The present study endeavours to explore the possible survival of British Latin in subRoman Britain. Through a detailed socio-historical analysis it argues that Roman Britain was much more deeply Latinised than hitherto assumed; widespread bilingualism with Latin, and in some cases outright monolingualism in Rome's language, is shown to have extended beyond the army, the upper classes and the cities, and to have in fact become common also in rural districts at least from the 3rd century onwards. To this end, deeply-entrenched beliefs on the nature of British Latin are discussed and dispelled through comparisons with the extant epigraphic evidence; estimates of the geographical distribution and quantitative impact of the alleged Latin speakers are also attempted. On the basis of archaeological evidence this research also undertakes to demonstrate that late 4th-century Germanic settlements may have taken place, in full accordance with Roman authorities, both near the so-called 'Saxon-Shore' fortifications and in the proximity of major settlements throughout the length and breadth of central-eastern Britain. The limited number of Brythonic words which these well-established groups transmitted to the later Anglo-Saxon wave in the 5th century is taken as proof of the essentially minor role which the Celtic tongue played in Roman Lowland Britain. On the strength of the various theories put forward over the last decades on British survival in the wake of the Anglo-Saxon conquest, this discussion maintains that a substantial proportion of the natives may have preserved their acquired language, and hypothesises that they may have relocated en masse to western Britain after their victory at Mons Badonicus around AD 500. In the context of this premise, a number of Latin loanwords into Old English are discussed as potentially transmitted without a Brythonic intermediary; the insight previously gathered on British Latin's true characteristics is also employed to examine whether certain English place-names (most notably those in '-cester', those of the 'Crutch/Crich/Churchill' type, and the so-called 'tautological' ones) may not in fact owe something to a Latin substrate.