A careful analysis of the three texts has satisfied me beyond question that while C is the most accurate in detail, it is marred by a peculiar tendency to omission on the part of its scribe. This, indeed, is its distinctive feature. Now B cannot be derived from C, because it supplies the latter's omissions. On the other hand, C cannot be derived from B, because it corrects, throughout, B's inaccuracies. Consequently they are independent. More difficult to determine is the genesis of A, the worst of the three texts; but as it virtually reproduces all the inaccuracies found in B (besides containing many fresh ones), without correcting any, it can only be inferred that B was its source. Thus we have on the one hand C, and, on the other B (with its offspring A), derived independently from some common source. And this conclusion agrees well with the fact that a long catalogue of lands abstracted from the House of Ely is found in C, but not in A or B,²¹⁸ and with the circumstance that the famous rubric ('Hic subscribitur inquisitio'), which heads the inquisition in A and B, is placed by C at the end of the lists of jurors.²¹⁹

Starting from this conclusion, let us now proceed to ask, what was the document from which B and C copied independently? Clearly, it was not Domesday Book, for outside the eastern counties they record the returns in full, like the Inq. Com. Cant. itself. Were they then taken from the original returns, or at least from the copy of those returns in the Inq. Com. Cant.? This point can only be determined by close analysis of the variants; if we find B and C containing occasionally the same errors and peculiarities, although copied independently, it follows that the document from which they both copied must have contained those same errors and peculiarities. Let us take the case of Papworth. The right reading, as given both in Domesday and the Inq. Com. Cant., I have placed on the left, and the wrong reading, in B and C, on the right:

Here are some further illustrations of errors in the I.E.:

Again, the clause 'Tost²²² pro viii. hidis et xl. acris', which ought to head the Hardwick entries, is wrongly appended in the I.E. (p. 110) to a Kingston entry with which it had nothing to do. So too, 'hoc manerium pro x. hidis se defendit [sic] T.R.E. et modo pro viii. hidis', which belongs to Whaddon, is erroneously thrown back by the I.E. (p. 107), into Trumpington, a Manor in another Hundred. It is singular also that all the MSS. of the I.E. read 'iii. cotarii' (p. 101), where D.B. and the I.C.C. have 'iii. bordarii' (p. 3), and 'x. cotarii' (p. 101), where they have 'x. bordarii' (p. 6): conversely, the former, in one place, read 'xv. bordarii' (p. 107), where the latter have 'xv. cotarii' (p. 63).

In comparing the text of the I.E. with that of the I.C.C., we shall find most striking and instructive variants in the lists of juratores for the several Hundreds. Take, for instance, the lists for the Hundreds of Cheveley and Staines, which follow one another in both MSS.

The second name on these lists can be conclusively tested. For the relative entry in the I.C.C. is 'Esselei tenet euerard[us]²²⁹ filius brientii de Alberico'. This proves that the I.C.C. is right in reading 'Euerard[us]', while the I.E. is right in adding 'homo Alb[er]ici de uer'.

These are the lists for Staines Hundred.

In these two lists the points to strike us are that Harold is placed first on one list and seventh on another; Aleran third on one list and first on another; and 'Fareman' distinguished more clearly in the I.E. than in the I.C.C. as a separate individual.

If we now collect from the other Hundreds some instances of instructive variants, we shall obtain important evidence.

It is impossible to examine the italicized variations in these parallel texts without coming to the conclusion that they must have been independently derived from some common original, an original containing more detail than either of them. On the other hand, the comparatively close agreement between the texts of the actual returns in the I.C.C. and the I.E. leads one to infer that these were copied with far more exactitude than the comparatively unimportant surnames of the jurors. For us the value of these variations in the jurors' lists lies in the evidence afforded to the origin of the existing MSS.

The object of this careful scrutiny has been to prove that as certain errors and peculiarities are found in two independent MSS., they must have existed in the original document from which both were copied, and which was neither the I.C.C. transcripts nor the original Domesday returns. What then was this document? It was, and can only have been, the true Inquisitio Eliensis, the date and origin of which I shall discuss below. Further, I should imagine this document to have probably been a roll or rolls, which—on its contents being subsequently transcribed into a book for convenience—was allowed, precisely as happened to the Domesday rolls themselves, to disappear. In perfect accordance with this view we find the whole contents of the Inquisitio arranged for a special purpose, and no mere transcript of the Domesday returns. Thus, after abstracting all the entries relating to the Cambridgeshire estates, and subjoining a list of houses held in Cambridge itself, it proceeds to add up all the items independently, and record their total values to the Abbey. This analysis is carried out for several counties (pp. 121-4), and is, of course, peculiar to the Inquisitio, although inserted between the abstracts of the Domesday returns for Cambridgeshire and Herts. So too the breviate or short abstract of the estates (pp. 168-173), which was part of the original document—for it is found in all the derived MSS.—must have been specially compiled for it, and so also was the Nomina Villarum (pp. 174-83).

Another peculiarity of the Inquisitio is the care with which it records the names of sokemen on the Abbey estates when omitted in the I.C.C. and D.B. This may lead us to ask whether its compilers supplied these names from their personal knowledge. We might think not, for in some cases they are recorded by the D.B. and the I.C.C., while in one (p. 106) the I.E. actually omits the name, reading only 'quidam sochemanus', where the other two documents (p. 46) supply his name ('Fridebertus'). From this we might infer that the names were probably recorded in the original returns, but deemed of too slight importance to be always copied by the transcriber. Yet the balance of evidence leads me to believe that the I.E. did supply names from independent knowledge. With the values, however, the case is clearer. The I.E. contains special and exclusive information on the value of socman-holdings, and must, I think, have derived it from some other source than the original Domesday returns. Here are some instances in point.

This last passage, of itself, is full of instruction. Firstly, the I.E. alone gives the value of the holding. Secondly, the I.E. preserves the 'sine socha' which qualifies the holder's right. Now D.B. gives the last clause as:

Hi omnes terras suas vendere potuerunt. Soca tantum hominis abbatis de Ely remansit æcclesiæ.

This qualification corresponds with the 'sine socha' of the I.E., and is, we should observe, wholly omitted in the I.C.C. Thirdly, the three versions of the original return employ three different words to express the same one—'recedere', 'vendere', 'dare'. Fourthly, the superiority of the C text of the I.E. over B (which makes two blunders in this passage) and of B over its offspring A (which adds a third) is here well illustrated. Fifthly, the phrase 'Picot vicecomes tenet eam sub abbate ely' differs notably from Domesday, which assigns the estate to Picot unreservedly, and still more from the I.C.C. which reads 'tenet Robertus de Picoto vicecomite in feudo regis'.

The next example is taken from the township immediately preceding.

I have said that in all these cases it might perhaps be held that the additional details found in the I.E. were not due to special information possessed by its compilers, but were derived from the original returns, though omitted by their other transcribers. It is possible, however, to put the matter to the test. If, anticipating for a moment, we find that we have, for the eastern counties, in Domesday the actual materials from which the compilers of the I.E. worked, we can assert that any additional details must have been supplied from their own knowledge. An excellent instance in point is afforded by Tuddenham, in Suffolk:

One knows not, truly, which blunder is the worst, that of the Domesday scribe, who has converted a probable 'S. æ',²³⁷ i.e. Ely Abbey, into 'Saxæ', or that of the compiler of the I.E., who, by interpolating the word 'viginti', has converted three shillings into three-and-twenty. But the point is that the latter could name the Abbot's sokeman (nameless in Domesday) and could supply his acreage and the value of his holding. The actual details seem to have been:

Domesday records the totals only.

Enough has now been said of the twelfth century transcripts in which alone are preserved to us the contents of the Inquisitio. We have seen that they point to the existence of some common original, which, while closely parallel with Domesday, as a record of the Abbey's possessions, contained certain special features and additional information. Why, when, and from what sources that original was compiled, I shall now endeavour to explain.

XVIII. THE ELY RETURN

The theory I propound for the origin of the so-called Inquisitio Eliensis is that it was the actual return ordered by that writ of the Conqueror,²³⁸ of which a copy is given in all three MSS. (A, B, C) and which is printed in Mr Hamilton's book, on p. xxi (No. VIII). I give the wording of the writ, followed by the heading to the Inquisitio with which it should be closely compared.

Willelmus Rex Anglorum Lanfranco archiepiscopo salutem.... Inquire per episcopum Constantiensem et per episcopum Walchelinum et per ceteros qui terras sanctæ Ædeldrede scribi et jurari fecerunt, quomodo jurate fuerunt et qui eas juraverunt, et qui jurationem audierunt, et qui sunt terre, et quante, et quot, et quomodo vocate [et] qui eas tenent. His distincte notatis et scriptis fac ut cite inde rei veritatem per tuum breve sciam. Et cum eo veniat legatus abbatis.

Return

Hic subscribitur inquisicio terrarum, quomodo barones regis inquisierunt,²³⁹ videlicet per sacramentum vicecomitis scire et omnium baronum et eorum francigenarum, et tocius centuriatus, presbiteri, prepositi, vi. villani [sic] uniuscujusque ville; deinde quomodo vocatur mansio, quis tenuit eam tempore R.E., quis modo tenet, quot hide, quot carruce²⁴⁰ in dominio, quot hominum, quot villani, quot cotarii, quot servi, quot liberi homines, quot sochemanni, quantum silve, quantum prati, quot²⁴¹ pascuorum, quot molendina, quot piscine, quantum est additum vel ablatum, quantum valebat totum simul,²⁴² et quantum modo, quantum quisque liber homo vel sochemannus habuit vel habet. Hoc totum tripliciter, scilicet tempore regis Æduardi, et quando Rex Willelmus dedit et qualiter modo sit, et si potest plus haberi quam habeatur.

Isti homines juraverunt, etc., etc.

Especially important is the fact that the return contains the jurors' names, in accordance with the express injunction to that effect in the Conqueror's writ.

Now if this theory meet with acceptance, and the writ be taken to refer, as I suggest, to the Domesday Inquest itself, it follows that the Bishop of Coutances and Bishop Walchelin were the heads of the Domesday Commission for this district. This, of course, has been hitherto unknown; but it adds to the presumption in favour of the facts that Bishop Walchelin is not mentioned in any of the Ely writs as taking part in the placita concerning the Abbey's lands, and that, therefore, the only Inquest in which he could have been concerned was the Domesday Inquest itself. It should be added, however, that these two Bishops may have been, respectively, the heads of two distinct commissions for adjoining groups of counties.

The heading to the Inquisitio Eliensis is so well known, and has been so often quoted by historians, that it is a gain to fix its status, the more so as it has been loosely described as the 'official' instructions for the Survey itself. We may also determine the date of the writ as the very close of the Conqueror's reign. For it must have been issued between William's departure from England, circ. September 1086, and his death (September 1087).

And now, how was the return compiled? It deals, we find, with six counties, arranged in this order: Cambridgeshire, Herts, Essex, Norfolk, Suffolk, and Hunts. For Cambridgeshire it copies, clearly, from the original returns. For Herts it must have done so also, because it gives full details, which are not found in Domesday Book. This conclusion is confirmed by the fact that, for these two counties, it gives the jurors' names (for the hundreds dealt with), which it could only have obtained from those original returns. For Essex, Norfolk, and Suffolk, on the contrary, it simply gives the same version as the second volume of Domesday Book, and omits accordingly the jurors' names. The case of the four Manors in Hunts I leave in doubt, because the version in the Inquisitio (pp. 166-7) has more details than that of Domesday, though the latter is here exceptionally full, and because it places first the Manor which comes fourth in Domesday (i. 204). The additional details (as to live-stock) are such as we might expect to be derived from the additional returns; but the names of the witnesses for the Hundred are not recorded, a fact to be taken in conjunction with the belated entry of these Huntingdonshire Manors not following, as they should, those in Cambridgeshire and Herts.

In addition to the Inquisitio itself, as printed by the Record Commission, there is a record, or collection of records, which follows it in all three MSS., and which is printed in Mr Hamilton's book (pp. 168-89). Although its character is not there described, it can be determined. For in the Inquisitio there are three references to the 'breve abbatis de ely' (pp. 123-4), all three of which can be identified in the above record (pp. 175-7). It is noteworthy that the record in question is only complete in C, which confirms my view that B and its offspring A were independent of C.

Though the word Breve in Domesday Book normally means the king's writ, there are passages which seem to have been overlooked, and in which it bears another and very suggestive meaning. One of them is found at the end of the Survey of Worcestershire and was foolishly supposed by the compilers of the index volume (pp. 250, 315) to relate to lands held by 'Eddeva' and entered immediately before it. The passage is an independent note, running thus:

In Esch Hund' jacent x. hidæ in Fecheham et iii. hidæ in Holewei et scriptæ sunt in brevi de Hereford.

In Dodintret Hund' jacent xiii. hidæ de Mertelai et v. hidæ de Suchelei quæ hic placitant et geldant, et ad Hereford reddunt firmam suam, et sunt scriptæ in breve regis (i. 178).

All four places are found on fo. 180b, 'Feccheham' and 'Haloede' [sic]²⁴³ together (under 'Naisse' Hundred²⁴⁴) as paying a joint ferm—'Merlie' (Martley) under 'Dodintret' Hundred and Suchelie (Suckley), now in Herefordshire, as 'in Wirecestrescire' (cf. i. 172).

It is clear then that Domesday here uses 'breve' of a return, not of a writ, and I venture to think the word may refer to the abbreviated entries made in Domesday Book itself as distinct from those in extenso found in the original returns.²⁴⁵

This usage is found in both volumes. We read of land at Marham, Norfolk, held by Hugh de Montfort; 'est mensurata in brevi Sanctæ Adeldret' (ii. 238), where the reference is to the 'Terra Sanctæ Adeldredræ' (ii. 212), and of Hurstington Hundred, Hunts, 'Villani et sochemanni geldant secundum hidas in brevi scriptas' (i. 203). The reference, in both cases, is to the text itself.

The former of these two phrases is repeated in the Inquisitio Eliensis,²⁴⁶ a fact of some importance if, as I venture to think, it is there meaningless. The point is worth labouring. We see that the phrase cannot have occurred in the original returns, where all the entries relating to Marham would have come together. But if it was only applicable to Domesday Book itself—where the fiefs were separated—then must the I.E. have copied from Domesday Book.

This, indeed, is the point to which I am working. For Essex, Norfolk, and Suffolk, I believe, the compilers of the Inquisitio (1086-7) must have worked from the second volume of Domesday as we have it now. We see it firstly, in the order of the counties; secondly, in the absence of the jurors' names; thirdly, in the system of entering the lands. With a fourth and minute test I have dealt just above.

But to make this clearer, we must briefly analyse the return. The Cambridgeshire portion extends from p. 101 to p. 120. It extracts from the original returns, Hundred by Hundred, all that relates to the Abbey of Ely. Following this is a note of its possessions in the Borough of Cambridge²⁴⁷ (pp. 120-1), and then summaries of the Abbey's estates, in dominium and thainland and socha, in all six counties, and of the lands held by Picot the Sheriff, Hardwin d'Eschalers and Guy de Raimbercurt, to which it laid claim as its own (pp. 121-4). Then we resume with Hertfordshire, the extracts from the original returns (pp. 124, 125). Both the Cambridgeshire and Hertfordshire portions close with the words, 'De toto quod habemus', etc., referring to the totals worked out by the Abbey from the entries in the original returns.

With Essex, we enter at once on a different system. This portion, which extends from p. 125 to p. 130 (line 8), is arranged not by Hundreds but by fiefs. It first gives the lands actually held by the Abbey (as coming first in Domesday), and then those of which laymen were in possession. To the latter section are prefixed the words: 'Has terras calumpniatur abbas de ely secundum breve regis'. From Essex we pass to Norfolk, the entries for which, commencing on p. 130 with the words 'In Teodforda', end on p. 141 at 'Rogerus filius Rainardi'. These again are divided into two portions, namely, the lands credited to the Abbey in Domesday (pp. 130-6), and those which it claimed but which Domesday enters under other owners (pp. 137-41). Between the two comes the total value of the former portion and a list of the Norfolk churches held by the Abbey. Last of the Eastern counties is Suffolk, which begins on p. 141 at 'In Tedeuuartstreu hund.', and ends on p. 166. This also is in two portions, but the order seems to be reversed, the alleged aggressions on the Abbey's lands coming first and its uncontested possessions last. The latter portion begins on p. 153, where the B text inserts the word 'Sudfulc'.

The following parallel passages are of interest as showing how closely the I.E. followed D.B. even when recording a judicial decision.

The one variation, the Bishop's style, has a curious parallel in Domesday Book (i. 165), where under the rubric 'Terra Episcopi Constantiensis' we read 'Episcopus de Sancto Laudo tenet', etc.

We may take it then that the compilers of the Inquisitio Eliensis worked for Cambridge and Herts from the original returns, but, for the eastern counties, from the second volume of Domesday. What are the corollaries of this conclusion? They used, for some reason or other, the second volume of Domesday, but not the first—if, indeed, it then existed. Speaking for myself, I have always felt not a little uneasy as to the accepted date for the completion of Domesday Book.²⁴⁸ Mr Eyton went so far as to write:

Imperial orders have gone forth that the coming Codex, the Domesday that is to outlive centuries, is to be completed before Easter (April 5th, in that year [1086]), when King William himself expects to receive it in his Court and Palace of Winchester (Notes on Domesday, 15).

And he explicitly stated that:

On any hypothesis as to the time taken by the different processes which resulted in Domesday Book, the whole, that is the survey, the transcription, and the codification, were completed in less than eight months, and three of the eight were winter months. No such miracle of clerkly and executive capacity has been worked in England since.²⁴⁹

But was it worked then? All that the chronicle says of the King is that the 'gewrita wæran gebroht to him', a phrase which does not imply more than the original returns themselves.

Of course, the chief authority quoted is the colophon to the second volume:

Anno millesimo octogesimo sexto ab incarnatione Domini vicesimo vero regni Willelmi facta est ipsa descriptio non solum per hos tres comitatus sed etiam per alios.

It seems to have been somewhat hastily concluded that because the Survey ('Descriptio Angliæ') took place in 1086, Domesday Book (which styles itself Liber de Wintonia), was completed in that year. The phrase 'per hos tres comitatus' proves, surely, that 'descriptio' refers to the Survey, not to the book.²⁵⁰

I have never seen any attempt at a real explanation of the great difference both in scope and in excellence between the two volumes, or indeed any reason given why the Eastern counties should have had a volume to themselves. For a full appreciation of the contrast presented by the two volumes, the originals ought to be examined. Such differences as that the leaves of one are half as large again as those of other, and that the former is drawn up in double, but the latter in single column, dwarf the comparatively minor contrasts of material and of handwriting. So, too, the fullness of the details in the second volume may obscure the fact of its workmanship being greatly inferior to that of the first. Of its blunders I need only give one startling instance. The opening words of the Suffolk Survey, written in bold lettering, are 'Terra Regis de Regione' (281b). I have no hesitation in saying that the last words should be 'de Regno'. Indeed, the second formula is found on 289b, as 'Terra Regis de Regno', while on 119b under 'Terra Regis', we read 'hoc manerium fuit de regno'. So also in the Exon Domesday 'Terra Regis' figures as 'Dominicatus regis ad regnum pertinens'.²⁵¹ The muddled order of the tenants-in-chief for Norfolk and for Suffolk—where laymen precede the church²⁵² —is another proof of inferiority, but only minute investigation could show the hurry or ignorance of the scribes.

Now, all this might, I think, be explained if we took the so-called second volume to be really a first attempt at the codification of the returns. Its unsatisfactory character must have demonstrated the need for a better system, which, indeed, its unwieldy proportions must have rendered imperative. So drastic and so successful, on this hypothesis, was the reform, that while these three counties had needed a volume of 450 folios, the rest of England that was surveyed—some thirty counties—was compressed into a single volume of 382 folios, and on a system which rendered consultation easier and more rapid. In every respect the first volume is a wonderful improvement on the second, but the authorities may have shrunk from ordering the latter to have been compiled de novo, when the work, though unsatisfactory, had once been done.

This, it must of course be remembered, is all hypothesis, a hypothesis suggested by the facts. If it were proved that at the time when the Ely return was made, the 'second' volume had been compiled, and the 'first' had not, I should have established my case. But it might be urged that the 'first' volume did exist at the time, and that the Ely scribes used the returns instead, because they contained fuller information. To this I reply, so far as the details of the estates are concerned, that neither the terms of the writ nor the heading of the Inquisitio involved the inclusion of such details as Domesday Book omitted. If the scribes inserted them, it must have been merely because they inserted everything they found in the records from which they copied. It might still be urged that they went to the returns for the names of the juratores; but why, if so, did they not do so for the three eastern counties? It certainly seems to me to be the most satisfactory explanation that the materials supplied for compiling this return, as being the recognized official records, were the so-called 'second' volume of Domesday, and (for the rest) the original returns.

XIX. FIRST MENTION OF DOMESDAY BOOK

No one nowadays should require to be told that the pseudo-Ingulf's dealings with Domesday are devoid of all authority. Some, however, may still believe in the tale found in that 'Continuatio' of his chronicle which is fathered on Peter of Blois. It is there that Ellis found (putting Ingulf aside) the only case of an appeal to its witness before the reign of John.²⁵³

With the 'Continuatio' I shall deal below,²⁵⁴ but I would observe, while on the subject, that the 'pseudo-Ingulf' (charters and all) was, I believe, largely concocted by the help of hints gathered from Domesday Book.

The absence of any authoritative mention, in its early days, of our great record gives a special importance to an entry in the Chronicle of Abingdon (ii. 115-6), where we read that Abbot Faritius was impleaded by certain men:

Sed is abbas in castello Wincestre coram episcopis Rogero Saresberiensi, et Roberto Lincolniensi, et Ricardo Londoniensi, et multis regis baronibus, ratiocinando ostendit declamationem eorum injustam esse. Quare, justiciarorum regis judicio obtinuit ut illud manerium, etc. ... sed quia rex tunc in Normanniâ erat, regina, quæ tunc præsens erat, taliter hoc sigillo suo confirmavit.

Then follows the Queen's writ, announcing the decision of the plea held in the royal 'Curia', together with the names of the 'barons' present. These names enable us to determine a certain limit for the date of the plea. 'Thurstinus Capellamus', for instance, implies that it was previous to his obtaining the See of York in 1114, while the presence of Richard, Bishop of London, places it subsequent to July 26, 1108. It must, therefore, have been held during the King's absence between July 1108 and the end of May 1109; or in his later absence from August 1111 to the summer of 1113.

The action of the Queen in presiding over this placitum illustrates a recognized practice, of which we have an instance in Domesday itself (i. 238b), where it is stated that Bishop Wulfstan, 'terram deplacitasse coram regina Mathilde in presentia iiii^or. vicecomitatuum'. The Queen's description of the Curia Regis as 'curia domini mei et mea' should be compared with the phrase employed by the Queen of Henry II, who, similarly acting in her husband's absence, speaks of the Great Justiciar as 'Justicia Regis et mea'.

But the essential portion of the passage before us is this:

Sciatis quod Faritius abbas de Abendona in curia domini mei et mea, apud Wintoniam in thesauro ... per Librum de Thesauro, diratiocinavit quod, etc.

The court was held 'in castello Wincestre', says the narrative, 'apud Wintoniam in thesauro', says the record. Both are right, for the Royal Treasury was in Winchester Castle.²⁵⁵

But what was the 'Liber de Thesauro'? I contend that it was Domesday Book, and can have been nothing else. For, passing now to the Dialogus de Scaccario (circa 1177), we there read in reply to an inquiry as to the nature of Domesday Book (which 'in thesauro servatur et inde non recedit'): 'liber ille de quo quæris sigilli regii comes est individuus in thesauro' (I. XV.). The connection of the Book with the Treasury is brought out strongly in the Dialogus, and leads to the presumption, as Mr Hall perceived, that the Treasury being originally at Winchester, the Book was there also—as indeed we see it was under Henry I.²⁵⁶ On the date of its removal to Westminster, there has been much discussion between my friend Mr Hall and myself.²⁵⁷ Mr Hall relies mainly on the Dialogus de Scaccario, and on the inferences he draws from it, for the early removal of Domesday to Westminster, and the establishment there of the royal Treasury. For myself, I claim for the Winchester Treasury greater importance and continuity than he is willing to admit. The leading records, of course, were stored there as well as treasure. We find William Rufus speaking of 'meis brevibus ... qui sunt in thesauro mea Wyntoniæ';²⁵⁸ and we read that, on his father's death, 'pergens apud Wincestre thesaurum patris sui ... divisit: erant autem in thesauro illo lx. m[ille] libræ argenti excepto auro et gemmis et vasis et palliis.'²⁵⁹ Heming's Cartulary describes the Domesday returns as stored 'in thesauro regali', and Henry of Huntingdon states that 'inter thesauros reposita usque hodie servantur'.²⁶⁰ Now, as the Treasury was in Winchester Castle at the time of the above suit, and as it had been in 1100²⁶¹ and 1087, so it was still at the accession of Stephen in 1135, and at the triumph of Matilda in 1141. This is absolutely certain from the Chronicles, nor do they ever mention any other Treasury. Moreover, the contents of this Treasury in 1135—'erant et vasa tam aurea quam argentea'—correspond with those described by the Dialogus forty years later: 'vasa diversi generis aurea et argentea'. Lastly, there is a piece of evidence which has not yet been adduced, namely, that in his Expugnatio Hibernica (1188), Giraldus, speaking of that ring and letters which John of Salisbury declared had been brought by him from the Pope, and were 'still stored in the Royal Treasury', writes of

Annulum aureum in investituræ signum ... qui statim simul cum privilegio in archivis Wintoniae repositus fuerat.

Giraldus certainly must have looked on the Royal Treasury at Winchester as the only recognized repository for all such objects as these.

Mr Hall, indeed, has gradually modified his original position that 'Ingulphus saw the Domesday register, as it now exists, at Westminster', and that it was sent there for good from Winchester 'early in the reign of Henry I',²⁶² but he still places the establishment of 'the' Treasury at Westminster, in my opinion, too early. It is the gradual decay of Winchester as the capital and seat of administration that makes it difficult to say positively when or how the national records, Domesday Books among them, were transferred to Westminster. We have seen at least that, in its early days, the 'Liber de Wintonia', as it styles itself, had its home within the walls of the Royal castle of Winchester; and I cannot but think, now as at first, that it began by visiting Westminster for Exchequer sessions only.²⁶³

In any case, we have seen its witness appealed to on a far earlier occasion than had hitherto been known. In my paper on 'An Early Reference to Domesday',²⁶⁴ I quoted an even earlier mention of the 'Descriptio Angliæ', but here again the reference seems to make rather to the Domesday Survey itself than to Domesday Book, the 'Liber de Thesauro'.

As an appendix to this paper, I give the pedigree of the Domesday MSS. according to the views I have expressed.²⁶⁵