To the probable antiquity and origin of the five-hide system I must recur, after glancing at the evidence for the northern and eastern districts of England.

VII. THE SIX-CARUCATE UNIT

One must dissociate at the outset this six-carucate unit from the 'long hundred', or Angelicus numerus, with which Mr Pell confused it. In Mr Stevenson's instructive article on 'The Long Hundred and its use in England',¹³² he has clearly explained that this reckoning only applied to a whole hundred, which, if a 'long' hundred, was really 120. Any lesser number was reckoned in our usual manner. This is seen at once in the test passage at Lincoln (D.B., i. 336a), where 1,150 houses are reckoned as 'novies centum et lxx.', because 'hic numerus Anglice computatur, id est centum pro cxx'.¹³³ The persistence, in Lincolnshire, of the long hundred is well shown in the Inquisitiones post mortem on Robert de Ros, 1311, among those printed by Mr Vincent.¹³⁴ We there read of 'c. acre terra arrabilis per majorem centenam que valent per annum lx. s. prec' acre vj. den.', at Wyville and Hungerton (on the border of Leicestershire); while at Claxby and Normanby (in the north of the county) we have 'cc. acras per minorem centenam et valent c. s. prec' acre vj. d.' Again, at Gedney (in the south-east), we have 'cc. acre terre arrabilis per majus centum et valent per annum xxiiij. li'. prec' acre ij. s. et iiij^xx. acre prati et valet per annum viij. li., prec' acre ij. s. Et cxiij. acre pasture per majus centum et valent per annum ix. li. xix. s. vi. d. prec' acre xviij. d.' On the same property there were due 'ccciiij^xx. opera autumpnalia cum falcis, et valent xxxvj. s. viij. d., prec' operis j. den.', so that these also were reckoned by the long hundred.

Mr Stevenson was not aware of this evidence, but admitted that as the Domesday passage refers to 'such a Danish stronghold as Lincolnshire, it is not free from the suspicion of Danish influence'. His own evidence from a sixteenth-century rental¹³⁵ is subject to a similar criticism. For the general use, therefore, of the 'long hundred' in England he is compelled to rely on the Dialogus de Scaccario and Howden's description of the new survey of 1198, the 'hide or ploughland' being described in both cases as of a hundred acres, where the 'hundred' must have meant 120. But I venture to think that the use of this reckoning for the ploughland, or archaic 'hide', does not establish its general employment. In Domesday, certainly, it is only at Lincoln that we find it actually recognized, houses being reckoned everywhere else on the usual system.

I think, therefore, that we fairly may hold the Anglicus numerus, or long hundred, to have specially prevailed in the 'Danish' districts, which were also assessed, we shall find, in sums of six and twelve. But what was the boundary of this Danish district? It was not the border between Mercia and Wessex, for Mercia was itself divided between the 'six' and the 'five' systems.¹³⁶ Of the two adjacent Mercian shires, for instance, of Leicester and Warwick (afterwards united under one sheriff), we find the latter decimal and the former duodecimal. The military service of Warwick and Leicester was arranged on the same method, yet Leicester sent twelve 'burgesses' to the fyrd where Warwick sent ten. But, it may be urged, the two shires were divided by the Watling Street, the boundary (under the peace of Wedmore) of Danelaw. Was then the Danelaw the district within which the systems prevailed? No, for the Danelaw, under this treaty, included all Cambridgeshire and other hidated districts. The answer, therefore, which I propound is this: The district in which men measured by carucates, and counted by twelves and sixes, was not the district which the Danes conquered, but the district which the Danes settled, the district of 'the Five Boroughs'.

Dependent on these 'Five Boroughs' were the four shires of Leicester, Derby, Nottingham, and Lincoln. For two of the Boroughs, Lincoln and Stamford, both belonged to this last shire, which was, indeed, the stronghold of the system.¹³⁷ Between Stamford and Cambridge we have the same contrast as between Warwick and Leicester, for while Cambridge was divided into ten wards ('custodiæ'), Stamford was divided into six. Lincolnshire, as I have said, was the stronghold of the system, and it is in Lincoln itself that we find Domesday alluding eo nomine to the Anglicus numerus, the practice of counting 120 as 100.

Now in the peculiar district of which I am treating there occurs an important formula which covers Lincolnshire, Yorkshire, Derbyshire, and Notts. Domesday has nothing like it for the other parts of England. Here are the three passages in which we find it recorded:

For comparison with these three passages we may turn to the charter of immunities confirmed to York Cathedral by Henry I, Stephen, and Henry II. We there read:

Si quis enim quemlibet cujuscumque facinoris aut flagitii reum et convictum infra atrium ecclesiæ caperet et retineret, universali judicio vi. hundreth emendabit; si vero infra ecclesiam xii. hundreth infra chorum xviii. ... In hundreth viii. libræ continentur.¹³⁸

As there were twelve carucates in the 'Hundred', so it paid twelve marcs, which, if we can trust the above explanation, themselves came to be termed a 'Hundred'. Moreover, the 'Hundreds' themselves were grouped in multiples of six. So too the Yorkshire thegn who held six Manors or less paid three marcs to the sheriff; if he held more than six, twelve marcs to the king (Domesday, i. 289b).

It is a special feature of the 'Danish' district that each territorial 'Hundred' contained twelve 'carucatæ terræ'. This point is all-important. Just as a 'Hundred' to an Anglo-Saxon suggested one hundred 'hides', so to the Danes of this district it suggested twelve 'carucates'. Nay, to the men of Lincolnshire there could be no more question that twelve carucates made a 'Hundred' than there could be now, among ourselves that twelve pence make a shilling. If we turn to the Lindsey Survey,¹³⁹ a generation later than Domesday, we obtain proof to that effect. We find that Survey, in three instances, adding up all the estates of a tenant within a Wapentake, and giving us the result in 'Hundreds' and 'carucates'. Here are the actual figures:

Now we must observe that these 'Hundreds' are not districts with 'a local habitation and a name'; they are merely sums of twelve carucates produced by compound addition. We further find, at the head of the survey of each Wapentake, a note that it is reckoned to contain so many 'Hundreds', with the explanation, in some instances that in each 'Hundred' were 'xii. carucatæ terræ'.¹⁴³ But even here the real unit is shown to be 'six carucates', for several Wapentakes contain an odd 'half-hundred', while in that of Horncastle this is actually entered as 'six carucates'.

Here are the nineteen Wapentakes, with the number of Hundreds assigned to each, and the number of 'carucatæ terræ' that such Hundreds would imply:

All the above, it will be seen, are multiples of the six-carucate unit. That the aggregate of recorded 'carucatæ terræ' appears to differ, though slightly, from the totals here given only shows how vain is the argument that, because the recorded aggregates of Hundreds may often be uneven figures, there could therefore have been no system at work such as I contend there was. Clerical error and special alterations have both to be allowed for.

It has never, so far as I know, been pointed out that these Lindsey Trithings were so arranged as to contain an approximately equal number of 'Hundreds'. So far as it is possible now to reckon them, the South Trithing contained 51½, the North Trithing 51½, and the West Trithing 49½. Fifty 'Hundreds' would represent 600 carucatæ; and it is, to say the least, a singular coincidence that, in the archaic territorial list that has hitherto baffled investigation, the North Gyrwa, South Gyrwa, and Spalda are reckoned each at 600 hides.¹⁴⁵

I shall now give some instances of Lindsey townships assessed on the basis of the six-carucate unit:

These instances will illustrate the value of the Lindsey Survey in enabling us to group the fractional assessments which appear in Domesday Book. Here are some other varieties:

Lastly, to complete the parallel with the Leicestershire Hundreds infra, we may take this case (cf. p. 63, note 122):

Precisely the same system prevailed in Holland as in Lindsey, for the 'Testa de Nevill' preserves for us the constituents of a Holland Wapentake, that of 'Elhou':

The Lindsey Survey would describe such a Wapentake as containing 'Seven Hundreds'.

Crossing the border from Lincolnshire into Rutland (i.e. the Rutland of Domesday), we find the same system at work that meets us in the Lindsey Survey. We read:

In Alfnodestou Wapent' sunt ii. Hundrez. In unoquoque [sunt] xii. carucatæ ad geldum.... In Martinesleie Wap' est i. hundret, in quo xii. carucatæ ad geldum.—D.B., i. 293b.

On analysing the contents of these Wapentakes, we find this statement fully borne out, the former containing twenty-four, and the latter twelve, 'carucatæ terræ'. These are carefully contrasted throughout with the 'terra carucæ' or areal measure.¹⁴⁶

In Yorkshire, Notts and Derby, we have less direct evidence. Sawley, in Derbyshire, has indeed been alleged to be entered in Domesday as a Hundred of twelve carucates, but Domesday does not justify this assertion being made.¹⁴⁷ I would rather trust to the notable formula, which, as I explained at the outset, is common to these counties for proof that they also were arranged in 'Hundreds' of twelve carucates.

The prevalence, however, of assessment by sixes, threes, and twelves, meets us on every side, as does, in hidated districts, the assessments by fives and tens. At the outset, for instance, of the survey of Yorkshire we have the district 'gelding' with the city assessed at eighty-four (12×7) carucates (which would be described in Lincolnshire as seven 'Hundreds'). We have two lists of the details, which are given here.¹⁴⁸

These lists have a value independent of their illustration of the arrangement in threes and sixes. They show how Domesday breaks down, when it supplies a check upon its own evidence, by failing to make its details agree with its total; and they further show by the discrepancy between them how easily error may arise, and how rash it must be to argue from a single case.¹⁴⁹

Yorkshire presents other traces, in its Hundreds, of the same system. Thus the townships in the Hundred of 'Toreshou' follow one another in this order: 18, 18, 20, 6, 18, 8, 12, 12 (8+4), 6, 18, 8, 18, etc. (infra, p. 80).

But my strong evidence is found in an invaluable survey of Leicestershire, unknown till now to historians,¹⁵⁰ which does for the carucated districts just what the Inq. Com. Cant. does for the hidated ones. Here we find the townships grouped in small blocks of from six to twenty-four 'carucatæ terræ', as a rule with almost monotonous regularity. And these blocks are further combined in small local Hundreds, of which the very existence is unknown to historians and antiquaries,¹⁵¹ and which are usually multiples, like the Lincolnshire Wapentake, of the six-carucate unit.

It will be remembered that in the case of Cambridgeshire, I selected for my first two examples a Hundred of 50 hides, composed of 5 Vills assessed at 10 hides each, and a Hundred of 70 hides, composed of 7 Vills, assessed at 10 hides each. In Leicestershire, precisely in the same manner, I shall begin with the simplest forms and select Hundreds of 36 and 48 carucates, composed of Vills uniformly assessed at 12 carucates each.

From these we may advance to other combinations:

With these types as clues we are in a position to assert that where the total assessment of a Hundred varies but slightly from a multiple of six, there must have been some slight error in one of the figures. Thus Hundreds of 35½, 34¹³⁄16 carucates, etc., may be safely assumed to have been Hundreds of 36 carucates; those of 41, 43⅞, etc., would be of 42 carucates; those of 48⅞, 50, etc., would be of 48 carucates. These slight discrepancies, precisely as in Lincolnshire, are accounted for by Vills of 6 or 12 carucates, being entered as of 5⅞, 5¹³⁄16, 6¾, or 11⅞, 13, etc. Thus:

The most usual Leicestershire Hundreds are those of 36, 42, and 48 carucates, which, be it observed, would be described in the language of the Lindsey Survey as 'Wapentakes' of 3, 3½, and 4 'Hundreds' respectively. The name may be different: the thing is the same.¹⁵²

It will have been seen by this Survey that the 'Vills', single or grouped, were assessed precisely as in Cambridgeshire, save that there the assessment was reckoned in fives and tens, while here it was in sixes and twelves.

VIII. THE LEICESTERSHIRE 'HIDA'

The case of Leicestershire introduces us to a very curious point. Leicestershire is not one of those counties to which the singular formula that I discussed above refers. This suggests that it was not arranged in 'Hundreds' of twelve carucates. The above Survey confirms this, for it shows us Hundreds resembling in character those found in the hidated districts. But although the twelve-carucate unit of the 'Hundred' is not found in Leicestershire, we do find in it a group-unit, and that unit is the hida. Just as we have seen the Hundred used in two wholly different senses, so also was the 'hida'. The quite peculiar way in which 'hida' occurs in Leicestershire (which was not a hidated but carucated district) completely baffled Mr Eyton, and was misunderstood by Mr Pell.¹⁵³ Both writers failed to observe not only that the use of 'hida' is here of a peculiar character, but also that the normal 'hida' of Domesday (from which they could not emancipate themselves) would be quite out of place in this carucated district.

The first point to grasp is that this Leicestershire 'hida' was a term which, locally I mean, explained itself. It is used at least a dozen times in the Survey of Leicestershire without any mention of its contents. Those contents must have been, therefore, familiar and fixed. But what were those contents? Three incidental notices enable us to determine them:

231 (a), 2: 'Ibi est i. hida et iiii^ta. pars i. hidæ. Ibi sunt xxii. car' terræ et dimidia.'

236 (a), 1: 'II. partes unius hidæ, id est xii. car' terræ.'

237 (a), 2: 'II. partes unius hidæ, id est xii. car' terræ.'

Just as the 'Hundred' of Lincolnshire was a sum of twelve carucates, so the 'Hide' of Leicestershire was a sum of eighteen carucates.¹⁵⁴ Working in the light of this discovery (for as such I claim it), we find that the other 'hides', thus interpreted, give us an aggregate of 'carucates' obviously suitable to the recorded ploughlands.¹⁵⁵ It may, however, be fairly asked why Domesday should speak in one place of half a 'hide', and in another of nine 'carucates'; in one place of a hide and a third, and in another of twenty-four carucates. The answer is that the singular love of variety which distinguishes Domesday in Cambridgeshire (as we saw) is at work here also. For instance, two equal estates are thus described: 'Willelmus iiii. car' terræ et dimidiam et iii. bovatas, et Rogerus iiii. car' terræ et vii. bovatas' (fo. 234a). The same instinct which led the scribe to enter these seven bovates as half a carucate plus three bovates, led him also to enter ten and a half carucates as half a hide plus a carucate and a half (fo. 237a).

But to the rule I have established there is a single exception. We read of 'Medeltone' in this shire: 'Ibi sunt vii. hidæ et una carucata terræ et una bovata. In unaquaque hida sunt xiiii. carucatæ terræ et dimidia' (fo. 235b). The actual formula employed is unique for the shire, and the figures are specially given as an exception. But, with singular perversity, Domesday students have always been inclined to pitch upon the exceptions as representing the rule, forgetting that it was precisely in exceptional cases that figures had to be given. In normal cases they would have been superfluous.

Several years have elapsed since I wrote the above explanation, but I have decided to publish it exactly as it originally stood. In the meanwhile, however, Mr Stevenson has dealt with the subject in an article on 'The Hundreds of Domesday: the Hundred of Land' (1890).¹⁵⁶ He has advanced the ingenious theory that the Leicestershire 'hida' was only a clerical error for H[undred], and that it was really that 'Hundred' of twelve carucates which we meet with in the Lindsey Survey. To prove this, he reads an entry on 236a, 1, as 'Ogerus Brito tenet in Cilebe de rege ii. partes unius hidæ, id est xii. car[ucatæ] terræ', and claims that this gloss defines the 'hida' as a 'hundred' of twelve carucates. I confess that to me such a rendering is in the highest degree non-natural. If we speak of 'two-thirds of a yard, that is twenty-four inches', we should clearly imply that the yard itself was thirty-six inches, not twenty-four. Similarly, I claim to render the 'gloss' as implying that the 'hida' itself contained eighteen carucatæ, not twelve.¹⁵⁷ If I am right, Mr Stevenson's suggestion that this 'hida' was really a 'Hundred' also falls to the ground.

After careful study of the Domesday Survey of Leicestershire, I definitely hold that in that county 'carucata terræ' was the geld-carucate and 'terra x car[ucis]' the actual ploughlands.¹⁵⁸ Now there are only three instances in which the Survey records the assessment both in terms of the 'hida' and in 'carucatæ terræ', and in all three the figures support my own theory. The Abbot of Coventry's Burbage estate (231a, 2), where a 'hide' and a quarter equates 22½ 'carucatæ terræ', is a test-case, and Mr Stevenson there takes refuge in a suggested 'beneficial hidation'. The exact formula, no doubt, is peculiar, but reference to the text shows that 's[un]t' has been interpolated between 'ibi' and 'xxii.' I suspect that the scribe had written 'ibi' (from the force of habit) when he ought to have written 'id est'.

I close this portion of my essay by applying my own theory to the case of 'Erendesbi' (Arnesby). The relative entries are:

'Episcopus Constantiensis tenet in Erendesber iii^as. car[ucatas] terræ et dim. et unam bovatam (231).'

'W[illelmus] Pevrel tenet dim. hidam et iii. bovatas terræ in Erendesbi (235).'

Put into figures they work out:

So that Arnesby was a typical Vill assessed at twelve carucates.¹⁵⁹

IX. THE LANCASHIRE 'HIDA'

There is one other case of a peculiar 'hide' in Domesday. This is that which is found in the land 'between Ribble and Mersey', that district of which the description offers so many peculiarities. We find it divided into six hundreds, and of the 'hides' in the first, that of (West) Derby, we read: 'In unaquaque hida sunt vi. carucatæ terræ' (i. 269b). Whether or not that explanation applies, as is believed, to the whole district, we have here again a 'Danish' place-name brought into direct relation with the six-carucate unit. On the opposite bank of the Mersey lay the Wirral peninsula, in which this system of assessment cannot be traced.

Mr Green alluded to the Danish 'byes' as found, by exception, 'about Wirral in Cheshire',¹⁶⁰ and held that Norsemen from the Isle of Man had founded 'the little group of northern villages which we find in the Cheshire peninsula of the Wirral'.¹⁶¹ I cannot find them myself. In his 'Notes on the Domesday Survey, so far as it relates to the Hundred of Wirral'¹⁶² (1893), Mr Fergusson Irvine, in a paper which shows, though somewhat discursive, how much can only be done by intelligent local research, has collated all the Domesday entries. 'Raby' is the one place I can there find in the peninsula with the 'bye' termination; while out of fifty-one entries twenty refer to places with the English termination 'tone', and the Anglo-Saxon test-words 'ham' and 'ford' are found in four others. There were, doubtless, Norse elements in the peninsula, but they were not strong enough to change the place-names or divide the land on their own system. In the same way, Chester had its 'lawmen', though it was not one of the Five Boroughs, nor is what I have termed the Scandinavian formula applied to Cheshire in Domesday. So, too, there were lawmen at Cambridge, and their heriot included eight pounds,¹⁶³ which occur in the above formula as the twelve marcs of the Danish 'Hundred'. Yet the whole system of Cambridgeshire was non-Danish. It was only, in short, where the northern invaders had settled down as a people that they were strong enough to divide the land anew and organize the whole assessment on their own system.

X. THE YORKSHIRE UNIT

We have seen that the unit of assessment for the carucated districts of England was 'vi. carucatæ terræ', just as five hides was the old unit in the south. We have also seen that the former reckoning extended over those districts which the Danish immigrants had settled. There remains the question whether the Danes had merely substituted six for five in the pre-existing arrangement, or had made a wholly new one for themselves based on actual area.

It is primâ facie not probable that they can have adopted the latter course, for the uniformity of their assessment proves its artificial character. Yet, in his remarkable paper on 'The Ploughland and the Plough',¹⁶⁴ Canon Taylor has arrived at the conclusion that:

The geldable carucate of Domesday does not signify what the carucate usually signifies in other early documents. The 'carucata ad geldum' is not as commonly stated by Domesday commentators, the quantity of land ploughed in each year by one plough, but it is the quantity tilled in one year in one arable field by one plough.¹⁶⁵

This 'novel and important proposition', as its author truly described it, was probably the most notable contribution to our knowledge that the Domesday Commemoration produced. The Canon's theory, which (so far as his own East Riding is concerned) he certainly seems to have established, is, at first sight, fatal to mine. But, on the other hand, my own theory can be proved no less clearly for Leicestershire, where the 'carucata terræ' and the ploughs are often connected in about the same ratio as in Yorkshire.¹⁶⁶ This leads us to inquire whether, even in the East Riding (where his theory works best), we may not find traces of that assessment by the six-carucate unit which I advocate myself. Such traces in Yorkshire we have already seen,¹⁶⁷ but there is other and stronger evidence.

If we take the modern Wapentake of Dickering (the first on Canon Taylor's list) and examine its three Domesday Hundreds of Turbar, Hunton, and Burton, we obtain these results:¹⁶⁸

The evidence of this last Hundred is so overwhelming that it cannot be gainsaid.¹⁶⁹

I claim, therefore, that my theory holds good even in Canon Taylor's stronghold, but I do so without venturing to dispute the accuracy of his own. How far they can be reconciled I leave to others to decide.

There are certain difficulties, however, which his brilliant suggestion must raise. It is the essence of his theory that in a two-field Manor the ploughland of 160 acres (half fallow) was assessed at one 'carucata terræ', while in the three-field Manor the ploughland of 180 acres (a third fallow) was assessed at two. This would be an obvious and gross injustice. Again, remembering that, according to the Canon, the proportion of 'carucatæ' to ploughlands should be either 2 to 1 or 1 to 1, what are we to make of such figures as these, taken at a venture from a page of the Leicestershire Survey (232a, 1):

It is certainly difficult to discover any regular or consistent assessment in a system where the ploughland was represented by anything from ½ carucata to 2¼ carucatæ. There is, however, in so many cases an approximation to an assessment of three carucatæ for two ploughlands, that there seems to have been some underlying idea, if we could only trace it out. But for this there is needed a special investigation of all the carucated counties, a work of great labour and requiring local co-operation. If we could have tables for each county, arranged Hundred by Hundred and Vill by Vill, showing in parallel columns the ploughland and the carucatæ ad geldum, we could then, and only then, venture to speak positively. Till that is accomplished we are not in a position to explain how a system of assessment, based on actual area, could result in aggregate assessments uniformly expressed in terms of the six-carucate unit.

XI. GENERAL CONCLUSIONS

Now if we follow the clue afforded by the Cambridgeshire evidence, and hold that the assessment was originally laid not on the Manor, nor even on the Vill, but on the Hundred as a whole,¹⁷⁰ it might be suggested that the Hundred itself subdivided the amount among its constituent elements. In practice, indeed, from the nature of the case, this principle must have prevailed in every town assessed at a Hundred or Half-Hundred, for where an urban community was assessed in 'hides' the burgesses must, as in later days, have settled among themselves the proportion to be borne by individuals or individual properties. If, then, they were able to do this, and if, as I hold, town and country were assessed on the same principle, as part of the same system, what was to prevent their neighbours, in the court of the rural Hundred, similarly distributing among its constituents their respective shares of the common burden?

We might even be tempted to go far further than this, and to carry our discoveries to a logical conclusion. If, as is asserted, direct taxation ('geld') began in England with the need for raising money to buy off the Danes, let us ask ourselves how the Witan would proceed when confronted with a demand, let us say, for £10,000. As there had been hitherto, ex hypothesi, no direct taxation, there would be no statistical information at their disposal, enabling them to raise by a direct levy the sum required. Their only possible resource, we might hold, would be to apportionate it in round sums among the contributory shires. Proceeding on precisely the same lines, the county court, in its turn, would distribute the quota of the shire among its constituent Hundreds, and the Hundred court would then assign to each Vill its share. As the Vills were represented in the Hundred court, and the Hundreds in the Shire court, the just apportionment of the Shire's quota would be thus practically secured. The arrangement would, moreover, be as satisfactory to the Witan as it was fair to the contributors inter se; for, by this gradation of responsibility, the payment of the whole was absolutely secured. This explanation is very tempting, and, indeed, such a system of apportioning liability is to be traced from time immemorial in the Indian village community.¹⁷¹ Moreover, if the ratio of 'hides' to ploughlands were found to vary to any marked extent, according to county, the hypothesis that the quota, in the first instance, was laid upon each county would duly explain the ratio assessment being higher or lower in one county than in another.

But such an hypothesis would imply that this assessment dated only from the days of Æthelred, or circ. 1000. Now the five-hide unit, on the contrary, was undoubtedly an old institution. Church lordships, the easiest to trace, appear to have retained their hidation unchanged from early times, and the 'possessio decem familiarum' of Bede seems to carry the decimal system back to very early days. Mr Seebohm, indeed—though, like others, he had failed to discover the existence of the five-hide system—saw in this 'possessio' of Bede a connecting link with the Roman decuria, just as he saw in the Roman jugatio the possible origin of English hidation. And we must, of course, trace its artificial arrangement (1) either to the Romans, (2) or to the Britons—assuming them to have had the same system as existed in Wales for the food-rents, (3) or to the English invaders.

Arrested at this point by the difficulty of assigning to the system I have described its real origin, I dropped these studies for some years in the hope that there might come from some quarter fresh light upon the problem. As I cannot, however, for lack of evidence, propound a solution capable of proof, I will content myself with indicating the line of research that offers, I venture to think, the most likelihood of success.

The proportionate sums contributed by the several counties to the Danegeld present a fruitful field of inquiry, but one, it would seem, as yet unworked. Mr Eyton, it is true, observed that 'in Devon and Cornwall the scope of the gheld-hide was enormous',¹⁷² that is, in other words, the assessment was strangely low, but it did not occur to him to seek the cause of the phenomenon he observed. If, as was the case, West Wales was assessed on quite a different scale to the counties adjoining it on the east, it may suggest a conclusion no less important than that, when the latter were originally assessed, West Wales was not yet a portion of the English realm. But, before concluding that the hide assessment is proved to be as ancient as this, we must see whether it is possible to detect any principle at work in the total assessments of the several counties, any relation between their area and the sums they contributed to the geld as entered in the Pipe Roll of 1130, our first evidence on the subject.

For such an enquiry it is especially needful to insist on breadth of treatment. In the first place, the modern area of the counties may vary more or less from the original extent;¹⁷³ in the second we have no proof that the assessment had always been the same, though the tendency in early days, no doubt, was to stereotype such figures. We must not, therefore attempt close or detailed investigation but if, on a review of the whole evidence, we detect certain broad features, uneffaced by the hand of time, we may fairly claim that we have in these the traces of a principle at work, the witness to a state of things prevailing in the distant past.

On comparing the contributions to a 'geld' at two shillings on the hide with the (modern) area of counties, we find that a rate of about a pound for every seven square miles prevailed widely enough to be almost described as normal.

The three eastern counties work out thus:

In all three cases the proportion to the square mile is between a sixth and a seventh of a pound. In Cambridgeshire it is just under, in Sussex, just over, a seventh:

Most remarkable, however, is this Midland group:

It is remarkable, not only for this agreement inter se, but also for the sharp contrast it presents to the groups of counties, lying respectively to the south-east and the north-west of it. The former approximates a rate twice as high, namely, two-sevenths of a pound to the square mile:

Taking this group as a whole, it paid £1,032 18s 1d, a curiously close approximation to the £1,021⁴⁄7 which my suggested rate of ²⁄7 would give. Middlesex was so exceptional a county, that one hardly likes to include it, but there also the rate was a little over two-sevenths.