It is manifest, however, that a table of this kind can have no permanent value, beyond giving a full and accurate statement of discoveries up to date, as further researches may not only change its numerical data, but give a totally different aspect to inferences based on the existence or absence of these remains in certain districts. Thus it is only within the last few years that Ayrshire could be included in the lake-dwelling area, so that, previously, the conflicting statement made by Chalmers,[40] that Galloway was colonised by the Irish about the eighth century, derived some countenance from the archæological discoveries in Loch Dowalton, and other lakes in the neighbourhood, when taken in conjunction with the prevalence of analogous remains in Ireland. Though we cannot, therefore, argue definitely from the present geographical distribution of Scottish Lake-Dwellings, the indications are so clearly suggestive of their having been peculiar to those districts formerly occupied by Celtic races that the significance of this generalisation cannot be overlooked. Thus, adopting Skene's division of the four kingdoms into which Scotland was ultimately divided by the contending nationalities of Picts, Scots, Angles, and Strathclyde Britons, after the final withdrawal of the Romans, we see that of all the crannogs proper, none have been found within the territories of the Angles;[41] ten and six are respectively within the confines of the Picts and Scots; while no less than twenty-eight are situated in the Scottish portion of the ancient kingdom of Strathclyde. Nor is this generalisation much affected by an extension of the list, so as to include those stony islets so frequently met with in the Highland lakes. On the other hand, that they have not been found in the south-eastern parts of Scotland may suggest the theory that these districts had been occupied by the Angles before Celtic civilisation--or rather the warlike necessities of the times--gave birth to the island dwellings. In that case we would suppose that their development dates back to the unsettled events which immediately followed the withdrawal of the Roman soldiers, to whose protection the Romano-British population in the south-west of Scotland had been so long accustomed. But this leads me to notice some of the historical phenomena associated with the localities thus referred to. SECTION II. _Historical and Traditional Phenomena associated with their Area of Distribution._ (Compiled chiefly from Dr. Skene's works.) In the year A.D. 79, Julius Agricola, with his legions, entered that portion of Britain afterwards known as the kingdom of Scotland by way of the Solway Firth, and, quickly subjugating the tribes occupying its northern shore, garrisoned the country as he advanced. The work of the following winter is thus described by Tacitus:--