(I) By a _fixed_ payment at the rate of one marc on the fee (old assessment). (II) By an _arbitrary_ payment of certain mysterious sums, which prove to be multiples of the unit 15s 3d. But there is no fixed proportion to be traced between the amount paid and the number of _servitia_ due. Numerous instances are found of a single knight's fee being charged with a sum equivalent to five of these mysterious units. Magnates, again, are found paying apparently strange sums, which prove on dissection to represent 50, 100, 200 and even 300 of these units. The clue to the mystery is found in an entry on the Pipe-Roll of the following year (12 Henry II), which proves that this unit was the pecuniary equivalent of a _serviens_, and that the various payers had 'promised' the king so many _servientes_ for the war in Wales.[161] Such 'promises' were evidently offers, made independently of the actual service due from the 'promising' party. Following up this clue, we see that the Abbot of Abingdon must, like the Bishop of Hereford, have promised 100 'serjeants',[162] that the Abbot of St Alban's must have done likewise,[163] while the Bishop of London must have promised 150, _in addition_, be it noted, to paying a scutage of a marc on each knight's fee (20) of his _servitium debitum_.[164] For the rolls of 1162 and 1163 prove that he had duly paid the scutage of the former year, and that this was a further payment. The varying form of these entries should be observed, for it was evidently quite immaterial to the clerks whether they wrote '5 serjeants' or their equivalent--76 shillings and 3 pence.[165] Taking the pay of the _serviens_ at 1d a day, the unit in question would represent six months' pay (for a year of 366 days).