Robert 'filius Regis' 100 (?) William de Traci 30 (?) William de Braose 25 (?) Oliver de Traci 25 (?) Abbot of Tavistock 15 15 William fitz Reginald 1 1 Ralf de Valtort 1 1 Robert fitz Geoffrey 1 1 Here we are supplied by this roll with four important _servitia_ which would otherwise be absolutely unknown to us. And they happen to be of special interest. For while the _carta_ of William de Braose returns twenty-eight fees, and that of Oliver de Traci twenty-three and a half (though he pays on thirty and a half),[119] their payments in 1165, by revealing their _servitium debitum_, show us that their fiefs represent the two halves of the Honour of Barnstaple (which, therefore, was assessed at 50 knights) then in their respective hands. Again, William de Traci returns his fees in his _carta_ as twenty-five and three-quarters, and says nothing about any balance on his _dominium_, as he should have done. Hence we should not have known his _servitium_ but for the roll of 1165. Swereford's extraordinary failure to understand this roll aright is possibly due to the fact that most of the relevant payments are entered without mention of their object. He seems to have been very dependent upon the rolls explaining themselves, and to have worked in the spirit of a copying clerk rather than of an intelligent student. One more example of his errors will suffice. In his abstracts from the aid 'ad maritandam primogenitam filiam regis' (1168), we read: Abbas Gloucestriæ de promissione, sed non numeratur quid; sed in rotulo praecedenti dicitur:--Abbas Gloucestriæ debet xxxviij. l. ij. s. vj. d. de veteri scutagio Walliae. Now (1) the amount of the abbot's contribution is duly entered on the roll ('xl. marcas de promissione de eodem auxilio'), and it is not paid in respect of fees, but is a voluntary proffer; (2) the phrase in the preceding roll is not 'de veteri _scutagio_', but 'de veteri _exercitu'_; (3) the payment there recorded represents a contribution of fifty _servientes_, and had nothing to do with scutage, for the abbot (as Swereford should have known) did not hold by military service, and ought not, therefore, to figure in his lists at all.[120] Let us turn, therefore, to the rolls themselves. Now, although the language of the exchequer was not so precise as we could wish, it is possible, more or less, to distinguish and classify these levies. Thus, we have of course a typical 'aid' in the levy for the marriage of the king's daughter (1168), while, on the other hand, we have an equally typical 'scutage' in 1156, in the payments made by the church tenants in lieu of military service.