It would seem impossible to resist the conclusion that in these widely differing rates we have traces of a polity as yet divided, of those independent kingdoms from which had been formed the realm. Kent, for instance, which had so steadily maintained, first, its independent existence, and then its local institutions, had succeeded in preserving an assessment that its neighbours had cause to envy. In the west, Cornwall similarly enjoyed a low, indeed a nominal assessment while that of Devon, though higher than this, was so significantly lower than those of Somerset and Dorset[175] as to remind us that here, in part at least, the 'Welsh' long held their own. If the incidence of geld were shown by shading a map of England, on the plan so successfully adopted in Mr Seebohm's great work, it would show that the heavily assessed counties were those which formed the nucleus of the old West-Saxon realm.[176] All round this nucleus the map would shade off sharply, another sudden change marking the Danish counties on the north, the Jutish kingdom on the east, and the British district in the south-west. It is, perhaps, worthy of remark that Shropshire was assessed twice as heavily as the adjoining county of Stafford, possibly because part of it was added, at a very early period, to the kingdom of the West Saxons. If Mr Eyton was right in his reckoning that Kesteven was assessed twice as heavily as Lindsey, and Lindsey, in turn, twice as heavily as Holland, it would illustrate the survival of local distinctions even within the compass of a modern county, as well as the 'shading off' tendency of which I have already spoken. The point I have here endeavoured to bring out is that if the system of artificial assessment were of Roman or British origin, we should expect to find it fairly uniform over the whole country, whereas we find, on the contrary, the very widest discrepancies. It might be urged, perhaps, that these were due to the differing conditions of particular counties, to their more or less partial reclamation, for instance, of the date when they were assessed. But this would not account for the grouping I have traced, and would imply that each county ought to differ indefinitely. Nor would it explain the case of Kent, where a county that must have been foremost in early development and prosperity enjoyed a phenomenally low assessment.