Readers will know that all coin collectors are in denial about the possibility that any of the objects they buy no-questions-asked came from an archaeological site. They ask us to believe that they do not buy "site finds" or off-site losses as they are not the quality they need (that's what they say, these folk most of whom have never dug any up in their lives so are just guessing what coins like these look like). They say their loot comes from undeclared hoards, and hoards were "never buried in archaeological contexts".
Except, that is, the ones that were..... Just a few weeks ago we had discussion of the Beau Street (Bath) hoard and the Seaton Down hoard near the Roman villa at Honeyditches. The Bredon Hill hoard was in a settlement area, The Frome Hoard was found on a 'productive' site and was interpreted as a communal votive deposit. The Bridgenorth hoard was interpreted in a similar manner due to the stratigraphy of the coins in the pot. On excavation of the findspot it was found it had been deposited in a settlement feature ("buried in a U-shaped gulley or ditch that formed part of an otherwise unknown Late Roman site"). The Wanborough coin deposit came from a votive site. The Cardiff Castle hoard was found in an excavation, then there's the Colchester Williams and Griffins one, the writer excavated a coin hoard himself in the Baths Basilica right in the centre of Wroxeter. The majority of hoards reported by metal detectorists in the UK are found where they are because the detectorist is exploiting (for collecting purposes) a 'productive site', in other words one which produces artefacts deposited as a result of human activity. This applies also to those found on rallies. It seems many collectors based in the US have problems envisaging this, they invariably state something like in the great majority of cases (they are misquoting PAS statistics here) coin finds are discovered within (or near) the disturbed layer in fields and other places away from historical settlements. In what way are "fields" away from "historical settlements" if the remains of those sites are actually found in what are today fields? It seems that people from the US with its landscape history of all of two centuries (and not being readers, one suspects, of any landscape history textbooks or reports) have problems conceptualising settlement pattern change over two millennia. In reality, the problem with attempting to discuss things with US coineys is with their own limited intellectual horizons and inability therefore to understand what it is they are being told.
Anyhow, to add to the growing list of hoards found on archaeological sites, we now have the "Somewhere-in-Galloway-biggest-Viking-hoard":
In the first full report on how a major Viking-age hoard was found and recovered in Galloway, south-west Scotland, we reveal that excavation suggests the treasure had been buried in the corner of a timber building over 12m (40 feet) long. The building stood within a bank-and-ditch enclosure, and may have been part of an early Christian monastic site. Research is still at an early stage, however, and the discoveries pose more questions than they answer. Among other news is that there were actually two hoards, one buried above the other.How many hoards which were not subsequently the subject of keyhole archaeological excavations were also associated with structures and stratification? Most of the hoards (even those published in various numismatic publications) are left 'floating' by the manner in which they "surfaced" (from "underground") on the market, we know nothing of their landscape context - let alone suite context of deposition. While those hoards hidden for safekeeping may have been deposited in remote spots away from human activity, more likely is that they were placed where an eye could be kept on the site - like the Williams and Griffins one right in the centre of Camulodunum. Certainly those hoards that were votive deposits, added to over a period were deposited at a place significant enough to promote that activity.
The only reason why the majority of hoard finds that coin collectors and dealers have handled are counted as not having been found on archaeological sites is because even when they have, that information is suppressed for the convenience of the no-questions-asked trade in them, thus totally distorting the "statistics".