The third point that I would raise is this. Even assuming that the amount of 'service' bore a fixed proportion--whether in pecuniary or territorial units--to the extent of possession, we are, surely, at once confronted by the difficulty that the owner of _x_ units of possession would be compelled, for the discharge of his military obligations, to enfeoff _x_ knights, assigning a 'unit' to each. A tenant-in-chief, to take a concrete instance, whose fief was worth £100 a year, would have to provide _ex hypothesi_ five knights; if, as was quite usual, he enfeoffed the full number, he would have to assign to each knight twenty librates of land (which I may at once, though anticipating, admit was the normal value of a knight's fee), that is to say, the crown would have forestalled Henry George, and the luckless _baro_ would see the entire value of his estate swallowed up in the discharge of its obligations.[38] What his position would be in cases where, as often, he enfeoffed more knights than he required, arithmetic is unable to determine. I cannot understand how this obvious difficulty has been so strangely overlooked. The fourth and last criticism which I propose to offer on the subject is this. If we find that under Henry II--when we meet with definite information--a fief contained, as we might expect, more 'units of possession' than it was bound to furnish knights (thus leaving a balance over for the _baro_ after sub-infeudation), we must draw one of two conclusions: either this excess had existed from the first; or, if the fief (as we are asked to believe) was originally assessed up to the hilt for military service, that assessment must, in the interval, have been reduced. In other words, Henry I--if, as Dr Stubbs in one place suggests,[39] he was the first to take a 'regular account of the knights' fees'--must have found the land with a settled liability of providing one knight for every five hides, and must, yet, have reduced that liability of his own accord, on the most sweeping scale, thus, contrary to all his principles, ultroneously deprived himself of the 'service' he was entitled to claim. Having completed my criticisms of the accepted view, and set forth its chief difficulties, I shall now propound the theory to which my own researches have led me, following the same method of proof as that adopted by Mr Seebohm in his _English Village Community_, namely working back from the known to the relatively unknown, till the light thrown upwards by the records of the twelfth century illumines the language of Domesday and renders the allusions of monks and chroniclers pregnant with meaning. 1. THE 'CARTAE' OF 1166