34 sites classified as duns have undergone some degree of excavation and as with forts only a minority of these, four, have associated radiocarbon dates. Kildonan Bay (NMRS No. NR72NE 5) was revisited after its initial excavation specifically to obtain radiocarbon dates, these indicating occupation between the 5th and 8th centuries AD. The dates were reinforced by the few diagnostic finds from the original excavations (Peltenburg and Hood 1979; Peltenburg; Barnetson and Turner 1984). Roman pottery, however, along with what is likely locally produced Iron Age pottery, was also recovered, indicating that the dun, as suggested in the original published report, ‘may have been built before the second century AD’ (Fairhurst 1939). More recently later 1st millennium BC dates have been obtained from the duns at Loch Glashan (4th – 1st centuries BC, NMRS No. NR99SW 8)), Barnluasgan (dun 4th -1st centuries BC and later enclosure 2nd century BC – 1st century AD, NMRS No. NR79SE 17) (Figure 91) and Balure (2nd century BC – 1st century AD, NMRS No. NR78NE 36), (Henderson and Gilmour 2011; Regan forthcoming) (see Case Study 8 :Balure Dun).
Again, as with forts, the dating of other dun sites relies on associated finds derived from the excavation of occupation material. Possibly one of the earliest dates of occupation comes from the vitrified dun at Rahoy (NMRS No. NM65NW 2), this dated to the 4th – 3rd century BC based on the recovery of a looped and socketed iron axe head and part of a La Tene bronze brooch, this relatively early date perhaps underlined by the recovery of saddle querns (Childe and Thorneycroft 1938). A saddle quern was also recovered from the excavation of Clachan Ard (NMRS No. NS05NW 3) on Bute, which might also suggest an occupation date before the 3rd century BC (Marshall 1934). A comb fragment recovered from Dun Scalpsie (NMRS No. NS05NE 4) might indicate a similarly early date if the suggested early Iron Age parallels are proven to be correct (MacCallum; 1959, 1963).
As at Kildonan, the recovery of datable Roman artefacts often proves to be the only diagnostic finds that date Iron Age occupation of dun sites. At Kildalloig, Glenramskill (NMRS No. NR71NW 11) for example, the recovery of a bronze fibula and a spiral ring date the occupation of the site to the 2nd century AD (Bigwood 1964), but the fibula at least was recovered from below external paving that may be secondary or even tertiary and there were numerous secondary internal structures. The Roman pottery, ring headed pin and strap end from Dun An Fhuerian (NMRS No. NM82NW 9) were examined sometime after the initial excavations (Anderson 1895a; Ritchie 1974). The finds derived from a midden below the dun and as such do not directly date the structural remains although they indicate occupation from the 1st – 2nd century AD to the middle of the 1st millennium AD. Roman pottery has also been recovered from Ardifuir and Dun Fhinn (NMRS No. NR63SE 10) which suggests a date of construction prior to the 2nd century AD (Christison 1905, Bigwood 1964), and at Ardifuir may relate to secondary occupation. The excavation at Dun Fhinn also produced a glass toggle bead and similar beads have also been recovered from the Dun at Ronachan Bay (NMRS No. NR75SW 7) and the fort at Dunagoil (Peltenburg 1979; Harding 2004b). More recently beads have been recovered from the dun at Balure (see Balure Dun: A Case Study) (Figure 92) and from an unenclosed settlement site at Kilninian (NMRS No. NM34NE 35) on Mull, both of these sites returning 2nd – 1st century BC dates (Regan forthcoming; Ellis pers. comm.). The resultant research on these beads and others recovered from other parts of Scotland, the Isle of Man and Ireland has shown that the beads are of localised manufacture re-using imported Roman glass.
A ring headed pin recovered from Dun Beag Vaul (NMRS No. NM04NW 4) perhaps suggests a date before the 4th century AD as does the recovery of decorated insular Hebridean pottery (MacKie 1963b). Decorated pottery has been recovered from excavations on Dun Nighean (NMRS No. NL94SE 1) and Dun na Cleite (NMRS No. NL93NE 5) on Tiree, while finds of similar decorated pottery have been recovered from An Dunan (NMRS No. NM04NE 9) and Dun Beag (NMRS No. NM04NE 3), also on Tiree, along with Dun Beic (NMRS No. NM15NE 5) and Dun An Achaidh (NMRS No.NM15SE 5), both on Coll (Piggott 1954; Holley; 1994a and b, 1996a and b). The recovery of undecorated pottery from Dunan Nan Nighean (NMRS No. NR49NW 5) on Colonsay led Piggott to suggest an occupation date of ‘the last century BC or early centuries AD’ (Piggott 1951). Small quantities of undecorated pottery have also been recovered from Leccamore South (NMRS No. NM171SE 2), Kildalloig (NMRS No. NR7NW 11), An Caisteal (NMRS No. NM32SE 2), Dun Mhic Choigil (NMRS No. NR63SE 21) and Balure (MacNaughton; 1891, 1893; Bigwood 1964; Fairhurst 1964; Hedges and Hedges 1977; Regan forthcoming) (see Balure Dun: A Case Study). The duns at An Caisteal and Leccamore also produced rotary querns which have also been recovered from a number of other sites such as Druim an Duin(NMRS No. NR79SE 1), Dun Chroisprig (NMRS No. NR26SW 5) and An Dun (NMRS No. NM92NW 3) and suggest occupation after their introduction sometime in the later half of the 1st millennium BC; but their use is a long one and without further dating evidence can not firmly place any of these sites in the Iron Age (McArthur 1873; Christison 1905; Newall 1966; Betts 1969). The excavation on Dunan Breac (NMRS No. NR85NE 17) produced iron slag and a few stone objects but no datable artefacts while that at Suidhe Chennaidh (NMRS No. NN02SW 1) produced only bones and charcoal (Graham 1915; Christison 1891). Similarly, excavations at Eilean Bhuide (NMRS No. NS07NW 04) along with smaller evaluations at Laganrure (NMRS No. NR78NE 38) and Castle Dounie (NMRS No. NR79SE 13) failed to produce readily datable artefacts (Balfour 1910; Maxwell 1941; Regan; 2006, 2011).
The recovered artefacts from the dun at Ugadale Bay (NMRS No NR72NE 9) indicate occupation of the site from the 8th century AD and perhaps sporadically into the late medieval period, although as the excavator pointed out the investigations were limited in nature and produced no firm dates for the actual construction of the dun (Fairhurst 1956). Similarly, the early historic artefacts recovered from Dunollie (NMRS No. NM83SE 11) and Eilean Righ I (NMRS No. NM80SW 5) along with the medieval finds from MacEwans Castle (NMRS No. NR97NW 1) while indicating occupation in those periods, may not necessarily date the primary construction or occupation of these structures (Alcock 1978; Brown and Cowie 1987; Marshall 1982).
Read the related case study Case Study 8: Balure Dun Excavation