Scattered, brief, but nonetheless, explicit literary references make clear that the Picts (and other northern tribes) continued to be troublesome throughout the 4th century. Although they are archaeologically invisible, Roman campaigns into Scotland took place in AD 305, in 342-3 (probably in response to trouble involving the areani, or scouts), in 360, and 367-8 (possibly in response to the so-called barbarian conspiracy), and in 382 and the 390s (references all listed in Hanson 1978). The origin of the Picts has seen considerable debate (eg. Mann 1974; Hunter 2007a), and it is generally agreed that Rome had some ill-defined role in fostering the emergence of this new polity, whether accidental or deliberate. The details remain a cause for debate and much needed future research. (see 6.3 and 6.5).
The late use of Roman forts is still to be confirmed structurally, although there are late Roman finds from a wide range of sites to the furthest reaches of any earlier Roman occupation, primarily in the form of stray finds of pottery and coins. These have never been fully synthesised, but later Roman pottery from Kintore, Cramond, Inveresk and Newstead (Wallace 2008), and late Roman coins from sites such as Birrens, Cramond, Inveresk, Newstead and Bearsden (Robertson 1983, table 2; Bateson and Holmes 2006, 162) should be noted. It is likely that some of the temporary camps in Scotland date to these late campaigns, or were re-occupied during these campaigns, but such temporary occupation is less likely to leave evidence that can be securely dated (Jones 2009a), although there are suggestions of such reuse from Kintore (Cook and Dunbar 2008, 351-3).
Coin evidence, in the form of hoards (e.g. Covesea, Fort Augustus and Balgreggan; Robertson 1978) and numerous stray finds, as well as the Traprain Treasure and a scattering of later Roman artefacts, attest contact with indigenous societies through the 4th century, though the precise character of that contact is a matter of interpretation (Robertson 1978; Hunter 2007a; Hunter 2010, arguing for the south-east of Scotland acting as a buffer zone). It has been suggested that most of the coin stray finds are modern losses (Casey 1984), although it has been argued that a reliable and useful core of finds can be extracted and used to demonstrate a Roman presence (Hunter 2010). Recent metal-detecting finds include a number of clusters of late-Roman coinage (from several tens to several hundred), for instance at Luce Sands and Springwood (Bateson and Holmes 2003, 248). Their date ranges are too broad to represent hoards and the nature of these finds is enigmatic and worthy of further study (Hunter 2010, 96-8).
The hoard of late-Roman silver from Traprain Law provides dramatic testament to continuing contacts with the late Roman world (Curle 1923). Its interpretation has been much debated; current research puts it into the mid-fifth century, but there is much still to do in order to obtain a better understanding of its nature and meaning (Painter and Hunter forthcoming; Hunter and Painter in press).
A review is needed of the extent and nature of late-Roman finds from military sites.
Fieldwork is required to follow up some coin scatters.
Completion of the full modern study of the Traprain Treasure in the context of other similar hoards would greatly assist its interpretation.