I recently published an article with one of my former @findsorguk volunteers @SalisburyMuseum where they tested metal detecting methodologies and also rewrote our understanding of the Roman small town of Sorviodunum (Salisbury)

https://www.academia.edu/43137329/A_controlled_metal-detecting_survey_Revising_the_Roman_numismatic_perspective_of_Sorviodunum

It was suggested that Roman occupation ran for over 1km along the Roman road. But the coin profile has always been a bit strange site with a number of early coins and few late Roman coins. This is unusual for Wiltshire with large numbers of coins of the House of Valentinian.

The survey compared what was found with a 10% and 100% metal detecting survey to evaluate if a 10% area provides an accurate insight into a site.

When comparing the coins from the surveys it is clear that the 10% survey did offer a good indication of the coins as a whole. The research though did show still very limited numbers of late coins that we might expect. Therefore, one of the key questions was why?

When all the coins from Sorviodunum are combined the profile is very different to what we knew before with a new substantial peak in the years 330-348. The site is in an area which Sam Moorhead notes declined post 350 and we know of no late Roman hoards from Sorviodunum

It is unusual for a site which is located on the crossroads of the road network not to have late Roman hoards and Barry Cunliffe describes Sorviodunum as a failed town. Might this is explain the lack of hoards or quantities of House of Valentinian coins?

It may be that this is a wider trend seen in roadside settlements in the area as the coin profiles for three of the most prominent are very similar. Instead of seeing Salisbury as a failed town is this evidence of change in the late Roman period but not exactly failure.

All of the finds have been recorded with the @findsorguk and the survey recommends a 20% percent area as this can be done with the same set up as the 10%.