Introduction to the Somerset Domesday

DOMESDAY SURVEY

The 'Exon Domesday,' p. 383. Assessment by hidation, p. 386. Areal measures, p. 388. The King's land, p. 393. Its firma, p. 395. The comital manors, p. 397. Lands of Queen Edith and of Wulfward White, p. 398. Taunton, p. 400. Other church lands, p. 405. Montacute, p. 409. Lands of the barons, p. 412. Serjeants and thegns, p. 414. The sheriffs, p. 419. The 'third penny,' p. 420. The towns, p. 420. The live stock, p. 423. Moor and meadow, p. 425. The peasantry, p. 425. Norman encroachments, p. 426. Customary dues, p. 428. Legal antiquities, p. 429. Miscellanea, p. 430.

If for our knowledge of the great Survey of 1086 we were dependent in Somerset on Domesday Book alone, we should find in it few points of general or special interest. Manor follows manor in regular, monotonous succession; the assessment, the number of ploughs required for working the land, the number actually possessed by the lord and by the peasants respectively, the number of these peasants arranged according to their classes, the water-mills, the amount, roughly reckoned, of meadow, pasture and woodland, and the value, past and present, of the manor, all these are duly recorded, together with the name of the Englishman who had held it before the Conquest and that of the baron or under-tenant to whom it had passed in what the record terms 'King William's time.' But we look in vain for those glimpses of history, those little incidental touches which bring us, in some counties, face to face, as it were, with real men and women, and lift for a moment the veil that still hides from our eyes the meaning of the Norman Conquest.

Happily, however, we possess for the south-west of England, including the county of Somerset, two auxiliary sources of information which supplement, on certain points, the evidence of Domesday Book. The first of these is known as the' Exon Domesday,' the other as the 'Geld Inquest' or Inquisitio Geldi. The nature of these records requires to be briefly explained, for a good deal of confusion on the subject appears to exist.

As was explained by Sir Henry Ellis, under whose superintendence they were printed by the Record Commission, (fn. 1) the MS. sheets containing both of them were 'bound up in two volumes about the end of the fourteenth or beginning of the fifteenth century,' but so carelessly that they had to be separated again and 'arranged as they are now printed, in the most obvious order,' for the Record Commission's edition. They were bound up in one volume, which is still in the custody of the dean and chapter of Exeter, whence is derived the name 'Exon Domesday.' Ellis distinguished quite rightly between (1) 'its main body,' which 'is supposed, so far as it extends, to contain an exact transcript of the original rolls and returns made by the Conqueror's Commissioners at the time of forming the General Survey from which the Great Domesday itself was compiled,' and (2) the 'Inquisitio Geldi or Taxation of the Hundreds,' which he discussed at some length. He also rightly connected the latter with the great levy of 'geld' (land tax) at the rate of six shillings on the hide 'at the end of the year 1083 or the beginning of 1084.' (fn. 2) He seems to have erred only in writing—

Certain it is that the Record itself bears evidence that the tax was raised at the time of the Survey; that it was connected with it; and that, at least in the Western Counties, it was collected by the same Commissioners.

But, in Ellis' time, the actual date of the great Domesday Survey seems to have been deemed uncertain; he himself knew from the colophon at the end of the eastern counties volume that 'it is evident that it was finished in 1086,' (fn. 3) and from the contents of both volumes that some parts of them could not be previous to '1085 or even later.'Yet the contradictions of chroniclers seem to have influenced his mind.

Mr. Eyton, who made such extensive use, in his monographs on Somerset and Dorset in Domesday, of the Exeter volume, wrote of that 'priceless document,' the geld-inquest, as follows—

The true and authentic title of this Record is 'Inquisicio Gheldi.' Because the place of its custody has happened for centuries to have been the same with that of the Exon Domesday, and because the older Record Commission caused it to be printed and bound up with the Exon Domesday, this Inquest is usually spoken of and quoted as part of the Exon Domesday. Nothing can be more erroneous, nothing more suggestive of further error. The Inquest is two years older than any Domesday.

A minute examination of this Inquest suggests that it was used, though not implicitly followed, by the Domesday Commission which afterwards visited the five Counties in question. Its date and nature, therefore, demand our immediate attention. . . .

The date of the Inquisicio Gheldi was therefore the first three months of 1084. (fn. 4)

From what has been said above, it will doubtless be seen that this is scarcely fair to Ellis, who had so carefully distinguished the Inquisitio from the rest of the MS., or indeed to the Record Commission. For the Inquisitio was not merely preserved in the same custody; it was actually part of the same MS., and a close comparison of the two portions, distinct though they are, suggests that their collocation had a definite meaning and purpose. (fn. 5)

On the other hand, if it was wrong to charge the Record Commission with having combined for the first time two distinct records, it is perfectly true that their publication under the single title 'Exon Domesday' tends to cause confusion. In the MS., as at present arranged, they are so 'sandwiched' together that they could hardly be published separately (fn. 6); but it would be better to speak of the whole as 'the Exeter Book,' (fn. 7) and to refer to its component portions, as will be done in this work, as the 'Exon Domesday' and 'Inquisicio Geldi' respectively.

Of this 'Exeter Book 'the importance can hardly be overrated so far as the Domesday Survey of the south-west is concerned. In Domesday Book, for instance, the usual hundredal headings are omitted in the case of Somerset, and we are therefore dependent for our knowledge of its Hundreds at the time, and for the identification of several doubtful manors, on the Inquisicio Geldi. Mr. Eyton, mentioning this, observes that 'the way in which the two Records, the Gheld-Inquest and Domesday, explain and supplement one another may be almost said to double the antiquarian value of the great Record.' (fn. 8) Of the other portion it is sufficient to say that a question arose, when deciding on the treatment of the Survey in Somerset, whether the text of the 'Exon Domesday should not be adopted in preference to that of Domesday Book, as approaching more closely to that of the original returns. It was only the desire to render the series of Victoria Histories as uniform as possible that led to the adoption here also of the great Exchequer volume as the standard text. As to the MS. itself, it was brought up to London for exhibition at the time of the Domesday Commemoration (1886), and Dr. De Gray Birch then described it as 'closely resembling Domesday Book in its general form and palæography,' and as 'of the eleventh century.' (fn. 9)

Enough, perhaps, has now been said of the tools with which we have to work in studying the Somerset Survey. It must, however, be borne in mind that the record of that Survey was made Hundred by Hundred, and that the original rolls in which that Survey was contained were lost at an early date. Even in the Exon Domesday their contents are entirely rearranged, while in Domesday Book they are rearranged on yet another system and also reduced by omitting certain details throughout.

As the primary object of the Domesday Survey was the ascertainment of the county's right assessment for the (Dane) geld, that is, the original land-tax, the principle on which it was assessed is the first point to be considered. Down to a few years ago Mr. Eyton's views held the field. Set forth in his Key to Domesday, (fn. 10) and repeated in his studies on the Somerset Survey, (fn. 11) his conclusions were as follows. He rightly rejected the view, formerly prevalent, that the hide was an areal measure, and recognized that it was a unit of assessment unconnected with area. In his opinion 'hidage was intended to be an index of one or more of three things, viz., of liability in the first place; of capacity or intrinsic value in the second; of adventitious or extrinsic value in the third.' (fn. 12) The hidation of Dorset was regulated, he held, by' royal favour, intrinsic wealth and extrinsic advantages.' (fn. 13) The only special feature distinguishing the Somerset hidation from that of Dorset was that 'the chief, we may almost say the only, intrinsic faculty which will have regulated the value of Somerset manors was fertility of soil; that this faculty was the very first to be considered when the hidation of any Somerset manor was originally prescribed is evident on every page of Domesday.' (fn. 14)

In my Feudal England (1895) I advanced a new theory on the subject based on the extraordinary number of vills assessed at five hides or some multiple of five hides. From this fact I argued that assessment was not regulated by area or value, but was of an arbitrary character, being so arranged as to apportion among the vills of a Hundred its total assessment in blocks of five or ten hides. This arrangement is much clearer in some counties than others, and Somerset is not a county in which it is obvious to the eye. I was able to show, however, that even here sufficient traces are found to show that this system prevailed. (fn. 15) This theory, which is altogether opposed to that of Mr. Eyton, was emphatically adopted by Professor Maitland in his Domesday Book and Beyond (1897), and has since been elaborately applied to Somerset by the Rev. E. H. Bates. (fn. 16) Mr. Bates contends that the local assessments can, in practice, be explained by my theory, if the county be 'arranged in twelve districts, containing one or more Hundreds apiece,' and each district 'sub-divided into blocks containing assessments of twenty hides, with a few double and triple instances.' Whether his conclusions be accepted or not, the number of Somerset vills assessed in terms, as a mathematician would say, of the five-hide unit is far too large to be accounted for by any theory but that of an artificial arrangement based upon that unit. We find, on examining Mr. Bates' tables, that, putting aside his own combinations, Somerset had no fewer than sixty vills assessed at 5 hides each, thirty-eight at 10 hides, ten at 15 hides, fourteen at 20 hides, three at 30 hides and two at 40 hides and 50 hides respectively. These, it will be seen, account between them for 1,290 hides, that is, for more than half the total assessment of the county. (fn. 17) But when to these indisputable cases we add, from Mr. Bates' tables, those in which the assessments of two or more adjoining vills have clearly been combined in a multiple of the fivehide unit—as was certainly done in Cambridgeshire (fn. 18) —we discover that apparent exceptions can thus be reconciled with the rule in a goodly number of cases, and that the evidence for its application is very considerably strengthened.

One of the interesting facts emerging from Mr. Bates' analysis is the marked difference in character of the assessments in different parts of the county. For instance, in the portion bordering on Gloucestershire (a county in which the prevalence of the 'five-hide unit' is marked) we find, to the west of Bristol, Wraxall assessed at 20 hides, Easton in Gordano and Portbury (adjoining it on the north) at 20 hides between them, Long Ashton (adjoining it on the east) at 20 hides, Blackwell and Barrow Gurney (adjoining them on the south) at 10 hides each, and Yatton (adjoining Wraxall) at 20 hides. Or again, to the south-west of Bath we have the adjoining vills of Twerton, English Combe, Newton St. Loe and Corston assessed, all four, at 10 hides each, while close by, to their south-east, are Hinton Charterhouse, Norton St. Philip, Road and Laverton, all similarly adjoining, and all similarly assessed at 10 hides each. Or, again, if we pass to the south of the county, along the Dorset border, we find a group of adjoining manors—North Perrott, (fn. 19) Haselbury and Hardington—assessed at 10 hides each, while Coker, Pendomer and Sutton Bingham, immediately adjoining them on the east, were assessed repectively at 15, 5 and 5 hides. (fn. 20) It is almost inconceivable that such figures as these, speaking so plainly as they do of an artificial system, should have failed to arouse in Mr. Eyton's mind grave doubts as to his own theory, or have done nothing to open his eyes to the true meaning of hidation. But he belonged, as we shall see, to that class, all too numerous, of antiquaries who can only look at evidence in the light of the theories they hold, and who strive to fit facts to theory rather than theory to facts.

Keeping, however, to the point of the differences presented by districts, we find, at the other end of the county, the district stretching, in Mr. Bates' words, 'from the mouth of the Parret to Exmoor Forest,' vills assessed not only at much smaller amounts, but also in irregular figures. Not merely the virgate or quarter of a hide, but even the 'fertine' or quarter of a virgate, occurs repeatedly in the record of assessment. It may even be said that in the western half at least of this district we lose sight altogether of the 'five-hide unit.' Arid though such investigation may seem, there can be no question that this contrast, well defined as it is, is very closely connected with the early history of the county. In the east we find a system of assessment resembling that which prevailed in the counties to its east and south-east; in the west we recognize a system approximating to that of Devon. Somerset, we are thus reminded, stands between Wessex and 'West Wales.'

From assessment we must now pass to measures of actual area. Having rightly abandoned the attempt of the older antiquaries to assign to the Domesday 'hide' a fixed area, Mr. Eyton devoted himself with ardour to the fascinating task of proving that its lineal and areal measures were so exact that they give us, when interpreted on his method, the actual acreage of the parish to which they refer. In the preface to the two volumes treating of the Somerset Survey he wrote as follows :—

After many months' study of the Somerset Domesday the Author finds nothing to disturb, but very much to support, those principles of criticism and those methods of analysis which were adopted in his 'Key to Domesday' as 'illustrated by the Dorset Survey.' Domesday thus examined, county after county, becomes a Science more and more exact . . . The same principles, the same rules, of Domesday mensuration and values will derive added strength from the Somerset Survey. (fn. 21)

Now this is a matter which must be discussed, for it stands at the threshold of our inquiry. Either Mr. Eyton discovered the true 'key to Domesday,' or the whole of his elaborate calculations, based as they are upon that 'key,' crumble into dust.

Let us first grasp clearly what Mr. Eyton held as to the Domesday phrase 'land for x ploughs,' a formula which meets us on every page. Even the casual reader could hardly fail to be struck by the arbitrary character of the numbers entered in Domesday as those of the ploughlands on many manors at the very outset of the Survey. Taking Mr. Eyton's own tables we find that on the king's land ten manors or groups of manors of ancient demesne (fn. 22) have the number of their ploughlands recorded. One is credited with 100 ploughlands, four with 50, one with 30 and one with 20. In only three cases are the numbers odd. (fn. 23) Yet it did not strike Mr. Eyton as at all peculiar that in no fewer than four cases the arable land should, as he reckoned it, be exactly 6,000 acres. Of the fifteen manors in which the king had succeeded the family of Godwine, one is assigned 60 ploughlands, one 50, two 40, two 15, two 10 and one 5. For those which had come to him on the death of Queen Edith the figures are : One 100, two 40 and one 10. (fn. 24) In all this Mr. Eyton saw nothing peculiar, nor were his suspicions aroused by the fact that the vast and scattered manor of Taunton is assigned 100 ploughlands, or that on the Glastonbury manors we get such figures as 40, 30, 20, 15, 10, 5 in some twenty cases, including four of 30 and six of 20. (fn. 25)

Of the area of such ploughlands Mr. Eyton wrote as follows :—

The normal capacity of the Plough-gang, or One-team-land, was 120 statute acres . . . the normal plough-land, the Terra unius carrucæ was 120 acres. (fn. 26)

For proof we are referred to his Dorset book, where we read as follows:—

The Terra ad unam carucam, or plough-gang . . . is, if not quite a definite expression, intended in Domesday to denote an area of arable land, nearly if not always constant . . . we may reasonably seek to determine the average (sic) contents of a Dorset plough-gang . . .

Let the Terra ad unam carucam or plough-gang stand then, for the present at least, as proximately (sic) implying 120 Domesday acres, and the same number of modern statute acres. (fn. 27)

Now I do not wish to press the point unduly—for it is not of great consequence—but on the basis of this conclusion as to the 'normal' or 'average' plough-gang, Mr. Eyton assumed throughout that each individual plough-gang with which he had to deal contained 120 acres, neither more nor less. (fn. 28)

Yet this inconsistency is of small or no account when compared with his conclusions on 'the more precise system of mensuration . . . by which the meadows, the pastures, the woodland and the wilds . . . were meted in Domesday.' (fn. 29) It is a wonderful instance of the power of self-delusion, and a further warning to those who subordinate facts to theory, that Mr. Eyton was able to convince himself that he had proved his case in Dorset, and that the Somerset evidence did but confirm its truth. For it is open to fatal criticism from three different quarters. The first and most obvious criticism is that when we find Domesday using as its unit of measurement the 'league' (leuga)—which Mr. Eyton (rightly, I think) deemed to be 12 furlongs or a mile and a half—it can only be making a vague estimate. At Winscombe, for instance, there was, we read, 'una leuga pasturæ in longitudine et latitudine, silva ii leugas long. et una leugam lat.' (fo. 90). But Mr. Eyton treated such measures as these as accurate to a yard, and worked them out as representing 120 acres of pasture and 2,880 acres of woodland. Adding 3,600 acres for thirty plough-gangs, and 60 acres of meadow, he arrived at a total of 6,660 for the' Domesday acreage' as against an actual 'parochial acreage' of 4, 140. (fn. 30) But such discrepancies appear not to have troubled him in the least. At Congresbury the woodland and pasture are reckoned by Domesday in the same crude fashion: 'pasture 2 leagues long and half a league wide; woodland 2½ leagues long and half a league wide' (fo. 87). Again accepting these measurements as absolutely exact, Mr. Eyton produced 'Wood-land 1830 acres' and 'Pasture 1440 acres,' which brought up the total Domesday acreage of Congresbury 'to 3,567 acres in excess of the aforesaid parochial measurements.' 'From this, 'he added, 'it follows' that the manor 'had attached to it a quantity of arable land, wood-land, and pasturage, which has never been brought within the parochial boundaries of Congresbury and Wick St. Lawrence.' (fn. 31) It was by such easy assumptions that he disposed of these discrepancies. In Hinton St. George, for example, he could find no room for the '720 acres' of woodland which the Domesday figures gave him, so dismissed them 'as being very probably in another parish,' and then claimed the remaining acreage as what he required. (fn. 32) Having stated his views on Domesday mensuration, he began his 'theories tested by examples' with the Dorset manor of Shillingstone. For this manor Domesday, he wrote:—

gives (according to our principles and calculations of measurement) an area of (1,920 acres of ploughland + 183 acres of meadow + 3,360 acres of pasture + 2,070 acres of wood) 7,533 acres to this Manor, while the present Parish of Shillingstone is only 2,223 acres. The explanation . . . is that the bulk of the pasture-land and all the woodland (say 5,310 acres) were mere adjuncts, probably, nay demonstrably, lying at a distance; in other Parishes, and, it may be, in other Hundreds. (fn. 33)

In his wonderful table of 'the exacter measures of the Dorset Domesday' he even went so far as to make his figures balance by assuming that '22,277¾' acres were 'designedly omitted in Domesday because irrelevant to the Survey,' (fn. 34) and even further assumed an area of '5,531¼ acres appurtenant to the Royal Forests.' (fn. 35) Misdirected ingenuity could scarcely go further than this.

The second criticism of Mr. Eyton's theories of Domesday mensuration admits of being briefly stated. Any student who investigates the matter will find that he could only arrive at the 'Domesday acreage' of his tables by assuming, however unconsciously, that the woodland, pasture, etc., was invariably composed of compact, strictly rectangular blocks like those of an American city. To imagine the county of Somerset or of Dorset laid out in such blocks as these is a conception that one can only describe as unthinkable. But the theorist has a way of shutting his eyes to the working of his theory in practice and of divorcing it from facts.

The third and perhaps the most fatal of these critical objections requires to be driven home. When Mr. Eyton accounted for discrepancies between his 'Domesday acreage' and the actual relative area by bold assumptions such as we have seen, he either had some evidence to support them, or he had not. In the latter case they were mere guesses, introduced only to account for his discrepancies. The test is found in those cases where he plausibly accounted for a difference that existed only in his own imagination. It is obvious that in these cases the explanation he offered must have been quite as imaginary as the difference it was invented to explain. I may venture to quote, from a paper of my own, an instance in point.

It should be noted that on the title-page of his Key to Domesday Mr. Eyton places prominently 'its exactitude of mensuration.' It might almost be described as having been a passion with him to account for every acre in each county on which he wrote.

At first sight, no doubt, his results are wonderful. Every item is exactly accounted for. Yet, alas ! this perfect system will not bear the strain of test. When Mr. Eyton was sure of the total for which he had to account, he was generally able to account for it. But sometimes the total turned out to be wrong after he had accounted for it quite satisfactorily under the impression that it was right. And this, I fear, is fatal. Take, for instance, the case of Drayton in his analysis of the Staffordshire Survey. Mr. Eyton considered that the area for which he had here to account was 3,315 acres. He reckoned up its two constituent manors, the king's and Thurstan's, which contained between them 9 ploughlands—that is, according to what he terms his 'oft-repeated theory,' 1,080 acres + '20 acres of meadow, and (as I,' he adds, 'compute the measurements,) 1,440 acres of wood' (p. 63). One can make nothing of this but 2,540 acres. Yet Mr. Eyton must have reckoned it, by an error of over 1,100 acres, at 3,653 acres, for he tells us that 'the Domesday measures' of these two manors 'imply 338 acres more' than the 3,315 that he had to account for. This supposed excess he at once disposes of by informing us that 'these 338 acres were in Cautwell' (p. 64). But, alas ! in his errata he frankly confesses that the existence of two manors at Drayton was on his part 'a delusion,' as one of the two had been included in error, and that his calculations in Drayton were consequently 'mostly inept'! (fn. 36)

Again, I have been able to show that in Staffordshire what was really the same estate was by him, in three separate places, reckoned at 616, 6,580 and 5,370 acres ! (fn. 37) In Somerset, owing to the same cause (his mis-reading of Domesday), we can apply the same criticism. He rather carelessly read Domesday as assigning to Milborne Port a liability to half a night's ferm (fn. 38) (instead of three-quarters), and then, finding it commuted for a sum equivalent to three-quarters, promptly produced an explanation to account for the supposed discrepancy. (fn. 39) Of this I observed in my Feudal England (p. 113):—

Great masses of Mr. Eyton's work consist of similar guesses and assumptions. Now, if these were kept scrupulously apart from the facts, they would not much matter; but they are so inextricably confused with the real facts of Domesday that, virtually, one can never be sure if one is dealing with facts or fancies.

The climax, perhaps, is reached in his treatment of the Dorset boroughs, which have to be compared with these on the royal demesne in Somerset. His singular misconception of the Domesday entries concerning them led him to make assumptions as to area which rested only on his error. (fn. 40)

It has been absolutely necessary to show, once for all, that all these assumptions have no real foundation, for there is reason to believe that they even affected Mr. Eyton's identifications, several of which have been found to need correction. It is not so much, I must repeat, on the ploughlands as on the woodland and pasture that Mr. Eyton went astray. Yet even on the ploughlands Mr. Bates rejects his theory that their area was always 120 acres, (fn. 41) and cites Professor Maitland's cautious conclusion that when that theory threatens difficulty it can almost always be prevented by the intervention of some plausible hypothesis about shifted boundaries or neglected wastes. More than this has not been done. Always at the end of his toil the candid investigator admits that when he has added up all the figures that Domesday gives for arable, meadow, wood and pasture the land of the county is by no means exhausted. Then the residue must be set down as 'unsurveyed' or 'unregistered,' and guesses made as to its whereabouts. Then further this method involves theories about linear and superficial measurements which are, in our eyes, precarious. (fn. 42)

As against 'Eyton's theory' Professor Maitland asserts that 'the teamland' was 'no areal measure.' (fn. 43)

The student, in fact, must dismiss from his mind, as a wholly baseless delusion, the idea that those responsible for the Survey set themselves the herculean task of measuring out the land. As even Mr. Whale has justly admitted:—

Domesday items of area are after all only roughly approximate; the great purpose of the survey was fiscal, other items incidental. Not a tittle of evidence exists of a survey of area. The whole work was completed in less than a year. When the jury . . . answered in multiples of plough lands, very rarely taking notice of a half, we may safely conclude that fragments of a plough land were disregarded; in other words, the question was answered without previous thought and with limited knowledge. (fn. 44)

A very little thought will show that certain information could easily be given and that other information could not. The assessment of a manor, or the amount of dues paid by custom to the Crown, were matters for exact statement; it was also possible to state accurately the number of plough-oxen on the land, the figures of the live-stock and the classes of peasantry. But it was not possible to give exact measurements of the land, nor would such measurements directly affect taxation. They would to a large extent be little better than guess-work, and to base upon them an exact acreage is obviously out of the question.

Another of Mr. Eyton's delusions needs to be here exposed, for it is one on which he expressed himself in terms of great confidence. We read in the preface to his Somerset Survey (p. 11):—

In his translation of Domesday entries Collinson uniformly adopted the error which confused the carucata terræ with the terra ad unam carucam. It is difficult to see how any one with the Exon survey of any county before him could remain in such an error. Nevertheless there are Domesday scholars, of deserved repute, still living and still writing, who abide by this inveterate misconception. Would they but study the original text more critically, would they but accept the precise interpretation furnished by the correlative expressions of the Exon Survey, they might at length learn that the Domesday Commissioners and scribes did not crowd nearly every sentence of their works with a vain tautology, nor yet indite matter which, well and literally translated, could fail to be significant in every word.

Oddly enough it was the writer himself who swept aside the language of the Exon Survey and its scribes when he found it in con flict with his own theory on the subject. Observing that at Bedminster the terra ad unam carucam of Domesday Book was entered in the Exon Domesday in one place (p. 83) as terra unius caruce and in another (p. 489) as I carrucata terræ, he promptly disposed of the latter phrase by writing—

Here the word carucata is improperly used. It was merely a clerical error, the clerk describing one thing in terms of another. (fn. 45)

The truth is that the scribe, on the contrary, knew what he was doing and simply employed a synonym after the manner of his kind.

The great lordship of Taunton affords us another illustration. Let us place the geld-roll for Taunton Hundred and the survey of Taunton in the Exon text side by side:—

Geld Roll

Habet adhuc episcopus W. in Tontona xx carrucatas terræ quæ nunquam reddiderunt Gildum. (fn. 46)

Exon Survey

Exceptis his predictis hidis habet ibi episcopus terram ad xx carrucas quæ nunquam redd [id] it gildum. (fn. 47)

Here the synonymity of the phrases is placed beyond doubt. And yet Mr. Eyton had no difficulty in persuading himself that the terram ad xx carrucas of the second passage should be read in conjunction with a subsequent entry of woodland, pasture and meadow 'to make the 20 plough-lands into true carucates' by the addition of '4,360 acres of other land' ! (fn. 48) But on turning to the Exon text we find that this woodland, etc., is not assigned, as he makes it, to these twenty ploughlands, but to the whole lordship of Taunton, (fn. 49) a very different matter.

Mr. Eyton's views on the 'carucates' of Somerset will be found on pp. 20, 28–9 of his Somerset Survey, (fn. 50) and were criticized by me on pp. 202–4 of Domesday Studies (1887), where I showed that the terra I caruce, or ploughland, of Domesday was equated in the Inquisitio Eliensis by I carucata or I carucata ad arandum. The geld carucate (carucata ad geldum) of northern England was quite another matter; but with that we have nothing to do in Somerset.

From this preliminary but necessary discussion we may now at length turn to the distribution of the lands of the county at the time of the great Survey.

Somerset was a county in which the king's lands were of great extent and importance. For, as was duly explained by Mr. Eyton, he held in 1086, not only the ancient demesne of the Crown, which had been held by Edward the Confessor, but the forfeited possessions of Harold and his relatives, and the broad estates of Queen Edith, which had escheated to the Crown at her death in 1074. To these were added a few manors which had been held by Wulfward the White (Albus), a thegn apparently of that queen, so that one cannot wonder that, in all, the manors in William's own hands occupy no less than five columns of Domesday Book.

Taking in succession these constituents of King William's holdings, we have first to deal with the 'ancient demesne,' which is not distinguished from the rest in Domesday Book itself, but which is defined in the Exon Domesday as 'D[omi]nicatus regis in Sumerseta' (p. 82). (fn. 51) This consisted, as is shown in Mr. Eyton's table, (fn. 52) of twelve manors: Carhampton, Williton, Cannington, North Petherton, South Petherton, Curry Rivel, Frome, Bruton, Somerton (with the 'borough' of Langport), Cheddar (with Axbridge), Bedminster, Milborne Port (with Ilchester). These manors possessed two distinctive features: in the first place they were not liable to pay 'geld' (or land tax), and were therefore not assessed in 'hides' (or units of taxation); in the second, they had formerly paid to the Crown, instead of a rent in money, the firma unius noctis (or supply in kind to the Court), and had been grouped, on an ancient system, for the purpose. (fn. 53) This grouping is so remarkable that I here give, from my Feudal England (pp. 111–2), the details of the money commutation for this ancient burden:—

I have already spoken of Mr. Eyton's error in taking the Milborne payment as commutation for half instead of (as it was) for three-quarters of a firma. (fn. 54) With his suggestion that Milverton, if not Bath itself, had formerly belonged to this group of manors of ancient demesne I cannot agree, (fn. 55) while his assignation to the twelve manors of certain estates as appendages was avowedly tentative. (fn. 56)

There was yet another feature by which these manors were characterized, although it was not, like the two others, distinctive and peculiar to them. This feature was that each of them, as Mr. Eyton explained, was the head of an old Hundred, to which it gave its name. (fn. 57)

One finds it very difficult to say on what ground Mr. Eyton, while rightly urging that to each of these manors was attached the lordship of the Hundred named after it and that the Court of the Hundred was probably held within its limits, insisted that the manor itself was always extra-hundredal. (fn. 58)

As I had occasion to observe, many years ago, 'Mr. Eyton had, unluckily, a provoking habit of occasionally omitting his evidence, and thus rendering it impossible to judge of the correctness of his induction'. (fn. 59) The witness of Domesday as to those counties in which it gives hundredal headings is dead against his assertion. The best parallel, perhaps, is afforded by the case of Hampshire, in which the royal demesne manors had been similarly held by King Edward, were similarly free from assessment to geld, and had similarly contributed on an ancient system to the 'day's ferm,' or rent in kind. Yet Domesday distinctly enters them as intra-hundredal. (fn. 60)

The issues of the king's manors, the Terra regis of Domesday, formed the nucleus of that firma comitatus for which the sheriff answered at the Exchequer and which constituted so important an element in the revenue of our Norman kings. Here one is again compelled to differ from Mr. Eyton's views, for the interpretation of a notable passage in the Somerset Domesday is at stake. With the king's manors of ancient demesne the sheriff, Mr. Eyton held, had at that time nothing to do:—

It is worth note that in no instance of the Twelve Royal Manors of Somerset is a manor said by Domesday to be in the ferm or custody of the Sheriff of the County.

Such manors were extra comitatum always. In William's time they were extra vice-comitatum also. . . . The Royal manors were fermed or managed by special Præpositi, not necessarily, nor always, resident, who were answerable to the Crown for fixed rents, whatever were the profits or losses of management. (fn. 61)

Evidence to the contrary could be produced from several other counties, but we need go no further than the immediately adjacent one of Gloucestershire to find in a single column of Domesday proof that the sheriff was responsible for the firma from manors of ancient demesne. (fn. 62) Next to Gloucestershire comes Worcestershire, where the language of Domesday on the subject is pre-eminently clear and decisive: 'de dominicis Maneriis regis reddit (vicecomes) cxxiii lib. et iiii sol. ad pensum' (fo. 172).

The reason for insisting upon this fact is that a passage in the Somerset Survey is of absolutely unique value as proving that, before Domesday and even before the Conquest, the twelfth century exchequer system of crediting the sheriff, at his annual account, with a fixed sum in respect of such royal manors or portions thereof as had been granted out to subjects was already in full operation. The passage in question is found in the entry on Cheddar:—

'De hoc Manerio tenet Giso episcopus unum membrum Wetmore, quod ipse tenuit de rege E[dwardo]. Pro eo computat Willelmus vicecomes in firma regis xii lib. unoquoque anno' (fo. 86) (fn. 63); 'sed episcopus tenuit de rege Edwardo longo tempore ante obitum regis E[dwardi]' (fn. 64)

I need not here repeat the argument in which I showed that computare was the technical word used at the exchequer for this system. (fn. 65) Ignorance of that fact, unfortunately, led Mr. Eyton, not only to miss the interest of the passage, but to give it an interpretation altogether erroneous:—

We have not noticed in this column an annual rent of £12 stated by Domesday to be paid by Bishop Giso for Wedmore as if it were a member of Cheddar. . . . Bishop Giso's payment of a rent for Wedmore was only a temporary incident. Elsewhere (p. 147, Exon) Domesday surveys the manor as the Bishop's in all integrity. . . .

Wedmore was not then, in any strict sense of the term, a member of Cheddar. It was not Royal Demesne; it was intra Vicecomitatum and intra Hundredum, and geldable. All that we can suppose is that, owing to some claim of the Crown, Wedmore had been seized by the sheriff, William de Moione; but Bishop Giso had been con tinued in formal possession, at a stipulated rent. The rent was received from him by the said Sheriff, who was probably instructed to hand it over to the King's Provost at Cheddar. The affair was quite ephemeral. Doubtless Giso had recovered full seisin even while the Domesday Commissioners were in Somerset. (fn. 66)

The whole idea is illusion. Domesday does not state that the bishop paid a rent for Wedmore, and the sheriff neither received this imaginary rent, nor made it over to that 'Provost at Cheddar,' whose existence rests solely on another erroneous assumption which I have exposed above. (fn. 67) Wedmore had undoubtedly formed part, as Domesday states, of the ancient demesne till Edward the Confessor granted it out to Bishop Giso some quarter of a century before the date of the Survey. Oddly enough, Mr. Eyton was fully aware of this grant. (fn. 68)

Other features of this group of manors, such as the arbitrary reckoning of its ploughlands, (fn. 69) the royal boroughs that it included, and its richlyendowed churches, are discussed under other heads, but attention may be drawn here to the cases of estates which had been added to these manors or had been subtracted from them. (fn. 70) The latter, as was pointed out by Mr. Eyton, appear in Domesday assessed in terms of the hide, in spite of the fact that the manor from which they had been withdrawn had not been so assessed.

At the close of that group of manors with which we have been dealing Domesday begins a fresh column with that section of the king's lands which had come to him by the forfeiture of the whole house of Godwine. These, as explained above (p. 394), are entered in the Exon Domesday in a separate place with a distinct heading, and, as Mr. Eyton pointed out, there is found at the head of a folio towards the close (but in a smaller handwriting) the title 'mansiones de comitatu.' He therefore styled them 'the comital manors,' a name which we may adopt. It has also been adopted by Professor Maitland, on the strength of this Somerset heading confirmed by the phrase comitales villæ, which he cites from the 'Pseudoleges Canuti.' (fn. 71) He observes that the existence of such manors held ex officio may account for the vastness of those estates which appear in Domesday as having formerly been held by the house of Godwine, and that 'the wealth of the earls is a matter of great importance.' (fn. 72) But neither he nor Mr. Eyton, it would seem, noticed the great peculiarity by which these manors are, in Somerset, distinguished. This consists in the entry of their rentals on a quite peculiar system: (1) They are entered as renders ('reddit'), not as values ('valet'); (2) the sums rendered are 'de albo argento'; (3) in at least ten out of the fifteen cases they are multiples of the strange unit £1 3s. (fn. 73) Whatever may be the explanation of this singular system, it differentiates the comital manors very clearly from others. I have shown that there are visible traces, in the adjacent county of Devonshire, of the rents of the comital manors being based on a duodecimal system, (fn. 74) but it is not nearly so peculiar as that of their Somerset rentals.

The difficult question of the 'third penny' (tercius denarius) as appurtenant to comital manors will be discussed below in connection with finance, but this would seem to be the fitting place in which to speak of those by whom the manors were held. Although it would be naturally expected that all comital manors would remain in the hands of Harold himself, it will be found that his mother Gytha, his brother Earl Tostig, his sister Gunhild, and his own son Godwine, were among their holders. We cannot tell, however, whether they had received them from Harold or from his father Godwine, his predecessor in the earldom. But it is the tenure of Crewkerne that is the real puzzle. As was duly pointed out by Mr. Eyton, who discussed the matter at some length, (fn. 75) Crewkerne was 'by far the most valuable of the comital manors of Somerset,' and yet it had been held by 'Eddeva,' a lady whose identity is not disclosed. Now it is one of the problems of Domesday, as Mr. Eyton observed, whether its Edith 'the fair,' 'the beautiful,' 'the rich,' was identical with Edith 'Swan's neck' (Swanneshals), Harold's mistress, and probably the mother, in Mr. Freeman's opinion, of his son Godwine mentioned above. As to this problem Mr. Freeman held that 'there is absolutely no evidence either way.' (fn. 76) Mr. Eyton, discussing the question as to who was the tenant of Crewkerne, summed up strongly in favour of its being 'Edith the Fair,' who was, he suggested, 'of the race of Godwin'; but I do not think his reasoning sound. Even if the manor had previously belonged, as he held, to King Edward—which seems to me doubtful (fn. 77) —it does not follow that the king must have given it to 'Eddeva,' and that, therefore, 'Eddeva' cannot have been Harold's mistress. Harold might well have obtained it from the king and then bestowed it upon her. Moreover, 'Edith the Fair' was in no case a daughter of Godwine, nor have we any reason to suppose that she was even 'of kin to him.' It is difficult, therefore, to see how she obtained the manor unless it was given her by Harold.

I agree with Mr. Eyton in thinking that the manors of 'Estalweia,' Banwell and Lullington, which had also been held by Harold, were not 'comital manors'; but I differ from his view that Temple Combe ('Come'), which had been held by his brother Leofwine, was one of them, for which supposition there seems to be no ground.

The third division of King William's lands consisted of those which had come to him on the death of Queen Edith. These have a separate heading in Domesday Book, while in the Exeter Domesday they are, further, entered in a separate place (p. 104). It is worth observing that the form of her name is 'Eddid' in the former and 'Editda' in the latter, which is more akin to our modern form than the usual 'Edeva' of Domesday or Mr. Freeman's 'Eadgifu.' The queen's Somerset possessions were exceptionally valuable and extensive, including, as they did, the lordship of Bath. Of those in the king's hands at the time of the Great Survey we learn from Mr. Eyton's table that the assessment was about 140 hides and the value the then large sum of some £320. But, in addition to the manors that had passed to the king, she had also held others at Luccombe (now Luckham), Selworthy, Combe Hay and Twerton; but these were, comparatively, of little value.

The fourth and last section of the king's great holding was composed of the lands of Wulfward White (Albus) and receives, in both texts, a separate heading like the lands of the queen, of whom Wulfward, we shall see, had acted as officer. So shadowy are the names of the great English thegns who meet us as holding lands on the eve of the Conquest, and so slight our knowledge of the distribution of their lands that it is worth our while to ascertain the history, position and estates of Wulfward so far as possible, the more so as he clearly escaped, doubtless through the queen's influence, the general wreck at the Conquest. The first point to attract our attention, in his as in other cases, is the scattered character and great extent of the lands held by an English thegn. Mr. Eyton did something to trace his estates in Domesday, (fn. 78) but his search was not exhaustive. In Dorset he had held at least Pentridge and Silton; in Hampshire (fo. 43b) he is entered as having held Hayling Island of Queen 'Eddid' as an 'alod,' Domesday adding that he had received half of it for life only from Queen Emma (Canute's widow) and had died in King William's time. (fn. 79) Far away in Kent he had held a manor with 'sac and soc' (fo. 1b, 9); while in Middlesex, as 'a thegn of King Edward,' he had owned three manors (fo.129–30). Tracing him from Somerset in another direction, he had held land in Wilts under the Bishop of Salisbury (fo. 66); in Gloucestershire Hatherop had been his (fo. 169), and in Oxfordshire a part of Cold Norton (fo. 160), while in Buckinghamshire he and his wife Eadgifu ['Eddeva'] had held land, between them, at some ten different places, which were assessed in all at about sixty hides. Even in distant Lincolnshire 'Wlward Wite' was one of those notabilities who enjoyed 'sac and soc' (337), a fact which proves that he was the 'Wlward' who had held a valuable manor at Butterwick in that county. I have also no doubt, myself, that he was the 'Wlward' whom Ernulf de Hesdin had succeeded in a valuable Berkshire manor, which is now Newbury (fo. 62b). We thus connect him with ten counties.

In Somerset itself, besides the manors of Corton Denham, Mudford and Pitney, he had held estates at Stanton Drew and Burnett under Queen Edith's lordship of Keynsham, and the latter was still in possession of his widow in 1086, a pitiful fragment of his wide estate. Mr. Eyton showed that Wulfward was present, in 1068, at William's court and at that of Queen Edith in 1072, witnessing in both cases Somerset grants. The geld roll implies that he was still living in 1084. From the Buckinghamshire entries we may gather that his daughter was given by Queen Edith to 'Alsi,' who was only allowed, however, to obtain land with her in that one county. (fn. 80)

From the king's land we may now at length pass to that of his subjects.

Awarding to the churchmen the precedence they invariably receive in Domesday, we have first to deal with the Bishop of Winchester. Pitminster and Rimpton he held as abbot of the Old Minster (St. Swithun's); Bleadon was apportioned to his monks; but these possessions become insignificant when compared with his lordship of Taunton. Three folios of the Exon Survey and about a column of Domesday Book are devoted to this lordship alone, and they supply material for a dissertation on many points of historical interest and institutional importance. While its local aspects have been discussed very fully by Mr. Eyton, it is Professor Maitland who has done most to illuminate its origin and character by his comments in Domesday Book and Beyond, in which work it is referred to in eight separate places. The first three points to be settled are: (1) How did the bishop come by Taunton? (2) What did 'Taunton' comprise? (3) Which were the places that owed suit and service to Taunton?

Walchelin, the Norman Bishop of Winchester, had obtained Taunton, Mr. Eyton tells us, by 'gift' or 'grant' of the Conqueror. It will be best to quote his own words :—

The great Somerset Manor of Taunton had been before the Conquest a possession of Stigand, sometime bishop of Winchester, sometime Archbishop of Canterbury, sometime occupant of both sees together.

After the Conquest, King William gave, or perhaps we should say confirmed, Taunton to Walchelin, the first Norman bishop of Winchester. The King conceded it, not as a personal and heritable feud, but as an endowment of the church of Winchester.

There are many records of this gift (sic): but we hardly appreciate its munificence and its fulness till we turn to Domesday. . . .

The elaborate care bestowed upon the Record and its unusual comprehensiveness are accounted for in a postscript which runs as follows:— 'De his terris' (the adjuncts of Taunton) 'semper jacuerunt consuetudines et servitium in Tantone, et Rex Willelmus concessit istas terras sancto Petro' (the church of Winchester) 'et Walchelino episcopo, sicut ipse' (Rex) 'recognovit apud Sarisberiam audiente Episcopo Dunelmensi, cui præcepit ut hanc ipsam concessionem suam in Brevibus scriberet.'

If we venture to translate somewhat freely, and to expatiate somewhat enthusiastically on this unwonted text, we trust at the same time to say no more than years of study and an ever growing tendency to prefer facts to views will warrant. (fn. 81)

. . . On him (the bishop of Durham), then, did William, King of the English, being at Salisbury, enjoin that he should take present and diligent note of the King's declared intentions touching the quality and extent of his grant of Taunton (sic) to Bishop Walchelin; and that when he (the bishop of Durham) should, in the course of his circuit, be in Somerset, he should enter a memorandum of such, the king's 'concession,' on the Rolls of the Survey—'in brevibus,' as the King named the elements of that, the contemplated Record, which men afterwards called Domesday Book. (fn. 82)

It will be seen at once that Mr. Eyton here speaks of the grant in one place as that of 'the adjuncts of Taunton,' and in another as that of 'Taunton'; and that his evidence for both statements is the 'postcript' cited above. Now I do not hesitate to say that this postscript referred neither to Taunton nor to its adjuncts, but to two estates in Lydeard (St. Lawrence) and (Angers) leigh, of which Domesday expressly tells us that they had been 'added' to Taunton, (fn. 83) and were held of Bishop Walchelin by grant (concessionem) of King William. (fn. 84) And this conclusion is in harmony with the evidence of the whole entry, which shows that Walchelin had but stepped at Taunton, as elsewhere, into his predecessor's shoes.

As against these additions the bishop's great lordship had suffered three encroachments, (fn. 85) though these were not serious. The Count of Mortain, John the Usher, and Alvred 'de Hispania' had thus secured between them 4 1/8; 'hides.' Mr. Eyton wrote of this as a 'grant made by King William . . . in qualification, or perhaps in non-anticipation of the King's more general grant of Taunton and its appurtenances to the Bishop of Winchester ' (fn. 86); but this latter 'grant' was, we have seen, an error, and as these encroachments are also found among the 'terræ occupatæ,' (fn. 87) there can be no reason for doubting that they were merely such aggressions as were frequent on church lands in the early days of Norman rule. In such cases the lands were rarely recovered by the church.

It is partly by comparing the Taunton lordship with those in other counties held by the Bishops of Winchester, and partly by invoking the evidence of Anglo-Saxon charters, that Professor Maitland has been able to bring the lordship of Taunton into relation with its fellows and to direct us on the right path for tracing its earlier history. In the first place he has drawn our attention, in Hampshire, to 'the great Chilcombe estate of the church of Winchester, which stretched for many a mile from the gates of the royal city,' and which was granted as 100 hides; in Wiltshire to its great Downton estate, entered in Domesday as 100 hides; and in Somerset to 'Taunton' as the name of a territory comprising several parishes. From these instances he draws the inference that 'whenever the West-Saxons conquer new lands they cede a wide territory to their bishop.' (fn. 88) In other words the bishops, he held, had acquired these territories 'in very early days.' (fn. 89)

When from the character of these great lordships, with the special privileges attached to them, we pass to what we can learn of their history, we find evidence which throws, I think, a curious reflex light on the Domesday Survey of Somerset. Although primarily interested in matters of justice and jurisdiction, Professor Maitland incidentally brings before us privileges of another kind. For the former he refers to a document of special interest as follows—

in a book (fn. 90) of fairly good repute we may read of the grand liberties with which in 904 King Edward endowed the Bishop of Winchester's large estate at Taunton— that estate which in subsequent centuries was to become the classical example of colossal manors. (fn. 91)

The critical study of such documents is not yet sufficiently advanced for us to say with certainty which of these ancient texts are wholly or partly spurious; but it is always a sound principle to accept admission of liabilities even if we look with suspicion on assertion of privileges and rights. I am, therefore, disposed to accept as genuine the important passage which follows, the more so as it is in harmony, we shall find, with known facts. Professor Maitland's rendering is this—

Taunton, which belonged to the Bishop of Winchester, had been bound to provide one night's entertainment for the King and nine nights' entertainment for his falconers, and to support eight dogs and a dog-ward, to carry with horses and carts to Curry and to Williton whatever the King might need, and to conduct wayfarers to the neighbouring royal vills. To obtain immunity from these burdens the bishop had to give the King sixty hides of land. (fn. 92)

We have here, in this grant of a thousand years ago, that firma unius noctis with which we have already met as the due payable from royal manors or groups of manors in Somerset, (fn. 93) and it is at least suggestive that the '100 ploughlands' of Taunton are parallel in Domesday Book with the '100 ploughlands' of the Williton group or of the Bruton and Frome group, each of which was liable to provide the firma unius noctis. But it is the further burden of providing for hawk and hound and for carriage of the king's stuff from one of his manors to another that gives to this evidence, I think, a genuine look. As Professor Maitland has said of such charters—

Apparently the king, the under-king, even the ealdorman, has a certain right of living at the expense of his subjects, of making a progress through the villages and quartering himself, his courtiers, his huntsmen, his dogs and horses upon the folk of the townships, of exacting a 'one night's farm' from this village, a 'two nights' farm' from that. (fn. 94)

In Somerset, indeed, Domesday mentions nothing beyond the 'night's farm,' but other counties supply us with examples, evidently archaic, of payments for hawk and hound, for sumpter horse or for carriage (avera). (fn. 95) A careful examination of the record reveals payments for hounds of £23 from Oxfordshire and from Warwickshire respectively (ff. 154b, 238), and of £42 from Northants (fo. 219), while in Gloucestershire three royal manors had to supply them with 3,000 loaves (panes) apiece, and in Beds two of them paid respectively £6 10s. 0d. and £3 5s. 0d. for the same purpose. (fn. 96) In six counties also we have identical payments for a hawk and for a sumpter horse. Looking at the whole evidence, I am disposed to think that the Somerset 'night's farm,' of which we read in Domesday, included certain 'customary' payments such as are mentioned in other counties, and indeed in the above Taunton charter, as additional. This might explain certain differences and oddly uneven sums in the totals of the money equivalents.

Another indication of the early date at which Taunton must have passed into the bishop's hands is found in the extraordinary privileges with which its lordship was endowed. In this respect it is instructive to compare it with the Bishop of Worcester's Hundred of Oswaldslow, in which that prelate enjoyed similarly exceptional privileges. (fn. 97) The 'customary dues ' enjoyed at Taunton by the bishop, its lord, included the profits of certain 'pleas of the highest class,' as Professor Maitland terms them, together with the local 'hundred-pence, Peter's pence and churchscot.' It was at Taunton that the pleas were held, and to Taunton that men were compelled to go when they had to make oath or to undergo the ordeal. (fn. 98)

As to the second portion of our inquiry, 'the vills implied in the Domesday survey of Taunton,' as Mr. Eyton has expressed it, (fn. 99) he made them to be Kingstone (13½ miles distant), Combe Florey, Nynehead Flory, Withiel Florey (16 miles distant), (fn. 100) Corfe, Orchard Portman, Ruishton, Staplegrove, Trull, Cothelstone, Bishop's Hull, Otterford and Bishop's Wood, and part of Bishop's Lydeard. (fn. 101) I have already discussed and rejected Mr. Eyton's system of obtaining from the Domesday figures exact measures of area, but one may point out that, applying them to the 'Taunton' of Domesday, he arrived at 16,659 acres as the total area implied. (fn. 102) But this was on the assumption that its ploughlands, containing 'about 120 statute acres,' (fn. 103) were '20 + 80 = 100.' But Professor Maitland read the text as 20 + 100 (fn. 104); which seems to me to be clearly right, especially if we examine Mr. Eyton's own table (ii. 33); and this gravely affects the total. He preferred the Hundred to the single vill as a test of his theories, and claimed that the Domesday Hundreds of Taunton and Pitminster, of which he reckoned the areas by 'the exacter measures of Domesday' at 33,814 statute acres, showed a discrepancy of only 8,672 acres 'between the two measurements, that of the 11th and that of the 19th centuries.' He therefore 'opined' that' in this district or province the Domesday surveyors found some 8,672 acres of moor or waste of which they took no notice.' As already explained, it was by such assumptions that he endeavoured to explain away the failure of his so-called 'key.'

Our third point, the suit and service due to the bishop as lord of Taunton from the vills beyond its limits, is one of great importance. Professor Maitland on this subject, which is one of special interest to him, observes that—

Within his immense manor of Taunton the bishop of Winchester has pleas of the highest class, and three times a year without any summons his men must meet to hold them. (fn. 105)

This is a somewhat serious slip, due to misapprehension. The Domesday Book text, which he cites as his authority in a note, is particularly clear upon the point. After closing its survey of the great manor of Taunton it begins a new paragraph thus:—

Istæ consuetudines pertinent ad Tantone, burgheristh, latrones, pacis infractio, hainfare, denarii de Hundret et denarii Sancti Petri, Circieti, (fn. 106) ter in anno tenere placita episcopi sine ammonitione, profectio in exercitum cum hominibus episcopi. Has denominatas consuetudines reddunt in Tantone hæ terræ:—

Then follows a list of the places (terræ) which rendered the above dues and services, places which were not 'within,' but without the manor of Taunton, and from which the men of other lords assembled for pleas, host, and so forth 'with the men of the bishop' (cum hominibus episcopi). In addition to the dues described above, the lords of all these places, we read lower down, had to be buried at Taunton—which brought, if one may speak colloquially, more grist to the bishop's mill (fn. 107) —and finally all who dwelt within them were bound, as already mentioned, to swear their oaths and undergo their ordeals at Taunton. In short the whole paragraph deals with consuetudines—the normal term in such cases —due to the great manor of Taunton from lands outside it and forming an addition to its value.

But some of these places enjoyed exemption from one or more of these duties, and of such exemption both the Surveys contain careful record. The variants here in their texts are of some interest and importance; both for instance agree in stating that Stoke (St. Mary) and 'Scobinalre' (or 'Scobindare') were exempted from the duty as to the host, while Bagborough was free from this and from the burial provision (sepultura) as well; but while Norton Fitzwarren, Bradford, Halse and Heathfield are entered together in Domesday Book among the places owing all the duties, the Exon Survey enters them apart as owing (only) attendance at the pleas, Peter's pence, and (apparently) Hundred pence. (fn. 108) Again, the Exon Survey includes 'Lediart et Lega' among the places liable to all the dues, while Domesday Book does not; but the latter duly asserts that they had been so liable when dealing with them lower down as having been 'added' to Taunton. (fn. 109) The Exon Survey here indulges in needless repetition.

There was no other lordship belonging to the church in Somerset comparable, even distantly, to Taunton in importance. But the great fief of the Bishop of Coutances fills some five and a half columns of Domesday Book. This fief, however, like that of the Bishop of Bayeux —which was represented in Somerset by one manor only—was akin, in reality, to the lay baronies and was so dealt with on the bishop's death. The fact that it did not descend, in later days, as a whole makes it difficult to trace the devolution and, therefore, the identity of its manors as recorded in 1086. Of the true church lands in the county the most important were those of the Bishop of Wells, and of the abbeys of Bath, Glastonbury, Muchelney and Athelney. In the early days of the Conqueror's reign the see of Wells had recovered the manor of Banwell, which had been wrested from it by Harold, but there are signs in Domesday that the imposition of knight-service on lands of the Church (fn. 110) was already making itself felt in the frequent mention of knights (milites) as quartered on the bishop's manors. But the famous abbey of Glastonbury was the chief sufferer in Somerset, and its chroniclers' complaints of the loss of its lands by their distribution among the Norman knights are confirmed by the Domesday Survey. Portions of its thegnland were annexed to the fief of the Bishop of Coutances, but its chief loss was caused by the imposition on the abbey of a quota of forty knights. (fn. 111) To supply this large contingent knights had to be enfeoffed, which involved the practical alienation of many a manor. Domesday shows a number of these in the hands of Roger de Courcelles, whose successors were responsible for ten knights, that is for a fourth part of the abbey's large quota. Muchelney was only called upon to provide one knight and Athelney escaped free.

Somerset was the only county in which St. Peter's of Rome re ceived an English manor as the reward of the Pope's encouragement of William's conquest. On it he bestowed the Puriton estate, and on his own abbey of St. Stephen of Caen the richly endowed church of his royal manor of 'Cruche' (Crewkerne). To the lot of the abbey of Jumièges there fell the much less richly endowed church of his manor of Chewton Mendip. Foreign churchmen shared the spoils; Maurice, who had just been appointed to the see of London, was holding in 1086 the churches of Ilchester (St. Andrew's), Congresbury, and North Curry, the first of which, with its valuable glebe, had been detached from the possessions of Glastonbury; Richer d'Andely, who had secured the church and glebe of Stogumber, was a Norman who held three churches in the neighbourhood of Southampton, and four houses in that town together with certain manors in the county. The wealthy church of Frome, as well as that of Milborne Port, was in the hands of Reinbald, a remarkable man of uncertain origin, who enjoyed the favour of William as of Edward, and whom I have elsewhere styled 'the first great pluralist.' (fn. 112)

In this county Domesday displays a singular lack of system in dealing with the churches in the king's gift. Some are dealt with in their normal place, under 'Terra regis,' but to others is assigned a special section, 'Clerici tenentes de rege,' which is oddly sandwiched in between the church tenants and the lay. From this special section we learn that Peter, bishop of Chester, who had died shortly before the survey, had held the royal churches of North Petherton, Carhampton, and Kilmersdon. Peter, who had been a chaplain of William's, and possibly of Edward's before him, (fn. 113) had also obtained a small estate at Buckland St. Mary. In Dorset he had secured a hide, probably glebe land, (fn. 114) while in Berkshire he was provided for out of a glebe estate at Wantage, his 'son' Reinbald also obtaining a manor. (fn. 115) Another chaplain was Stephen, who held Milverton church, while Erchenger, who had secured that of Cannington, must also have been a foreigner. An English priest, Alviet, is entered in this section, but not in connection with any church, while three other Englishmen included in it were apparently not priests, but holders of small estates in almoin, (fn. 116) as were certainly three nuns who are comprised in it. Indeed the Exeter Domesday arranges (pp. 178–181) these lands very differently, throwing together sections xi, xii, xiii, xv, and xvi of the Exchequer text into a confused group, to which it gives the heading: 'Terra (sic) que date sunt Sanctis in elemosina in Sumerseta.' In the Somerset returns at the great inquest of 1212 we find a small holding at Holnecote (in Selworthy) still held by the true almoin tenure (free from all service but prayer) and seem to be told that Edith the nun was really the widow of a man slain in the king's service. (fn. 117)

Nigel the physician, who, in this as in five other counties, had succeeded Spirtes the priest, had bestowed a manor in Somerset as in Wilts on the Norman abbey of Montebourg; and that of Grestain had obtained from its patron, the Count of Mortain, the manor of Norton-sub-Hamdon. The Earl of Chester had given Henstridge to his own abbey of St. Sever, but, on the whole, the alien churchmen had not obtained much. It was at a later date than this that Cluniac monks made their home in the priory of Montacute below the castle of its founder, the Count of Mortain.

On one result of the Norman Conquest on church lands in Somerset we obtain light from the survey, not of Somerset, but of Bedfordshire and of Bucks. The 'Old Minster' of St. Swithun, Winchester, had suffered heavily from the change; (fn. 118) but in Somerset its actual loss was confined to the manor of Crowcombe and to Heal, a dependency of Taunton which William's grasping brother, the Count of Mortain, had seized. Crowcombe had been given to the minster by Harold's mother— apparently after her husband's death (1053)—who gave Bleadon at the same time. (fn. 119) Bleadon duly appears in Domesday as apportioned to the support of the monks of Winchester after as before the Conquest ; but in another part of the great survey we find estates in Bedfordshire and in Bucks held by Geoffrey, Bishop of Coutances, 'in exchange for Bleadon.' (fn. 120) This would seem to imply that the bishop had laid hands on Bleadon, and had only disgorged it on receiving compensation in another part of the kingdom. Possibly the excuse in both cases was that the manors had belonged to the house of the dead 'usurper.'

Everywhere the English houses were liable to Norman spoliation; Athelney had lost to Ralf de Limesi its manor of Bossington in Porlock, although its issues were specially assigned to the monks' support; and Cerne Abbey had failed to regain an estate in South Cheriton which Alwold had leased from it for life, because, as in similar instances, it was confused with his own land. Another example of this fruitful source of loss to religious houses occurs at Wheathill, which Elmer had held of Glastonbury Abbey, 'nec poterat ab ea separari.' At the survey it is found in the hands of Serlo de Burci, evidently because he had obtained the lands of this Elmer or Almar. The nuns of Shaftesbury, however, had added to their solitary manor of Combe Abbas that of Kilmington, which was given them by this same Serlo, when his daughter took the veil among them.

With the church lands, as with those of the king, (fn. 121) Domesday betrays a lack of system. Under the fief of Roger 'de Corcelle' we read that his three hides at Puckington had formerly been held 'of St. Peter's church,' which Mr. Bates holds to be Muchelney Abbey. Yet under Muchelney's fief we cannot trace Puckington or find the abbey making any claim to this estate. In the entry which follows those on Puckington, we read that the 1 1/8; hide to which it relates was held of Muchelney Abbey, 'nec poterat ab ea separari.' In this case we have the further information that the land is 'de xx hidis de Draitune et est Tainlande.' We turn to Drayton, a Muchelney manor, but the only encroachment recorded under it is one of two hides, which the details prove to be distinct. But, with the help of the Exon Survey, we discover that Domesday only accounts for 11 5/8; hides of desmesne, 5 hides in the hands of the villeins, and 2 hides of encroachment. There remains, therefore, to be accounted for 1 3/8 hide, which would more than cover this holding of Roger. The fact that such encroachments are sometimes recorded under the religious houses affected, and sometimes not, makes it extremely difficult to trace them out.

A curious feature, in this county, of the Exchequer Domesday is the appearance, in the margin of the text, against the entries of certain manors, of a Maltese cross. We find it against the Count of Mortain's manors of Crowcombe and Heal, of which he had robbed St. Swithun, as of Tintinhull and Kingstone, of which he had deprived Glastonbury. It calls our attention to the fact that Roger 'de Corcelle' was holding at Long Sutton land which two thegns had held of Athelney Abbey, 'et non poterant ab ea separari'; it stamps Roger Arundel as the wrongful holder of Ash Priors, stolen from the bishop of Wells; and it stands against the record of Glastonbury's right to Brompton Ralph and Clatworthy, two manors which William de Moion had contrived to annex to his fief. (fn. 122) But it does not distinguish all the losses which had been inflicted on the Church. In its place, however, an accusing finger points to Stratton-on-the-Fosse and Middlecote, of which Glastonbury had been robbed by the grasping bishop of Coutances. Glastonbury and Athelney, which had suffered most at the hands of the newcomers, group together, at the end of their fiefs, some of their chief losses, from which we learn that Roger 'de Corcelle' had been preceded by his father, who had obtained Limington by giving in exchange five hides which he held of Glastonbury with no power to separate them therefrom. To Athelney's loss of Long Sutton we are indebted for one of those double entries, which sometimes prove so instructive:—

De eadem terra hujus Manerii tenet Rogerius de Corcel ii hidas invito abbate. Duo taini tenebantde æcclesia T.R.E. nec inde poterant separari. Terra est ii (sic) car. quæ ibi sunt in dominio, et vi (sic) acræ prati. Valet 1 sol. Duo homines tenent de Rogerio (91).

Dodeman et Warmund tenent de Rogerio Sutone. Duo taini tenuerunt T.R.E. de æcclesia Adelingi et non poterant ab ea separari et geldabant pro ii hidis. Terra iii (sic) car. In dominio sunt iii (sic) car. . . . Ibi viii (sic) acræ prati, Valet 1 sol. (94 b).

The 'thegnland' so often spoken of in Somerset was held of the church by thegns, one of whom, at the time of the survey, was still holding two hides at Chard of the bishop of Wells.

Of the lay tenants the Count of Mortain, half-brother to the Conqueror, was by far the most important, Mr. Eyton reckoning the assessment of his manors at over 342 hides and their annual value at £346 6s. 4d.

The figures for the fief of the Bishop of Coutances are slightly in excess of both these, but the count could claim, like the bishop, that he held nearly a tenth of the county, whether from the standpoint of assessment or from that of annual value.

A tenant-in-chief in twenty counties, the count can hardly be said to be associated specially with one, unless it was that of Cornwall, in which he practically reigned supreme; for whether he was actually its earl or not, he was virtually the only lay tenant-in-chief within its borders. With Somerset, however, he also had a connection of a special kind, for it was there that he raised his castle of 'Montacute,' (fn. 123) which became, on the breaking up of his fief, the head of a great 'honour' comprising his broad estates in the counties of Somerset and Dorset. Of him Professor Freeman wrote, speaking of the conquest of the West (1068):—

And there was one beyond all these, whose share of the spoils of England was greater than that of any other one man, and whose chiefest and richest rewards lay in the newly conquered lands. Robert, the son of Herlwin and Herleva . . . . now gathered in the richest spoil of all in the forfeiture of countless Englishmen within the Western shires . . . . The lord of the waterfalls (fn. 124) heaped together manor upon manor among the dashing streams of Devonshire and among the hills and islands of Somerset. And one spot came to him by an exchange with an ecclesiastical body, the possession of which, like the possession of Pevensey, seemed to mark him out as the very embodiment of the overthrow of England. The hill of Lutgaresbury, whence came the holy relic which had given England her war-cry, and which had been the object of the life's devotion of her king, now passed into the hands of one who was to wipe out its name and memory. The height, one of the peaked hills which form so marked a feature in the scenery of Somerset, was now crowned by a castle of the new Earl, which, under the French name of Montacute, became at once a badge of the presence of the stranger and an object of the bitterest hatred to the men of the Western lands. (fn. 125)

The 'exchange' by which the count obtained from its owner, the abbot of Athelney, that manor of 'Bishopston' on which he fixed as the site for his strong castle, was the manor of Purse Caundle, Dorset, which was only about half its size and one-third of its value. In the adjoining county of Devon the count, by a similar transaction, had given the Bishop of Exeter two manors 'in exchange for the site of his castle in Cornwall.' (fn. 126) Even the king himself, however, had only obtained a site for his important castle of Corfe (in Dorset) by giving to the abbess of Shaftesbury the advowson of Gillingham in exchange. (fn. 127) The count further improved his home estate by obtaining Tintinhull, close to Montacute, from Glastonbury Abbey, to which he gave 'in exchange' Camerton, south-west of Bath, a manor of not half its value.

But the churches which suffered at the count's hand were by no means always fortunate enough to obtain even an 'exchange.' In addition to the lands of which, as we have seen, he had despoiled St. Swithun of Winchester, (fn. 128) he had laid his hands in addition on Kingstone, not far from his castle, a manor which Glastonbury Abbey had held before the Conquest. Yet these aggressions were as nothing when compared with his 'wide and reckless spoliation of ecclesiastical bodies,' as Mr. Freeman describes it, in Cornwall. Nor was the church his only victim; the very manors of the king himself in Somerset had been encroached on or defrauded of dues by the count or by his vassals. (fn. 129) These vassals, as entered in Domesday, are of more than fleeting importance; for, on the count's fief escheating to the Crown, they became tenantsin-chief, and their tenures were converted, as in the case of the vassals of the count's brother, Odo of Bayeux, into separate baronies.

So ardent a student of feudal genealogy as Mr. Eyton has left one little that is fresh to record about these vassals. (fn. 130) Two of the most interesting are the count's butler, whose name, 'Alvredus,' suggests Breton extraction, and Dru (Drogo) 'de Montagud,' as Domesday styles him, each of whom occurs about a dozen times as a tenant of the count in Somerset. The holdings of the former were of great extent, for the count enfeoffed him not merely in the south-western counties, but in others so far afield as Sussex and Northamptonshire; his holdings were represented, in 1166, by a barony of ten fees which, although mainly in Somerset, was returned under Dorset, and which was held by his direct descendant, 'Richard the son of William.' The holdings of the Domesday 'Drogo' were represented similarly, in 1166, by a barony of ten fees held by his descendant 'Drogo de Monte Acuto.' (fn. 131)

It may, perhaps, be well here to caution the reader against the confusion surrounding the name of Montacute. The baronies held by members of the house which bore it were quite distinct from 'the fee of Montacute,' which was a phrase used alternatively with 'the fee of Mortain' for those baronies formed from the local portion of the great fief held by the Count of Mortain in Domesday. On the Pipe Roll of 1168 we find, under Somerset, such baronies described as 'of the fee of Montacute,' while on that of 1194 we actually find William de Montacute holding his barony 'of the fee of Montacute,' (fn. 132) the double use of 'Montacute' being here well illustrated. In 1166 we find, under Somerset, Richard del Estre making a return of the knights of his 'barony,' at the end of which he explains that 'three of them are of the barony of Montacute in Somerset, and the fourth of the Honour of Berkhampstead in Northamptonshire.' (fn. 133) Here 'the barony of Montacute' is used in its wider sense; it is not merely that which was held by the family of the name, but the whole of that extensive honour of which the castle of Montacute was the head, just as 'the Honour of Berkhampstead' was that portion of the count's fief which had the castle of that name for its caput. It is in this sense that Bernard Pullus returns himself, under Dorset, in the same year as holding one fee 'of the Honour of Montacute,' and the lord of Harptree as holding there a knight's fee 'of the Honour of Montacute of the fee of the Count of Mortain.' (fn. 134) In this latter return we have both phrases used, but in that of Richard son of William, the representative of Alvred the Count of Mortain's butler, his barony is described only as ten fees 'of the fee of Mortain.' (fn. 135)

Four centuries at least after the date of Domesday we still read of a knight's fee (or a small knight's fee) 'of Mortain,' in spite of the early dissolution of the count's immense fief. The reason for this persistence of the name is that the knight's fees that belonged to it enjoyed the peculiar privilege of paying only two-thirds as much scutage as others. This reckoning, however, was not absolutely exact; investigation shows that when the ordinary fee paid one pound (240 pence) or one marc (160 pence) the fee 'of Mortain' paid respectively 150 pence and 100 pence.

Before leaving the name of Montacute we must glance at the question of its origin, which presents some difficulty. In Domesday the castle itself is styled 'Montagud,' a form preserved in its later name. Dru (' Drogo') was one of the count's four tenants who held a hide apiece there in 1086. But although he is styled in the survey 'Drogo de Montagud,' (fn. 136) and Ansger (another of the count's great tenants) Ansger 'de Montagud,' (fn. 137) the latter at least is in no way associated by it with the castle, and it is scarcely conceivable that Dru can have taken his name from his lord's castle, with which moreover he is not shown as more closely connected than his three fellow-tenants on the count's manor of 'Bishopston.' (fn. 138) The family name appearing normally in its Latinized form alone, there is a doubt as to what it was; but on the Pipe Roll of 1167 we find under Somerset the 'Monte Acuto' of the chancellor's roll represented by 'Muntagu.' In the present department of La Manche, in which Mortain is situate, we have a Montaigu and a Montaigu-lesbois, from one or the other of which the name may well be derived. (fn. 139)

The castle itself presents a further difficulty, for although its name is undoubtedly French, like those of Richmond, Rougemont, or Belvoir, the 'peaked hill' from which it rose had its parallel at Mortain itself. (fn. 140) Its name, however, is due, doubtless, to its own situation. Of it Mr. Freeman has finely said:—

From the Peak which had now taken the name of Montacute, the fortress of the stranger Earl looked down like a vulture's nest on the surrounding hills and on the rich valleys at its foot. Of the castle itself not a stone is left. (fn. 141) . . . But the wooded height still covers the fosses which marked the spot which the men of Somerset and Dorset in those days looked on as, above all others, the house of bondage. . . . We read how the West-Saxons of Somerset, Dorset, and the neighbouring districts besieged the castle of Montacute. (fn. 142)

Whether the Count of Mortain obtained his widespread estates in the district in 1067 or 1068 the fact that it was already a stronghold in 1069 shows that he had lost no time in erecting it. And the fact that 'the two porters of Montacute' appear as holding Stert in Babcary tells us that its lord had made the same permanent provision for their support as he had been careful to do at his Sussex castle of Pevensey. The Montacute entry is of further interest as an illustration of the practice of assigning to leading tenants small estates in the immediate vicinity of the lord's castle. Thus we have here Alvred (the butler), Dru (de Montacute), Bretel (de St. Clair) and Donecan, all holding a hide apiece in the manor of Bishopston, in which stood the castle. Bretel (or Britel)—from whom Ashbrittle derives (as Mr. Eyton observes) its name—must have been called from St. Clair, (fn. 143) as was Hubert the count's tenant at Kingstone. His other principal tenants in the county were Mauger de Carteret, Ansger le Breton, who seems to have been also styled Ansger de 'Muntagud,' and who occurs in eight entries; and Robert Fitz Ivo, otherwise the constable, predecessor, we learn from Mr. Eyton, of the Beauchamps of Hatch Beauchamp. Among the count's tenants of less importance one may name William 'de Lestra,' who held of him at Bickenhall and Pointington in Somerset, and Catherston, Durweston, and Corscombe in Dorset, because his descendant Richard del Estre appears as a Somerset baron in 1166. (fn. 144) There is some confusion about his name, which appears in various forms; but it is clearly derived from Lestre near Valognes. (fn. 145)

Next in importance to the fief of the Count of Mortain were those of Roger 'de Corcelle,' Roger Arundel, Walter 'de Dowai,' and William de 'Moion.' The first of these, whose holdings cover more than five columns of Domesday, held so largely as a tenant-in-chief or under-tenant in the county, that ' there were not,' in Mr. Eyton's words, 'six Hundreds in Somerset of any capacity, in which this ubiquitous and omnivorous Feudalist had not some interest.' The persistent endeavour to make him the founder of the house of Churchill was very properly rejected by Collinson and, after him, by Eyton, (fn. 146) but I am prepared to go further. Although the name appears in Domesday as 'Corcelle' or 'Curcelle' normally, and only once (fo. 72b) as 'Corcelles,' it is found regularly enough in the next century as 'Corcelles' or 'Curcelles,' and I have elsewhere (fn. 147) derived it from Courseulles in the present 'Calvados,' north-north-west of Caen, on the coast. Roger is remarkable not only for the number of his Somerset tenures in capite, but also for that of the manors he held as an under-tenant, especially if, as Eyton asserts, he was the Roger 'Witen' who held of the Bishop of Coutances. The devolution of is wide estates remains, unfortunately, as subject to doubt as when Eyton investigated the problem. (fn. 148) Here one can say no more than that the baronial Malets succeeded him at Curry (Mallet) and in the bulk of his barony and his tenancies, but apparently under a fresh grant after a forfeiture of his fief, rather than by inheritance and descent.

The extensive barony of Walter (or Walscin) de Douai presents similar difficulty as to its devolution. Broadly speaking, however, it is found in the next century divided between the 'Honour of Castle Cary' (Somerset), of which his castle there was the head, and the ' Honour of Bampton,' which Devonshire manor was the head of this portion. (fn. 149) The former came into the hands of the Lovels, the latter into those of the Paynels.

Roger Arundel's fief, of thirty knights' fees, had passed to Gerbert de Percy in 1166 (fn. 150) and was subsequently divided equally between the families of Newburgh and Fitz-Payn. (fn. 151) Alone among these great Domesday barons William de 'Moion ' was destined to transmit his extensive fief, of which Dunster was the head, (fn. 152) to descendants of his own name until the line ended in heiresses about three centuries after the great survey. (fn. 153) Sheriff of the shire at the time of that survey, William derived the name, which became anglicized as Mohun, from Moyon (now in the ' Manche'), which lay some seven miles south of St. Lo. (fn. 154)

Neither of the two earls who were tenants-in-chief in Somerset had extensive estates in the county. Earl Hugh of Chester, as Mr. Eyton explained, had for his only vassal one of his Cheshire barons, William Malbanc; Count Eustace of Boulogne's chief tenant was Alvred of 'Merleberg,' lord of Ewias (Harold). This Alvred had for his predecessor, in Somerset as elsewhere, a certain Carle, whom he had succeeded at Chelwood, the only estate in the county that he held in chief; in accordance with a common Domesday practice he held the rest of Chelwood as a tenant of Count Eustace. Kingweston, held by the Countess of Boulogne, was subsequently bestowed by her on the Cluniac priory of Bermondsey. (fn. 155)

Next in importance to the lay tenants dealt with above were William de Eu, Turstin Fitz Rou and Serlo de Burci, who held between 50 and 60 hides apiece. William's predecessor in all but one of his Somerset manors was that great Wiltshire thegn, Alestan of Boscombe, whose estates, scattered over England, he obtained, as Mr. Eyton has shown, after their intermediate tenure by Ralf de Limesi. They were forfeited, however, on his fall, not long afterwards. Turstin is alleged to have been identical with the bearer of the Norman ' gonfanon' at the battle of Hastings. (fn. 156) The devolution of his fief eluded even the researches of such a master of feudal history as Mr. Eyton proved himself to be; but it is now known that it must have been forfeited and then regranted to Winebaud de Ballon, a baron from Maine, who took his place as a lord of the March, and the heiress of whose house brought his estates in Somerset, as in other parts, to a branch of the line of Newmarch. (fn. 157) Serlo, who derived his name from Burcy near Vire (in the 'Calvados'), was father-in-law of William de Falaise, another Somerset baron. His barony, as Mr. Eyton has explained, is afterwards found in possession of the Martin family, and was usually known, from its chief manor, as the barony of Blagdon (fn. 158); but considerable obscurity surrounds its devolution, as well as that of the fees which, like Roger de Courseulles, he held under Glastonbury Abbey.

The only other lay barons deserving of special notice are William de Falaise, whose chief holdings were in Devon, and whose daughter Geva brought Stoke to the Courcys; Alvred 'de Hispania,' (fn. 159) whose barony was shown by Mr. Eyton to be identical with that which was returned, under Somerset, in 1166 as held by Philip de Columbers; Edward de Salisbury (sheriff of Wilts), ancestor of the first Earls of Salisbury; Baldwin of Exeter, lord of Oakhampton and sheriff of Devon; Robert Fitz Gerold, whose estates passed to his nephew; Matthew de Mortagne, whose estates ranged through six counties from Somerset to Essex; and Ralf Paynel, a Yorkshire baron, who owed his Somerset manors to his obtaining the lands of Merlesweyn, a great thegn, the chief seat of whose power was in Lincolnshire, but who was a holder of land also in Gloucestershire, Devon and Somerset.

The last two pages of the survey of the county (fols. 98b, 99) present peculiar difficulties, difficulties duly recognized, but scarcely explained by Mr. Eyton. The normal practice in Domesday Book is to class together, where they exist, at the end of each county (1) the king's serjeants, who held their land by serjeanty; (2) the king's thegns, that is the few Englishmen who had either been allowed to hold some pitiful estate in the place of their former possessions, or found favour in Norman eyes and received a manor or more from the spoils. These 'thegns' are a class of peculiar interest as Englishmen, and are fairly numerous in the counties of Devon, Somerset, Dorset, Wiltshire and Hants. There was, however, in addition to these recognized classes, a third and limited class for which the Exchequer clerks had no definite name. These are they who appear at the end of the Somerset Survey as 'certain others.' Mr. Eyton's statement of the difficulty is as follows:—

The Exchequer Domesday is by no means free from the blemish of misarrangement of Fiefs. It commences a survey of Humphrey Chamberlain's estates with a proper title, viz. ' Terra Hunfridi,' therewith treating his fief as the last of Somerset baronies, yet still as a barony. It enters two manors proper to this category, and then, without giving any fresh title or note of interruption, lapses into a list of the estates of King's Serjeants. Then it gives a chapter entitled Terra Tainorum Regis, which proves, on examination, to be of the lands of the Angli Taini only, who are thus placed in an unusual precedency. (fn. 160)

Lastly, and out of all sequence and propriety, it gives a chapter entitled 'Item Hunfridi Terra et Quorundam aliorum.' Here it resumes its interrupted survey of Humphrey Chamberlain's estates, and registers the particulars of three manors. Then again, without any fresh title or note of interruption, it lapses into the estates of the Franci Taini, as they are called in the Domesday of other Counties.

The true arrangement of the Exchequer text would have been to have inserted this last chapter immediately after the two manors entitled Terra Humfridi, and then, when Humphrey's manors had all five been registered, to have inserted a new title, such as Terræ Francorum Tainorum. (fn. 161)

Franci Tegni.—The perverse arrangement of both editions of the Somerset Domesday has led to the suppression of this Title in both Records. Nevertheless, collating the two, we are enabled to make out a Somerset list of those who in other counties would have come under this category . . . the 'French Thanes.' (fn. 162)

In no county is found a category of ' French Thanes'; indeed such a phrase might be described as a contradiction in terms. (fn. 163) One can only suppose that the writer must have had a vague recollection of the phrase 'Hugo de Luri et alii Franci' in the heading to the survey of Dorset, which is contrasted with ' Gudmund et alii taini' in the same heading (fo. 75). On comparing the arrangement under other counties, it seems tolerably clear that Domesday had in this matter no fixed practice, and that when it grouped, after tenants-in-chief, several 'Frenchmen' who held a manor or two apiece, it did not thereby imply that either in status or in tenure they differed from those who were entered, under separate headings, before them. In Somerset, for instance, the survey closes with the two-hide holding of Ansger 'de Montagud' at Preston Plucknett, Ansger being here grouped with 'certain others.' But under Devon this same Ansger is entered among the great tenantsin-chief (fo. 116), merely, it would seem, because he there held seven different estates. So too Ralf 'de Berchelai,' who figures in the same group, under Somerset, where he held half a hide, is placed, with a separate heading, as a tenant-in-chief, in Gloucestershire (fo. 168), where his two estates were assessed at 5½ hides. Keeping still to neighbouring counties, we have seen that in Dorset Hugh de Luri heads a group similar to that at the end of the survey of Somerset. Yet in Northamptonshire, where his solitary holding was a very much smaller one, we find him entered separately among the great barons (fo. 224b).

We have here then a further instance of Domesday's capricious practice. (fn. 164) I cannot agree with Mr. Eyton in treating the group of Normans found both in Somerset and in Dorset as differing in status or tenure from their fellows, though I entirely concur with his view that 'the true arrangement of the Exchequer text would have been to have inserted the Somerset group in continuation of Humphrey the chamberlain's land.' But I cannot admit that the 'confusion probably arose in the indeterminate and complex position of Humphrey Chamberlain himself . . . as his very name suggests, he was or had been a king's sergeant, and in that quality the clerks gave him a third post at the head of the Servientes Regis.' (fn. 165) It was not Humphrey, but Robert 'de Odburvile,' who headed the king's serjeants (fn. 166); and Humphrey (whose brother, Aiulf the chamberlain, was sheriff of the adjoining county of Dorset) was a tenant-in-chief in no fewer than nine counties. Three out of five of his Somerset estates had been added, we read, to Brictric's lands, which gives us a clue as to how he came by them. For the bulk of the lands held by that great western thegn, Brihtric the son of Ælfgar, had been given to Queen Matilda, with whom legend has associated his name, and various entries in Domesday show that Humphrey had been her chamberlain and had received estates from her. (fn. 167) Among the small tenures in capite which follow those of Humphrey is Tadwick, held by William 'Hosed' and by Ralf brother of Roger de Berkeley. (fn. 168) The former is identified by the Exon Domesday as the tenant of Serlo de Burci at Ridge Hill, of the Bishop of Coutances at Pitcombe, and of Bath Abbey at Charlcombe. Next in order is 'Hugolin' the interpreter, who held three estates, and who is probably identical with the Hugh 'latinarius,' who had held an estate in the New Forest (fo. 50b). He was an under-tenant of Bath Abbey at Bathampton, and had secured for himself a valuable house in Bath. In the geld-roll he appears under the noteworthy alias of 'Hugolinus legatus,' an alias which suggests that the 'Richard' who appears with him in this group as holding a hide in Road (where he was also an under-tenant of the Bishop of Coutances), and who appears in the Exon Domesday as Richard 'Interpres,' was the Richard 'legatus' who held, as a tenant-in-chief, the valuable manor of Tormarton in Gloucestershire (fo. 168b). The question as to his title at Road is obscurely stated in Domesday, and I have elsewhere suggested that the 'tenuit' of the Exchequer record is a slip of the scribe for 'emit,' the word in the Exon Domesday, which makes sense of the passage. (fn. 169) With Ansger and Dru 'de Montagud' I have already dealt. There remain only in the group 'Schelin,' who gave his name to Shillingstone, his Dorset manor; Hugh (de Valletort) who held at Fodington of the Count of Mortain as well as of the king, and who probably derived his name from Vautort (Mayenne) some 25 miles south of Mortain, Odo the Fleming, and 'Eldred' (the Aldret' of the geld-roll) who, at Brockley and Crandon, still held the lands which had been his before the Conquest. Even Mr. Eyton, who studied so long and to such good purpose both the Norman and the English holders of lands, did not attempt to identify Eldred; but I believe him to have been identical with a man whose case was as interesting as it was exceptional. If so, he was the Eldred or Edred who appears in Devon to the west as holding three estates, two at least of which had been his before the Conquest (fo. 118), and the 'Aldred' who, in Wiltshire to the east, held five, two of which had been similarly held by himself before (fo. 73b), and finally the Eldred 'brother of Odo of Winchester,' who held land in Hampshire which he had similarly contrived to retain. (fn. 170) In all three of these counties he appears among the native thegns, as is shown by his name to be his right place; and I conclude that only an obvious error has placed him in Somerset, among Normans.

Next in order are the king's serjeants. Robert de 'Odburvile,' who heads the list, was, says Mr. Eyton, 'a king's forester, and to foresters of King Edward he succeeded in some estates.' (fn. 171) I have not found the evidence for his office, and in only one of his five estates is he recorded to have been preceded by foresters. He is followed by John the usher, who had six estates in this county and two in Wiltshire. Next to him are Ansger 'Fower' (the hearth-keeper (fn. 172) ) with four small estates and Ansger the cook, who, in addition to his own holding at Lilstock, had obtained possession of land belonging to the royal manor of Martock. This cook appears in Wiltshire also among the king's tenants by serjeanty, and is doubtless identical with the man of the name who had a small estate at Aveley far away in Essex. The name of Anschitil the parker speaks for itself. The appearance of one who must have been an Englishman among these foreign serjeants is noteworthy; Edmund the son of 'Pagen,' who held three estates, is entered in Hampshire as an English thegn holding land in the New Forest district, with which he and his father seem to have been associated; but the strange thing is that he is also found in Suffolk, where he had succeeded to a substantial manor held by his father 'Pagan' before the Conquest. The last tenure by serjeanty is that of the wife or widow of Manasses the cook, who himself appears as an under-tenant at Stalbridge in Dorset.

Of the 'king's thegns,' that is the Englishmen who, in 1086, were still allowed to hold land, (fn. 173) Harding son of 'Elnod,' or 'Alnod,' was 'clearly the greatest.' (fn. 174) He has been the subject of much discussion, rather because he was the probable ancestor of the historic race of Berkeley than because he was certainly the founder of the Somerset house of Meriet. (fn. 175) In the geld-roll of Crewkerne Hundred (1084) he is styled 'Hardinus de Meriet,' taking his name from his chief manor as did his descendants. Mr. Freeman, who was much interested in the subject, established the identity of this Harding, son of 'Elnod' or 'Alnod,' with the 'Heardinc' or 'Hierdinge' son of 'Eadnoth,' who is found in Anglo-Saxon documents, and with the 'Herdingus' son of 'Ednod,' who was alive when William of Malmesbury wrote, (fn. 176) and whose father, that historian tells us, fell in repelling the descent on Somerset by Harold's sons in 1068. This identifies the latter with the 'Eadnoth Stallere' of the Chronicle, the 'Eadnothus Haroldi Regis stallarius' of Florence, who commanded, they tell us, William's troops on that occasion. The Domesday holder of Meriet is also, clearly, the Harding 'filius Elnodi' who acted as a justice itinerant for Devon and Cornwall in 1096. (fn. 177)

The great estates of Eadnoth did not pass to his son; they are found in Somerset, as in other counties, in the hands of Hugh, Earl of Chester. A Berkshire entry (58b) enables us to prove the identity by styling the earl's predecessor 'Ednod stalre.' The succession of the Count of Mortain to a manor which had been Ordulf's, and to three (fn. 178) which Edmeratorius had held, is of interest because he had obtained the whole estates of both in Devonshire, as Mr. Eyton points out. The second of these great thegns, who is only styled 'Edmer' in the Exchequer Domesday, is, in the Exeter Book, 'Edmeratorius,' which enables us to identify him with the Count's predecessor in his great lordship of Berkhampstead and other Hertfordshire lands. The Exeter book applies the term Honor to the vast holdings of these two thegns.

Restrictions of space and the usual difficulties attending their identification preclude any detailed description of those English thegns whom the Normans had dispossessed; but some mention should be made of Mærleswegen the sheriff. Although Harold, it is said, had left him, after the battle of Stamford Bridge, in charge of the North, he is found witnessing as sheriff (i.e. of Lincolnshire) an early charter of William's. (fn. 179) But he must soon have gone into exile, for we find him among those who landed in the Humber in 1069 to join in the siege and storm of York. Although the seat of his power was in Lincolnshire and Yorkshire, another portion of his great possessions lay in Somerset, Devon and Cornwall. Apart from his distinctive name we have clear proof of identity in the fact that the whole of the estates of Ralf Paynel in Somerset and Devon had previously belonged to Mærleswegen, and that Ralf was also Mærleswegen's successor in his mansio at Lincoln itself and in his Lincolnshire lands. Of this great noble's Somerset estates four manors, which are not found in the hands of Ralf Paynel, were secured by Walter de Douai, while the only other had been annexed to the king's manor of South Petherton.

In the absence of a distinctive name or of a definite succession the identification of English thegns is apt to be rash work. On the one hand, the Domesday scribes were loose in the spelling of their names; on the other, the student may confuse names which are wholly distinct. There can, virtually, for instance be no question that the Alverd, Aelvert, or Ailvert, who was Roger Arundel's predecessor in four Somerset manors was also the Alvert, Ailvert, or Aielvert who preceded him in three Dorset ones. Nor need we hesitate to see in him the 'Æilferth Minister' who was a witness, as Mr. Eyton points out, to Edward the Confessor's grant of Ashwick to Wulfwold, abbot of Bath in 1061. (fn. 180) But when that writer treats the 'Alward,' 'Elward,' and 'Olward' of Domesday as variants of his name, nay, when he further identifies him with Ælfred (Aluredus) sheriff of Dorset and lord of Lulworth before the Conquest, (fn. 181) it is time to call a halt. Again, when he suggests that 'Aluredus', a 'dapifer' of Queen Edith, who had held of her at Tiverton (Twerton), may have been the 'Alvered or Alverd' who held Bathwick, as a king's thegn in 1086 (and, similarly, before the Conquest), and also with 'Alured,' who held of Count Eustace, (fn. 182) it becomes necessary to point out that this last was no other than the well known Alvred (Aluredus) of Marlborough, a tenant-in-chief in six counties.

But with Tofig the sheriff at least we stand on sure ground. Not only before the Conquest, but even after, he acted as sheriff of Somerset. Mr. Freeman made him sheriff 'between 1061 and 1066,' (fn. 183) but he also, as Mr. Eyton points out, was the 'Tong (i.e. Tovig) Minister' who appears among the witnesses to the charter by which King William, in the summer of 1068, restored Banwell to the church of Wells. (fn. 184) And, one may add, he is actually the sheriff to whom King William addresses the charter by which he restored Winsham to Giso and the church of Wells. (fn. 185) His lands in Somerset were at Keynsham, (fn. 186) Buckland St. Mary, Bradon, Capland, Freshford, Discove in Bruton, Berkley and Lopen. Most of them are found at the time of the Survey in the hands of Harding of Merriott, the son of Eadnoth the Staller.

Tofig's Norman successor, at the time of the Domesday Survey, was that William de Moion whose seat and successors are the subject of a learned monograph. (fn. 187) We are here chiefly concerned with him as sheriff. Although the Survey gives us but indirect information on the county's administration and finance, I have here shown that William— in direct contradiction of Mr. Eyton's conclusion—was farming the royal demesne of Somerset in 1086. (fn. 188) He was also receiving and paying to the Crown—no earl having been appointed—the rent of the 'comital manors.' (fn. 189)

Dealing as we are here with administration and finance, one ought to say something of a difficult and obscure question, that of 'the third penny' (tertius denarius). In the adjoining county of Dorset we read under Puddletown, a comital manor which Harold had held: 'Huic etiam manerio Piretone adjacet tercius denarius de tota scira Dorsete.' Here we seem to recognize what is afterwards familiar as 'the earl's third penny,' that is the third part of the profit from the pleas (placita) of the shire. (fn. 190) But in Somerset the entries are perplexing. Under (King's) Brompton—which had been held by Harold's mother Ghida, and which heads the Comital Manors—we read that it no longer receives, as it used to do, the third penny of Milverton, a manor which had reverted to the king on Queen Edith's death. A little lower down, under Old Cleeve, which had been held by Harold, we have the puzzling entry: 'Huic Manerio adjacuit tercius denarius de Burgherist (et de) Caretone et Willetone et Cantetone et Nortpereth.' We distinguish the names of two units of royal demesne (1) Carhampton, Williton, and Cannington; (2) North Petherton; but what are we to make of 'Burgherist'? We have seen, under Taunton, that this term, as 'burgeristh,' remains as yet obscure to us; nor are scholars likely to accept Mr. Eyton's ingenuous view that it was 'burglary,' and connoted the Placita Coronæ. (fn. 191)

We must, I think, distinguish carefully the 'third penny' spoken of above and associated with rural manors (fn. 192) from that which Domesday, as I read it, associates only with the local boroughs. Nor do they differ only in the character of the places from which the payments are due; they differ also in that the latter is not due to a comital manor. Mr. Eyton, it is true, reckons as appurtenant to Henstridge the third penny of six towns, (fn. 193) a due which the record enters by itself; but in spite of this he, quite justly, pointed out reasons for doubting whether it was so appurtenant. The fact that, as we shall now see, it was distinctively burghal is further ground for rejecting that view.

In boroughs, or—to speak with more precision—in towns of a burghal character, Somerset appears in Domesday exceptionally rich. But the line of division between the manor and the nucleus of a rising town is here so difficult to draw that students of the Survey may well differ in the classifications they adopt. Mr. Eyton, Professor Maitland, (fn. 194) and Mr. Ballard (fn. 195) have each their own.

We may well start from that simple list which Domesday itself suggests. Standing by itself is the 'borough' of Bath, which was in the king's hands in 1086, and of which Edward the sheriff (of Wilts) was paying him 'the third penny.' Next to it would come the group of six places of which Domesday shows us William the sheriff (of Somerset) paying 'the third penny' to the king ; these are Ilchester, Milborne Port, Bruton, Langport, Axbridge and Frome. All the six have ' burgesses' in Domesday (fn. 196) except Frome. which, however, has a market. Lastly we have Taunton, where all the burgesses belonged to its sole lord, the bishop of Winchester.

Mr. Eyton practically adopted this classification, but more recent students appear to have rejected it. Professor Maitland, after mentioning Ilchester and Bath, proceeds :—

Axbridge, Langport, and Milborne seem to be boroughs. Axbridge and Langport occur in that list of ancient fortresses which we have called the Burghal Hidage. Wells was an episcopal, Somerton a royal manor; we have no reason for calling either of them a borough. (fn. 197)

Mr. Ballard constructs a definite list, consisting of six 'composite' boroughs—Axbridge, Bath, Bruton, Ilchester, Milborne, Langport, and one simple borough Taunton. (fn. 198) To the mint as a test of burghal rank he has given special attention, and he points out that while Domesday speaks of mints at Bath and Taunton, (fn. 199) there are coins from pre-Conquest mints at Bath, Bruton, Ilchester and Langport. (fn. 200) But neither at Axbridge nor at Milborne is there any trace of a mint, while at Somerton, which was not a borough, there is. (fn. 201) On the whole his own evidence would seem to 'prove,' not 'that a mint was a necessary factor in the making of a borough,' but that it was not; for Milborne Port was a market town, as its name implies, and ranked, in the number of its burgesses, easily third among the towns of Somerset. Frome was probably on the border line between manor and market town.

A feature which has struck all students is that, while in Dorset to the south Domesday places four towns all by themselves at the head of the Survey, it does not accord this distinction to any town in Somerset. Mr. Eyton, in his chapter dealing with 'the royal burghs of Somerset,' wrote :—

The Burghs of Dorset were, doubtless, Royal Burghs; but they were distinct, and they were kept distinct in Domesday from the estates of Ancient Crown-Demesne. . . . But in Somerset, where every Burgh, except Bath, was but a mere appendage of some estate of Ancient Demesne, no such distinction could be made, and certainly none was attempted by the Domesday Commissioners. (fn. 202)

Again, Professor Maitland writes :—

If we compare the first page of the survey of Somerset with the first pages that are devoted to its two neighbours, Dorset and Devon, we shall probably come to the conclusion that the compilers of the book scrupled to put any Somerset vill on a par with Exeter, Dorchester, Bridport, Wareham and Shaftesbury. (fn. 203)

Mr. Ballard similarly notes that at the head of the Somerset survey there is only 'a blank space,' (fn. 204) which contrasts with Dorset. It is no doubt the case that the infant towns of Somerset were smaller than those of Dorset. With the exception of Bath, Ilchester was much the largest, having 108 burgesses; and these were fewer than those of Bridport, the smallest of the Dorset towns. Yet Malmesbury in the neighbouring county of Wiltshire is assigned the distinctive position at the head, although its burgesses were considerably fewer than those of Ilchester. We can also, I think, detect in Domesday the beginning of a Somerset town of which I have not yet spoken. This is Yeovil, of which we read that to William d'Eu's manor of Yeovil there had been 'added' twenty-two messuages (masuræ (fn. 205) ), formerly held pariter by the men who dwelt in them. Obscure as is the previous status, we have clearly here the nucleus of a town which is unmentioned by the writers I have named, but of which the twenty-two householders compare not unworthily with the thirty-four burgesses of Langport, the thirty-two of Axbridge, or the seventeen of Bruton. It was not size alone that gave a town its importance. Nor indeed is size the test that Professor Maitland himself selects. That test he finds in 'tenurial heterogeneity,' that is to say in the fact that the burgesses do not all hold of one lord. (fn. 206) But when we apply this test to the Somerset and Dorset boroughs we find that this 'heterogeneity' is wholly absent at Bridport, a borough which Domesday emphatically places 'above the line,' while Bath, which stands below it, is 'a good example,' as Mr. Ballard observes, (fn. 207) of 'tenurial heterogeneity.' Other Somerset towns, as he says, display this feature, although obscurely entered as part of the Terra Regis.

This is perhaps the most important point about the Somerset boroughs, for it suggests that the compilers of Domesday Book—with that lack of system and love of variety which seems to us so strange— may after all have had no definite practice as to the placing of boroughs 'above' or 'below' the line. (fn. 208) Mr. Eyton found the distinction in the fact that the Somerset boroughs were 'interned' in groups of royal demesne, while the Dorset ones occupied a wholly different and independent position. But I have maintained that this view was wholly mistaken, and that Dorchester and Bridport, for instance, formed part, much as did Ilchester and Milborne Port, of a royal demesne group. (fn. 209)

The most important element, the centre of life, in these little towns was the market. Its value is not always recorded, but that of Ilchester, 'with its appurtenances,' was worth £11 a year, that of Frome £2 6s. 8d., and that of Taunton £2 10s. Crewkerne, although apparently only a rural manor, had a market worth £4 a year, while that of Milverton was worth ten shillings. Ilminster market brought in twenty shillings. The mention of eight smiths at Glastonbury hints at a small iron industry, of which plough-shares, probably, would be the chief product. But ironstone would not be found in the lias of the Polden hills. (fn. 210) Nor can we connect with ironstone districts the curious dues in unwrought iron entered in the county. In the south, near Crewkerne, we have mention at Seaborough, Cricket St. Thomas, and Bickenhall, of a due from each free man of a 'bloom' of iron (fn. 211) ; at Alford, near Castle Cary, eight such 'blooms' are receivable from 'the villeins'; and at (White) Staunton, on the Devonshire border, four 'blooms' are due in respect of 50 acres of pasture. Lexworthy (in Enmore) appears to be the only place where Domesday speaks of plumbæ ferri; its three mills rendered between them four of these plumbæ

In Somerset, as in other counties at this early period, agriculture was the dominant occupation; the field and the meadow, the woodland and the stream, were the paramount sources of wealth.

Like that volume of Domesday Book which deals with the three Eastern counties, the Exeter book records the number of the live stock on the manors, which, although comprised, it is believed, in the original returns throughout the country, are omitted in the digest thereof in the larger volume of Domesday. The abstracts in this larger volume are concerned only with the plough-oxen, the teams of which were accounted the most important element in the agricultural system. Mr. Eyton, in his short, but very useful section on 'farming stock,' (fn. 212) tells us that the plough-teams in the county were notably fewer than the plough-lands :—

The numerical inadequacy of the teams-in-stock to the plough-lands, specified to be available, was more constant in Somerset than in Dorset.

But the subject of this relation and of its actual meaning is still for Domesday students one of notorious difficulty. (fn. 213)

In a table constructed to show the proportion of plough-teams to other live stock for those counties in which we have material for such comparison, Professor Maitland selects these figures for Somerset:

But partial figures such as these prove nothing. Even the figures for a whole county may be in practice misleading, for the proportion would naturally vary in different districts. Swine we should expect in the woodlands; sheep are associated with marshlands; and cows we look for in the meadows down in the river valleys.

Very interesting in the records of the Survey are the wild (silvaticæ) or unbroken (indomitæ) brood-mares, a class of stock not mentioned above, though met with frequently enough in Somerset.

The valuable tables of the Hundreds so laboriously constructed by Mr. Eyton tempt us to see whether it is possible to localize, through their mention, the industry of horse-breeding in the county. Certainly in and about the Hundred of Bempstone—that is between the Mendips and the mouth of the Parret—these mares are often mentioned; we read of them at Wedmore (6), Tarnock in Badgworth (8), Burnham and Highbridge (12), Huntspill (6 + 14), and we also find them at Stoke Rodney (20). On the other hand, it seems to me that Mr. Eyton is right in holding that horse-breeding depended on the barons who held the fief. 'The bishop of Coutances,' he writes, 'the Comte of Moretain, and William of Moione were the chiefs who seem most to have encouraged this class of stock.' Analysis shows that the bishop is credited with about 130 mares, of which he kept 60 at Long Ashton alone. The Count's chief stud was at Cloford, near Frome, where he had 38, besides 34 elsewhere. That of William de Moiun was at Cutcombe, south-west of Dunster, where his mares were 39 in addition to 7 in Luccombe and 22 at Brewham. The bishop of Wells and the abbot of Glastonbury also did a little horse-breeding.

Milch-cows, as Mr. Eyton says, are not so often mentioned 'as to indicate that the county was to become famous for its dairy farms.' At Cheddar, it is true, we read of one cow only, but at Winscombe, north of the hills, there seem to have been sixteen. It is not, however, at all certain, in my opinion, that cows were not sometimes among the 'animalia,' which we can only translate as cattle. The entries of one cow are, I think, significant, as suggesting that it was kept for the use of the lord of the manor. Of cheese there is hardly any mention, nor were cows needed for it then, as ewes' milk was largely employed. (fn. 214)

Horses—as distinct from brood mares—are divided, Mr. Eyton writes, into two classes, caballi and roncini. Of the former, the riding horses, from whose name is derived the Spanish caballero (cavalier), 'Alured of Spain,' (fn. 215) he asserts, 'had a stock.' The reference is, I find, to the abbot of Glastonbury's fief, on which, at Woolavington, Alured had 'xiii caballos' inter alia. But the phrase is interlined, and over it is written the word 'roncinos,' (fn. 216) so that even this occurrence is open to question. The roncini—Chaucer's 'rounceys'—were most probably pack-horses, specially useful where the roads were few or bad.

The ewe-sheep, says Mr. Eyton, were 'common to nearly all Somerset estates,' and the she-goats 'more frequent in, but not peculiar to, the hill districts.' The large number of she-goats recorded at the Survey may come as a surprise to those who forget the important part they still play in lands where cows' milk is scarce or unprocurable. (fn. 217) The kid would be valued for its skin and its flesh, but milk would be the product of the goat.

The swine were fattened in the woodlands on the mast. In Somerset, as in Devon, the swineherds are a feature, the great majority of those entered in Domesday Book being found in these counties. To the king's manor of North Petherton no fewer than twenty swineherds contributed between them five pounds; at King's Brompton fifteen of them paid thirty-five shillings; and at East Coker ten swine were payable by the swineherd. Again, 'in William de Moione's forestal manor of Cutcombe six Porcarii paid an annual rent of thirty-one hogs.' (fn. 218) It was ingeniously suggested by Mr. Eyton that, comparing the payment at North Petherton with the render in kind of five swine, to the King, by the Bruton swineherd, the commuted value of a fat hog was one shilling. But this was only a 'surmise,' and cannot be taken for more.

The moorlands, which have always formed so prominent a feature of Somerset, and of which so many traces are preserved in its local nomenclature were, of course, far more extensive at the time of the great survey. But as they were of little or no value, Domesday was not much concerned with these moræ, as it terms them. Mr. Eyton points out, for instance, that at Wedmore the Exeter book speaks of moors which are worth nothing, and which Domesday Book ignores. In the north-east, 'moors' are mentioned in districts liable to inundation, at Yatton and Weston-in-Gordano. We read of them also to the south of Bridgwater, at Newton Huntworth and Edgeborough, to its west at Spaxton, and again, north of this, at Fiddington. A considerable moor which is entered at Wells would probably, as at Wedmore, be undrained land; but that which is spoken of at Milborne Port in the south-east of the county may have been of a different character.

Meadow-land (pratum) that was mown for hay was a possession of some value, but pasture (pastura) occupied an indeterminate position between profitable and worthless land. The water-mills and (weir) 'fisheries' in the streams need not detain us, save for the solitary mention of ten fishermen (piscatores) at Glastonbury. The three vineyards, in the heart of the county at Glastonbury, Pamborough, and Meare, were doubtless introduced by Thurstan, the innovating abbot, together with his Norman archers. But there were also seven acres of vineyard at North Curry and a small one at Muchelney on the demesne. At Langport there is mention of a garden (ortus) from which the singular due was 50 eels, a render more suggestive of a 'fishery.'

Of the agricultural classes the most noteworthy are 'the small but interesting class of coliberti,' (fn. 219) otherwise styled in Domesday buri or burs. Professor Maitland deemed them 'distinctly superior to the servi, but distinctly inferior to the villeins, bordiers and cottiers.' (fn. 220) They are found in a group of counties extending from Buckinghamshire and from Shropshire to Cornwall, and in Somerset are fairly numerous, 218 being entered, as against 32 in Devon and 33 in Dorset, though in Wiltshire there are 260, and in Gloucestershire 103. (fn. 221) No fewer than 150 of those in Somerset are found on the King's manors, and here, it should be observed, they are systematically entered after the serfs and, like them, as belonging to the lord's demesne. (fn. 222) Indeed, at Frome, where there were no serfs, while there were six coliberts to the three demesne teams, I suspect that they took the place of the serfs in their normal proportion of two to each demesne team. Another interesting class are the holders by a money rent, the seven gabulatores who pay between them, at Cheddar, seventeen shillings. For this would seem to be their only mention by that name in Domesday.

We may now turn to the remaining point deserving of comment in this county.

The Domesday Survey brought to light in various parts of the country wholesale encroachments on the king's manors by the great barons and their vassals. His own reeves, at times, proved but unfaithful stewards, though they sometimes added to the king's demesno the small estates of private men, probably hoping to 'farm,' thereby, a larger area without having to pay a larger sum to the Crown.

Thus from Somerton, which stands at the head of the Terra Regis, half a hide had been filched by Alvred 'de Hispania,' while the estates (terræ) of three thegns had added five and a half hides to the manor. It is one more illustration of the lack of system in Domesday that, although Alvred's encroachment on Somerton is recorded under that royal manor, his similar diminution of (North) Petherton by the lands he had obtained at Wolmersdon and West Bower is recorded not under that royal manor, but under his own fief. In each case the land had been secured by 'Alwi,' his predecessor in title; and in that of Wolmersdon, at least, Domesday records that the (king's) reeve had lent (prestitit) the land to Alwi in King Edward's time, reminding us that the king's interests might thus suffer loss even before the Conquest.

The king was even robbed of lands and dues by his own greedy brother, the Count of Mortain. From King's Brompton he had annexed a hide at Torrels Preston, 'quæ fuit de dominica firma.' Here we have a good instance of these double entries in Domesday, which are often so instructive. (fn. 223)

Rex tenet Brunetone. Ghida tenuit T.R.E. . . . De hoc manerio tenet Comes Morit' i. hidam in Prestetune, quæ fuit de dominica firma T.R.E. Terra est iiii. car. Ibi sunt ii. car. Valet xl. sol. et valuit (86b).

Robertus tenet de Comite in Prestitone i. hidam. Hanc tenuit Heraldus Comes. Terra est iiii. car. In dominio est dim. car. cum i. servo et vi. vill. et ii. bord. cum ii. car. . . . Valet et valuit xxx. sol. Hæc terra jacuit in Burnetone M[anerio] regis cum firma (92).

Now what are we to say of such a case as this? Here are two entries describing the same estate, differing not only in their whole arrangement, but in the valuation past and present, in the number of ploughs upon the land, and in the name of the previous holder. (fn. 224)

It is clear surely that these entries cannot be taken from a common original; and yet how can there have been more originals than one? The answer, I take it, is that the entry found under the Count's fief is derived from the general return, while the entry under Terra Regis is derived from a special and separate return—breve regis?—made for the king's land. This conclusion is strongly supported by the fact that in the only transcripts of original returns which have come down to us, the entries for the king's manors are wanting. (fn. 225)

Let us see if the Exeter Domesday helps us. Under the Count of Mortain's fief we have, as usual, a slightly fuller entry, which expands the name of Robert into Robert son of Ivo, who was the Count's constable. (fn. 226) But in the entry of King's Brompton we have a far more important addition; for we there read that the hide at Preston was held of the Count by Hugh de Valletort ! (fn. 227) Here we have a really flagrant contradiction in addition to those already discovered in the text of the Exchequer Domesday. (fn. 228) If I am right in suggesting two independent returns, all the contradictions become explicable. And this suggestion is remarkably supported by the fact that King's Brompton appears in the Exchequer Domesday as 'Brunetone' under the King's land, and 'Burnetone' under the Count's, while it is similarly in the Exeter Domesday 'Brunetona' under the King's land (pp. 94, 474), and 'Burnetona' under the Count's (pp. 252, 478). This would appear to prove that the name was spelt differently in two independent returns.

The entries relating to the annexation of Eastham by the Count from the king's manor of Crewkerne, formerly Edith's, and its tenure under him by Turstin present no contradiction, but this is not inconsistent with the theory of two returns; and other entries of aggression under 'Terra Regis' are difficult, if not impossible, to discover on the fiefs of the alleged aggressors.

The alteration of the king's manors by the processes of accretion and diminution (fn. 229) find their parallels on the manors in private hands. Thus, for instance, William de Moiun had secured the valuable manor of Brewham, which had been held by Robert Fitz Wimarc the Staller, one of the Confessor's favourites; (fn. 230) but it had suffered diminution by the subtraction of Witham, which Domesday, it should be noted, reckons in one place as two, and in another as three hides, and which, though inalienable from Brewham, had been grabbed by Roger 'de Corcelle.' (fn. 231) On the other hand the Brewham estate had been increased by the addition of three virgates which Almar, an Englishman, had held. All these cases speak of the confusion that accompanied the Conquest and of the need for such a check as must have been supplied by the compilation of the Survey.

The curious customary dues withheld from the king's manors are by no means consistently entered. Under South Petherton, a manor of ancient Crown demesne, we read that Cricket St. Thomas (a sixhide manor), held by Turstin of the Count of Mortain, had failed to pay its annual due of six ewes and lambs and a 'bloom' of iron from each free man since the count obtained that estate. It heads the account of his own fief, but there is nothing there of this liability. So too, Dulverton, one of Harold's manors, which was held by the king, was entitled, we read, to 24 sheep a year from Brushford, which was held from the count by Mauger, who withheld that due 'per comitem.' Under Brushford itself there is no mention of this due. Yet conversely, on the count's fief, the very next manor (also held by Mauger) is entered as owing a customary due to the king's manor of Curry. The same is recorded of six other estates on the same page, the due being always a ewe and lamb for a hide, (fn. 232) save in one instance, where a money payment suggests that the commuted value of the ewe and lamb was sixpence. Here the point is that there is no mention of this due under Curry itself. The same remark applies to the bishop of Salisbury's manor of Seaborough, in the account of which we read that it had been appurtenant to Edith's manor of Crewkerne, now the king's, from which its holder could not separate it, and to which he owed the customary due of twelve sheep with as many lambs, and from each free man a 'bloom' of iron. (fn. 233) But under Crewkerne itself there is no mention of this.

So, again, Ralf de Limesi is charged, under his own fief, with withholding 12 sheep a year due from his own manor of Allerford (in Selworthy) to the King's manor of Carhampton, though under the latter manor we hear nothing of that due. Exactly the same remark applies to Ralf de Pomerei's manor of Oare, lying near it in the wild north-western corner of the county. It had similarly been held by Edric, and owed the same due, which was similarly withheld.

As against these withdrawals of dues we find, under Terra Regis, the statement that there has been added to Williton an annual due from a manor of Alvred (de Hispania), which was not due in King Edward's day. One would expect to find complaint of this under that manor itself, but we find there nothing.

Apart from the record of the Taunton franchise there is singularly little in Somerset to illustrate legal antiquities; but the phrase 'lord of the manor' (dominus manerii) occurs under Congresbury and Henstridge among Harold's lands. There are also two notable entries—one of them marginal—of estates, held by Robert de 'Odburvile.' The first of these relates to Withypoole, of which we read that Robert used to contribute in respect of it twenty shillings towards the revenue derived from the king's manor of Winsford, but that now it has been proved to be 'thegnland' (diratiocinata est in tainland) which had been held by Harold's brother (Tostig). Under Winsford itself we find another entry of this estate, which is there admitted to have been an addition (addita), but of which the retrocession and Robert's tenure are alike ignored ! Moreover, its three former holders are described as 'thegns,' not as 'foresters.' It is, further, explained that they used to render customary service to the reeve of Winsford, but gave him no rent (absque omni firma donante). Again, there is the marginal entry of Robert's nameless land, which it will be well to collate with the text of the Exeter book.

EXCHEQUER.

Hic Robertus habuit unam virgatam terræ quam tenebat Dodo libere T.R.E. Hec addita fuit Dolvertone manerio Regis. Modo dijudicata est esse tainland. Valet x. sol.

EXETER.

Rodbertus de Odburvilla habuit i. virgam terre quam tenuit Dodo pariter . . . . Hec addita fuit mansioni Regis que vocatur Dolvertona. Modo iterum dijudicata est esse Teglanda et valet per annum x. sol.

Here we have the libere pariter equation, both terms, evidently, excluding subjection to the manor of Dulverton. And, as in the preceding case, we find the law strong enough, or King William just enough, to secure the restoration of the land to its former free status, when that status had been proved. I do not, however, follow Mr. Eyton's observation that 'there are many similar appearances in Domesday of recent question and settlement of De Auberville's estates and position,' (fn. 234) for I cannot find Robert as a holder of land elsewhere than in Somerset.

The marriage portion may be said to occur in connexion with Serlo de Burci, for he, who had given with one of his daughters a manor when she became a nun, gave to another, on her marrying William de Falaise, the manor from which Woodspring Priory, in after days, derived its name. (fn. 235)

Cases of equally divided vills are always deserving of notice; of such vills Puckington is an instance. It is the subject of two consecutive entries, of which each records an assessment of 1½ hides, which in each was apportioned alike, viz. 1 1/8; hide on the demesne, and 3/8 hide on the rest. Each portion had 6 acres of pasture and 66 of woodland and one and a half ploughlands, while the only difference was in the acreage of meadow and in the number of peasants and their ploughs. All this points to Puckington having been one, and subsequently divided into two moieties, probably by two brothers. At the Conquest it was once more thrown into one manor by Roger 'de Corcelle', to whom both moieties had been granted.

The language in which Domesday describes this operation has an important bearing on its use of the term 'manor.' Prof. Maitland held that Domesday employs it as 'a technical term'. . . 'an accurate term charged with legal meaning.' If the words 'pro manerio' were added to 'tenuit,' this, he urged, distinguished the tenure and imparted a special meaning. As against this view I have been able to show that the terms manerium and terra were used indifferently, and that the addition of 'pro manerio' made no difference to the meaning. (fn. 236)

Now of the two Puckington entries in the Exchequer text the first has simply 'Leving tenuit' and does not speak of 'manor,' yet the second refers to it as 'huic manerio.' In the second we read 'Alward tenuit pro 'manerio,' and yet Domesday proceeds at once to speak of these two 'manors' as 'has ii terras.' In the Exeter text, on the other hand, the word 'mansio' (i.e. manor) is found not only in both the entries, but also in the rider, which runs 'has ii mansiones.' The perfect indifference with which Domesday employed or omitted the word is thus made manifest.

To the genealogist the Exeter text is a record of the greatest value; for it enables us at times to identify those of whom the Exchequer text gives us but the Christian names. From it, for instance, we obtain the surnames of such men as William d'Aumary (de Almereio), Richer d'Andeli, Bernard Pauncevolt, Britel de St. Clair, and Walter and William Hussey (Hosatus). From it also we learn the distinctive names of some Englishmen, such as Ælfwine 'Banneson,' whom we could not otherwise trace. Again, the Beatrice, who holds of Ralf de 'Pomerei' at Nether Stowey is entered in the Exeter book as Ralf's sister. Mr. Eyton pointed out that she also held of him a Devon manor, while she further held in that county two of William 'Capra,' who is similarly entered as her brother. On the strength of this he asserted that Ralph and William were brothers, (fn. 237) and although this may seem not absolutely clear, it is interesting to note that Roger 'Capra' and William his son were benefactors, in the next generation, to the Pommeraye abbey of St. Mary Du Val. (fn. 238)

Ralf, though of small account in Somerset, was a great man across its western border. Coming, as his name reminds us, from among the apple orchards of Normandy to make his home amidst those of Devon, he left his castle of La Pommeraye to gain a mightier lordship, and to found that rock fortress, in the heart of English woodlands, which still preserves his name in that of Berry Pomeroy.

More famous, if less enduring, is the name of Ranulf Flambard, who held at Woodwick in Freshford under the abbot of Bath.

It is always a matter of considerable interest when we can identify in Domesday serjeanties which meet us in the records of the 13th century. Wigborough, for instance, in South Petherton is one of the six estates held in Domesday by John the Usher (Hostiarius) among the king's serjeants. And in 1212 (fn. 239) Helen, holder of the ushership, is returned as holding Wigborough with her other lands in Somerset by usher-service. (fn. 240)

Long as this introduction may seem, it would be far longer if space permitted of discussing in detail the returns for the Hundreds to 'the geld-inquest,' as Mr. Eyton styles it. With infinite labour and patience he brought together in his work the entries in these returns and those in the Domesday Survey. To that work, therefore, the reader is referred should he desire to study this aspect of the Survey. Space also does not allow of discussing the relative merits of the Exeter and Exchequer texts or of the genesis of the former and the object with which it was compiled.

We may sometimes be called upon to make our choice between the readings of the Exeter and of the Exchequer Domesday. A notable case in point is discussed in my Feudal England (pp. 425–6), where I have urged that Richard the interpreter claimed to have bought (emit), not held (tenuit) Road from Reinbald the priest and chancellor. This involves our rejecting the Exchequer text's tenuit, but one may point out that, if it were not for the 't,' 'emit' and 'enuit' might very easily be confused by the scribe. It is held by some that the Exeter text is the more exact as well as the fuller, and although the errors of its rival cannot be of much consequence, it is possible that their view is right.

It has already been observed that lack of system in dealing with encroachments on the King's lands, as also on those of others, is a weak point in the Exchequer text. (fn. 241) Mr. Eyton claimed for the Exon text superiority in this respect, because it groups together such aggressions as 'Terræ occupatæ in Sumerseta' (pp. 471–489). But although it is convenient to have them thus brought together, there is no attempt at arranging them in any order, the lands of the King and of his subjects being mixed up in hopeless confusion.

Without entering on the many problems raised by the Exeter book (fn. 242) attention should be drawn to one of its features by which students, I think, have been misled. Its normal system of grouping estates is to deal in turn with those of each tenant or class (1) under Devon (2) under Somerset. Thus we find grouped together:—

Terræ Francorum militum (sic) in Devenesira.

Terræ Francorum tegnorum (sic) in Sumersete Syra (pp. 421–433).

Terræ servientium regis in Devenesira.

Terræ servientium regis in Summerseta (pp. 439–445).

Terræ Anglorum tegnorum in Devenesira.

Terræ Anglorum tegnorum in Sumersetæ syra (pp. 445–457).

Here we have the origin of the misconception dealt with above. (fn. 243) The 'Anglo-thegns,' as writers on Devon and Somerset portions of Domesday call them, are the only true thegns, and are rightly recognized as such in the Exchequer text. The Exeter text is here not even consistent; for its French 'thegns' of Somerset are French 'knights' in Devon. The one term is, if one may say so, no less absurd than the other. For among the so-called 'knights' in Devon is no less a man than 'Richard, the son of Count Gilbert,' one of the mightiest of Domesday nobles and the founder of the house of Clare. Two other Domesday barons, Ralf Paynel and Ralf de Limesi, both of them lords of vast fiefs, are no less strangely here classed among the 'knights' of Devon and the 'thegns' of Somerset. Why they should have been selected from among their fellows for this treatment, and why their respective manors should be entered, not by themselves, but hotch-potch among those of smaller men, is one of the unsolved mysteries, as yet, of the 'Exon' Book. All that can be said is that, here at least, the Exchequer volume appears to advantage, classing, as it does, these magnates among their fellows, and entering their manors under their names instead of making them masquerade as French 'knights' or French 'thegns.' It is greatly to be hoped that the 'Exon' Book will some day receive at the hands of a trained scholar the full and critical treatment of which it stands in need and which may yet reveal its character, its origin, and its object.