No / Although some limited consideration has been given to the possibility of links between the early medieval ceramic traditions of the Western Isles and the souterrain ware of north-east Ireland, these have tended to be framed in the context of supposed Dalriadic cultural infl uence fl owing from Ireland to Scotland. A re-evaluation of the possible relationships between these pottery styles suggests that souterrain ware might instead be seen as part of a regional expansion of western Scottish pottery styles in the seventh¿eighth centuries AD. This raises the question of what social processes might underlie the cross-regional patterning evident in what remains a vernacular, rather than a high-status, technology.

Shelf seas represent an important source of iron (Fe) to the open ocean. Additionally, shelf seas are highly productive environments which contribute to atmospheric carbon dioxide drawdown and support large fisheries. The work presented in this thesis describes the seasonal cycle of Fe in the Celtic and Hebridean Shelf Seas, and determines the physico-chemical speciation of Fe supplied from oxic margins. The results from repeated field surveys of the central Celtic Sea showed a nutrient type seasonal cycling of dissolved Fe (< 0.2 µm; dFe), which is surprising in a particle rich shelf system, suggesting a balance of scavenging and remineralisation processes. Coincident drawdown of dFe and nitrate (NO3-) was observed during the phytoplankton spring bloom. During the bloom, preferential drawdown of soluble Fe (< 0.02 µm; sFe) over colloidal Fe (0.02-0.2 µm; cFe) indicated greater bioavailability of the soluble fraction. Throughout summer stratification, it is known that NO3- is drawn down to < 0.02 µM in surface waters. This study revealed that both dFe and labile particulate Fe (LpFe) were also seasonally drawn down to < 0.2 nM. Consequently, it is hypothesised that the availability of Fe seasonally co-limits primary production in this region. At depth both dFe and NO3- concentrations increased from spring to autumn, indicating that remineralisation is an important process governing the seasonal cycling of dFe in the central Celtic Sea. In spring, summer and autumn, distinctive intermediate nepheloid layers (INL) were observed emanating from the Celtic Sea shelf slope. The INLs were associated with elevated concentrations of dFe (up to 3.25 ± 0.16 nM) and particulate Fe (up to 315 ± 1.8 nM) indicating that they are a persistent conduit for the supply of Fe to the open ocean. Typically > 15% of particulate Fe was labile and 60-90% of dFe was in the colloidal fraction. Despite being < 50 km from the 200 m isobath, the concentration of dFe was < 0.1 nM in surface waters at several stations. Broadly, the concentration of nutrients in surface waters described an oligotrophic environment where co-limitation between multiple nutrients, including Fe, appears likely. Over the Hebridean shelf break, residual surface NO3- concentrations (5.27 ± 0.79 µM) and very low concentrations of dFe (0.09 ± 0.04 nM) were observed during autumn, implying seasonal Fe limitation. The dFe:NO3- ratio observed is attributed to sub-optimal vertical supply of Fe relative to NO3- from sub-surface waters. In contrast to the shelf break, surface water in coastal regions contained elevated dFe concentrations (1.73 ± 1.16 nM) alongside low NO3-. Seasonal Fe limitation is known to occur in the Irminger and Iceland Basins; therefore, the Hebridean shelf break likely represents the eastern extent of sub-Arctic Atlantic seasonal Fe limitation, thus indicating that the associated weakening of the biological carbon pump exists over a wider region of the sub-Arctic Atlantic than previously recognised. These key findings demonstrate that the availability of Fe to phytoplankton may seasonally reach limiting levels in temperate shelf waters and that oxic margins persistently supply Fe dominated by colloidal and particulate fractions to the ocean.

This study has collected and analysed a database of place-names containing potential ethnonymic elements. Competing models of ethnicity are investigated and applied to names about which there is reasonable confidence. A number of motivations for employment of ethnonyms in place-names emerge. Ongoing interaction between ethnicities is marked by reference to domain or borderland, and occasional interaction by reference to resource or transit. More superficial interaction is expressed in names of commemorative, antiquarian or figurative motivation. The implications of the names for our understanding of the history of individual ethnicities are considered. Distribution of Walh-names has been extended north into Scotland; but reference may be to Romance-speaking feudal incomers, not the British. Briton-names are confirmed in Cumberland and are found on and beyond the fringes of the polity of Strathclyde. Dumbarton, however, is an antiquarian coining. Distribution of Cumbrian-names suggests that the south side of the Solway Firth was not securely under Cumbrian influence; but also that the ethnicity, expanding in the tenth century, was found from the Ayrshire coast to East Lothian, with the Saxon culture under pressure in the Southern Uplands. An ethnonym borrowed from British in the name Cumberland and the Lothian outlier of Cummercolstoun had either entered northern English dialect or was being employed by the Cumbrians themselves to coin these names in Old English. If the latter, such self-referential pronouncement in a language contact situation was from a position of status, in contrast to the ethnicism of the Gaels. Growing Gaelic self-awareness is manifested in early-modern domain demarcation and self-referential naming of routes across the cultural boundary. But by the nineteenth century cultural change came from within, with the impact felt most acutely in west-mainland and Hebridean Argyll, according to the toponymic evidence. Earlier interfaces between Gaelic and Scots are indicated on the east of the Firth of Clyde by the early fourteenth century, under the Sidlaws and in Buchan by the fifteenth, in Caithness and in Perthshire by the sixteenth. Earlier, Norse-speakers may have referred to Gaels in the hills of Kintyre. The border between Scotland and England was toponymically marked, but not until the modern era. In Carrick, Argyll and north and west of the Great Glen, Albanians were to be contrasted, not necessarily linguistically, from neighbouring Gaelic-speakers; Alba is probably to be equated with the ancient territory of Scotia. Early Scot-names, recorded from the twelfth century, similarly reflect expanding Scotian influence in Cumberland and Lothian. However, late instances refer to Gaelic-speakers. Most Eireannach-names refer to wedder goats rather than the ethnonym, but residual Gaelic-speakers in east Dumfriesshire are indicated by Erisch­-names at the end of the fifteenth century or later. Others west into Galloway suggest an earlier Irish immigration, probably as a consequence of normanisation and of engagement in Irish Sea politics. Other immigrants include French estate administrators, Flemish wool producers and English feudal subjects. The latter have long been discussed, but the relationship of the north-eastern Ingliston-names to mottes is rejected, and that of the south-western Ingleston-names is rather to former motte-hills with degraded fortifications. Most Dane-names are also antiquarian, attracted less by folk memory than by modern folklore. The Goill could also be summoned out of the past to explain defensive remains in particular. Antiquarianism in the eighteenth century onwards similarly ascribed many remains to the Picts and the Cruithnians, though in Shetland a long-standing supernatural association with the Picts may have been maintained. Ethnicities were invoked to personify past cultures, but ethnonyms also commemorate actual events, typified by Sasannach-names. These tend to recall dramatic, generally fatal, incidents, usually involving soldiers or sailors. Any figures of secular authority or hostile activity from outwith the community came to be considered Goill, but also agents of ecclesiastical authority or economic activity and passing travellers by land or sea. The label Goill, ostensibly providing 178 of the 652 probable ethnonymic database entries, is in most names no indication of ethnicity, culture or language. It had a medieval geographical reference, however, to Hebrideans, and did develop renewed, early-modern specificity in response to a vague concept of Scottish society outwith the Gaelic cultural domain. The study concludes by considering the forms of interaction between ethnicities and looking at the names as a set. It proposes classification of those recalled in the names as overlord, interloper or native.