Britannia is endowed with a comprehensive numismatic record. The quality and quantity of numismatic data available for study for the late 4th and early 5th centuries AD is remarkable, comprising not only hoard data, but excavation assemblages and 'site' finds. Records exist for 232 hoards with a terminus post quem of AD 388 (note: the main sources for hoards are Robertson 2000, the Coin Hoards from Roman Britain series (Vols 1-13), Coin Hoards 1–7 (Royal Numismatic Society 1975-95), and summaries in the Numismatic Chronicle (1994-2011) and British Numismatic Journal (2012 onwards), the Treasure Annual Reports (1997-) and for Wales, Guest and Wells 2007. A thorough listing and analysis of gold hoards is in Bland and Loriot 2010.). Although this is in part a reflection of Britain's established record of reporting hoards, the diocese does proportionally have a higher number of hoards from this period than any other province in the Roman Empire (note: 62% of early 5th-century precious metal treasures and 58% of all hoards containing silver coins from the period AD 300-500 come from Britain; Britain also had 24% of all bronze hoards from the period (Guest 2005, 28); 80% of all known silver coin hoards from the period 388-410 come from Britain (Abdy 2002, 62).)(Bland 1997a; Guest 1997, 411; 2005, note 33; Hobbs 2006). A large number of these hoards, for instance the Hoxne, Haynes and Coleraine hoards, include other objects such as precious metal jewellery, plate, ingots and spoons, illustrating the complex relationship between bullion and currency in this period (Guest 2005; Johns 2010; Inscker and Orna-Ornstein 2009; Robertson 2000, 405-6, 1621).
| Hoard composition | Number of hoards |
|---|---|
| Gold | 16 |
| Gold and silver | 31 |
| Silver | 90 |
| Silver and bronze | 34 |
| Bronze | 56 |
| Gold, silver and bronze | 5 |
| Total | 232 |
The data provided by hoards can be supplemented by numismatic material from archaeological excavations. Although such data are scattered through excavation reports and often not published fully enough to enable meaningful analysis, the summaries produced by Richard Reece and Philippa Walton for 140 and 368 sites respectively provide a very useful overview (note: Reece 1991 records 27,736 coins of the period AD 388-402, of which 22,822 were found at Richborough; Walton 2012 records 29,073 coins of the period AD 388-402; other corpora include Moorhead 2001 for Wiltshire, Shotter 2011 for the north-west and Penhallurick 2009 for Cornwall.). In addition, a selection of assemblages with a major Theodosian element has been generated by Moorhead (Moorhead et al. in press).
Hoards and site find data can also be supplemented by unstratified 'stray' losses, found by metal detector users and other members of the public. Since the foundation of the Portable Antiquities Scheme in 1997, over 225,000 Roman coins of this type have been recorded throughout England and Wales. These include 7532 Theodosian coins of the period AD 388-402: 8 solidi, 323 siliquae and 7198 nummi (note: http://www.finds.org.uk; the PAS data includes 52,804 coins from Wales (Guest and Wells 2007). The denominational composition of the PAS assemblage as of 23 April 2014 taken from query at http://www.finds.org.uk/). A major analysis of late Roman coin finds recorded with the PAS has been published by Walton and the recently published volume on the Traprain Treasure includes a detailed analysis of late Roman silver coins on the PAS database (Walton 2012; Bland et al. 2013).