mound summit.

Our examination of Mounds 6 and 7 sugges<sup>t</sup> that remnants of these features are present and can be represented geophysically by high magnetic susceptibility. Topographic signatures from these features show they have eroded measurably, with Mound 7 in danger of vanishing completely. The combined topographic and geophysical data for Mounds 8 and 9 are hard to interpret. The LiDAR-derived imagery shows small rises in these locations that are slightly lower than the 1917 heights and roughly 20 cm larger in base diameters, suggestive of deflation from plowing. The gradiometer data show several isolated magnetic highs associated with these small rises. The EMI data over Mound 8 identified a high conductivity feature at the center of the rise but none of the datasets clarify whether these are mounds.

Work related to Mound 10 allowed us to revise the footprint of the monument, from round to rectangular; however, its height is reduced by approximately 30 cm, probably related to plowing indicated by the several plow scars traversing the mound in the gradiometer data. The gradiometer survey also revealed internal features that our preliminary excavations have confirmed were associated with use of the Johnston landscape before Mound 10 was constructed.

Other aspects of our work that allow us to create a new map of Johnston include the circular embankment feature along the western edge of the Johnston landform and several activity areas denoted by both sets of high surface magnetic susceptibility values. Clusters of isolated high magnetic features that likely represent subsurface pit remnants or large posts also adds to our new map of the Johnston Site and the broader understanding of land use. The potential circular structure west of Mound 10 is another important feature we have identified at the site. Our new awareness of the Johnston landscape, illuminated by our integrated remote sensing approach, allows us to formulate new research questions for the site.

#### *5.2. Beyond Never-Never Land: Developing Future Questions for the Johnston Site*

Our research has identified several di fferences between the 1917 map of Johnston and its present condition. However, for the most part, we can correlate topographic rises or measurable geophysical trends with features mapped at the site in 1917. Only the 'streets' and 'walls' linking Mounds 4, 5, 8, 9, and 10 are indistinguishable using the methods we employed. Therefore, we think it is important to move beyond questions that undermine the validity of Buck and Myers' work and move forward with new research questions for the site. To this end, we propose that additional geophysical survey methods have the potential to help elucidate issues related to the parallel embankments running north-south between Mounds 1 and 2 and Mound 4, as well as the streets and walls we cannot identify at all. Recent large-area GPS-guided EMI surveys have shown to be e ffective in teasing apart buried archaeological features that are not detectible using magnetometry [100]. The ability to map patterns of earthen conductivity across the entire Johnston Site might provide some evidence of the parallel embankments if they were built using soils high in clay. Even residual clays from an embankment construction would retain more moisture than surrounding soils today, potentially making them detectable via conductivity surveys. Additional geophysical surveys, employing methods like ground-penetrating radar (GPR), over the summits of mounds at Johnston has the potential to identify features associated with the uses of the summits. This would be particularly useful at Mounds 4 and 5 to determine if any mound-top structures are evident, or to identify features related to feasting activities.

Beyond additional geophysical surveys, we find the rate of erosion evident along the northern and western edges of the site troubling. If Mounds 2 and 7 are indeed indigenous constructions, some amount of these earthen mounds have probably eroded away already. To both determine whether these mounds are indigenous and examine how much is left, we propose that future 'cut-bank' examinations would be important. Cutbank geophysics that employ magnetic susceptibility have already proved to be successful in identifying cultural layers and archaeological features eroding into river courses [101]. Adding cutbank geophysics to well-known geoarchaeological methods like sequential loss-on-ignition, particle size analyses, and soil micromorphology is likely to confirm the nature of Mounds 2 and 7.

An additional avenue of future research we propose is important for better understanding the Johnston Site, as well as the larger SFFDR landscape, which relates to the temporality. The mounds, activity areas, pit clusters, and subsurface structural remnants we have identified here appear to represent a long history of landscape use in this area of western Tennessee. There are good reasons to believe that a lot of this evidence for human occupation at Johnston is not all Middle Woodland in age. For instance, the hickory nut roasting pits we have identified north of Mound 5 may relate to a pre-Middle Woodland hunter-gatherer use of Johnston. The clay cooking ball recovered from construction fills at Mound 10 is indicative of Late Archaic occupations at the site. Moreover, we have already identified a relative chronology for shifting forms of space relating to the change from enclosure to platform mound at Mound 5. However, we do not ye<sup>t</sup> know when that shift took place, or how long each monument was in use. Therefore, untangling the uses of Johnston through time should reveal significant changes related to social complexity and the palimpsestic history of human occupation on this landform.
