included in the Cambridgeshire Inquest but is included in the Inquisitio
Eliensis. We speak of the Isle of Ely. There are entries which, having
told us how many hides a manor contained, proceed to allot these among
their various occupants, and, as in some of these cases a calculation by
acres is mixed up with a calculation by hides, they hold out a hope that
we may be able to discover how many acres were reckoned to the hide. We
will begin with Ely itself. 'Ely defends itself for 10 hides.... In
demesne there are 5 hides ... and there are 40 villeins with 15 acres
apiece ... and 18 cottiers and 20 serfs[1555].' Now if from the total of
10 hides we subtract the 5 that are in demesne, this leaves 5 others,
and if we divide these 5 among the 40 villeins this gives to each
villein 1/8th of a hide; but we are told that each villein has 15 acres;
therefore it follows that 120 acres make a hide. We reckon that in eight
other cases[1556] the same method of computation is followed, though in
one of these a hide divided among 17 villeins is said to give them 7
acres apiece and this shows us how a single acre may be neglected in
order to avoid a very ugly fraction[1557]. Against these cases must be
set seven which give less pleasing results[1558]. In at least one of
these no possible theory will justify the arithmetic of our record as it
stands[1559], and there is no accord between the remaining five.
[Evidence from Middlesex.]
At first sight the survey of Middlesex seems to offer materials similar
to those that come to us from Cambridgeshire. Very curious and
instructive they are. A Middlesex entry will usually give us the number
of hides (_A_), the number of teamlands (_B_), the number of teams
(_C_), and also certain particulars which state the quantity of land
that there is in demesne and the quantities held by divers classes of
tenants. The sum of these particulars we may call _P_. Now we begin by
hoping that _P_ will be equal to _A_, and, since the particulars often
contain acres as well as hides and virgates, we hope also to discover
the equation that is involved in the sum. As an example we will take a
case in which all goes well. At Cowley a manor defends itself for two
hides; in demesne are one and a half hides; two villeins have a half
hide. Here _A_ = 2 H. and _P_ = 1-1/2 H. + 1/2 H.; so all is as it
should be. But we soon come upon cases in which, though we make no
assumption about the relation of the acre to the hide, our _P_ refuses
to be equal to our _A_. Then perhaps we begin to hope that _P_ will be
equal to _B_: in other words, that the sum of the quantities ascribed to
lord and tenants will be equal to the number of teamlands. But this is
more fallacious than the former hope. We will put a few specimens in a
table[1560].
Hides Teamlands Sum of particulars
Harrow (Abp. Canterbury) 100 70 46-1/2 H. + 13 V. + 13 A.
Stepney (Bp. London) 32 25 18-1/2 H. + 48-1/2 V.
Fulham (Bp. London) 40 40 41-1/2 H. + 30 V.
Westminster (Abbot) 13-1/2 11 10 H. + 14-1/2 V. + 5 A.
Sunbury (Abb. Westminster) 7 6 4 H. + 10-1/2 V.
Shepperton (Abb. Westminster) 8 7 3-1/2 H. + 17 V. + 24 A.
Feltham (C. Mortain) 12 10 6 H. + 16-1/2 V.
Chelsea (Edw. of Salisbury) 2 5 1 H. + 4 V. + 5 A.
[Meaning of the Middlesex entries.]
We seem to have here three independent statements, and, though
throughout the county _P_ shows a tendency to keep near to _A_, still we
must not make calculations which suppose that the 'hide' of _A_ is the
'hide' of _P_. Take Chelsea for example. We must not say: 2 H. = 1 H. +
4 V. + 5 A., and therefore four virgates and five acres make a hide. No,
it seems possible that in these Middlesex 'particulars' we do at last
touch real agrarian arrangements. At Fulham the bishop has 13 hides in
demesne; 5 villeins have 1 hide apiece; 13 villeins have 1 virgate
apiece; 34 have a half-virgate apiece; 22 cottiers have in all a
half-hide; Frenchmen and London burgesses have 23 hides; so there are
41-1/2 hides and 30 virgates. That we take to be the real arrangement of
the manor, though we are far from saying that all its hides are equal.
But it gelds for only 40 hides. A virgate can not be a negative
quantity. Therefore we need say no more of these Middlesex entries, only
in passing let us observe that the discrepancy between _P_ and _B_ is
often considerable, and this seems to show that the teamland of these
Middlesex jurors is not in very close touch with the agrarian and
proprietary allotments.
[Evidence in the Geld Inquests.]
To yet one other quarter we have hopefully turned only to be
disappointed, namely, to the so-called Geld Inquests, copies of which
are placed at the beginning of the Exeter Domesday. They tell us of a
geld that obviously is being levied at the rate of six shillings on the
hide, and sometimes they seem to tell us expressly or implicitly the
amount that an acre pays. For a moment we may think that we are
obtaining valuable results. Thus at Domerham we find that 14 hides
minus 4 acres pay £4. 3_s._ 8_d._ We conclude that each acre is taxed at
one penny and that 72 A. = 1 H.[1561]. Then at Celeberge 20 H. minus 4
A. is taxed at £5. 19_s._ 6_d._ We conclude that each acre is taxed at
three-half-pence and that 48 A. = 1 H.[1562]. But we soon come to sums
which are absurd and discover that as regards small quantities these
documents are for our present purpose quite useless. For the Wiltshire
hundreds we have three different documents. They do not agree in their
arithmetic. Probably they represent the efforts of three different
computers. Indubitably one or more of them made blunders. To give one
example:--one of our documents begins its account of Mere by saying that
it contains 85 hides, 1/2 a hide and 1/2 a virgate; the other two
documents say 86 hides, 1/2 a hide and 1 virgate[1563]. This is by no
means the only instance of such discrepant results. But mere clerical or
arithmetical errors are not the only obstacle to our use of these
accounts. It soon becomes quite evident that small amounts are dealt
with in an irregular fashion. Thrice over we are assured that 15 H. 1/2
V. paid the king £4. 11_s._ 0_d._[1564]; but they should have paid £4.
10_s._ 9_d._, if four virgates make a hide. Thrice over we are assured
that 64-1/2 H. paid £19. 6_s._ 10_d._[1565]. All suppositions as to
acres and virgates apart, 64-1/2 H. should have paid £19. 7_s._ 0_d_ In
Somersetshire the calculations do not speak of acres, but they introduce
us to the _fertinus_ or farthing, which is certainly meant to be the
quarter of a virgate. Numerous entries show us that 4 _fertini_ = 1
virgate, and yet when a mass of land expressed in terms of hides,
virgates and farthings is said to pay a certain sum for geld, we find
that the odd farthings are reckoned as paying, sometimes 3_d._,
sometimes 4_d._, sometimes 4-2/3_d._, sometimes 5_d._, sometimes 6_d._
per farthing[1566]. So again, when additions are made, odd acres are
ignored. We are told that in a certain hundred the barons have 20 hides
in demesne, and then that this amount is made up by the following
particulars, 8 H. + 1 V. + 3 H. + 3 V + 4-1/2 H. - 4 A. + 3-1/2 H. It is
obvious that these particulars when added together do not make 20 hides,
though they may well make 20 hides and 4 acres[1567]. A study of these
Geld Inquests has brought us reluctantly to the conclusion that, though
they amply prove that 4 V. = 1 H., they afford no proof as to the number
of acres that are reckoned to the virgate[1568].
[Treatment of small quantities.]
One word to explain that the apparent rudeness with which small figures
are treated is not due to any persuasion that they may be safely
disregarded, but is rather the natural outcome of a partitionary method
of taxation. Little quantities are lost in the process. It is known that
a certain hundred should have, for example, 80 hides and a certain vill
5 hides: but when you come to add up the particulars you can not bring
out these round figures, perhaps because many years ago a small error
was made by some one when an estate of 2-3/4 hides was being divided
into 7 shares. If a mistake be made, it can never be corrected; the
landowner who has once or twice paid for 47 acres will refuse to pay for
48 and will tell you that the deficient acre does not lie on his land.
[Result of the evidence.]
The ignes fatui which dance over the survey of Middlesex and the Geld
Inquests of the south-western counties have for a while led us from our
straight path. We have seen that in Cambridgeshire the equation 1 H. = 4
V. = 120 A. is employed on at least twenty occasions. Now as to the rest
of England it must at once be confessed that we have no such convincing
evidence. In many counties acres of arable land are but rarely
mentioned; parcels of land which geld for less than a hide are generally
expressed in terms of hides and virgates; we read, for example, not of
so many acres, but of the ninth part of a hide or of two third parts of
a virgate. Thus we are compelled for the more part to fall back upon the
presumption that the treasury has but one mode of reckoning for the
whole of England.
[Evidence from Essex.]
But we would not rest our case altogether upon probability. In Essex we
find one fairly clear case in which our equation is used[1569].
Sometimes, again, we read that a tract of land is, or gelds for, or
defends itself for _x_ hides and _z_ acres, or for _x_ hides, _y_
virgates and _z_ acres. Now in any entry which takes the first of these
forms we have some evidence that _z_ acres are less than one hide, and
from any entry which takes the second of these forms we may infer that
_z_ acres are less than one virgate. Of course from such a statement as
that '_A_ holds 90 or 115 or 240 acres' we draw no inference. It is
common enough in our own day to speak of things costing thirty shillings
or eighteen pence. But we never speak of things costing one pound and
thirty shillings, or one shilling and eighteen pence, and we should
require much proof before we thought so meanly of our ancestors as to
suppose that they habitually spoke in this clumsy fashion.
Let us use this test. Happily in Essex we very frequently have a tract
of land described as being _x_ hides and _z_ acres.
Now we read of
a half hide and 30 acres[1570],
a hide and a half and 31 acres[1571],
a half hide and 35 acres[1572],
a half hide and 37 acres[1573],
a hide and a half and 40 acres[1574],
a hide and a half and 45 acres[1575],
a half hide and 45 acres[1576],
two hides and a half and 45 acres[1577],
a half hide and 48 acres[1578],
_x_ hides and 80 acres[1579],
nine hides and 82 acres[1580].
We have here cited twenty instances in which, as we think, the hide
exceeds 60 acres (we might have cited many others) and twelve in which
it exceeds 80 acres. We might further adduce instances in which our
record speaks of a virgate and 10 acres, a virgate and 15 acres, and
even of a virgate and 20 acres[1581], and when we read of two hides less
30 acres and two hides less 40 acres[1582] we infer that a hide probably
has not only more but considerably more than the 30, 40 or 48 acres
that are allowed to it by Kemble and Eyton. Our argument is based on the
belief that men do not habitually adopt extremely cumbrous forms of
speech. From a single instance we should draw no inference, and
therefore when we just once read of 'three hides and a half and 80
acres' we do not infer that 80 acres are less than half a hide[1583].
[Evidence from Essex continued.]
But more can be made of these returns from Essex. We will take a large
number of tracts of land described in the formula '_x_ hides and _z_
acres'; we will observe the various numbers for which _z_ stands, and if
we find some particular number frequently repeating itself we shall be
entitled to argue that this number of acres is some very simple fraction
of a hide. We will take at hazard 100 consecutive entries which contain
this formula--'_x_ hides + _z_ acres,' where _x_ is either an integral
number or 1/2. The result is that in 37 cases _z_ is 30, in 12 it is 15,
in 8 it is 40; then 35 and 20 occur 5 times; 80, 50, 45, 37, 18, 10
occur thrice, and 38 and 15-1/2 twice; eleven other numbers occur once
apiece. There can we think be but one explanation of this. The hide
contains that number of acres of which 30 is the quarter, 40 the third,
15 the eighth[1584].
[Further evidence.]
But Essex, it must be confessed, lies next to Cambridgeshire, and for
the rest of England we have less evidence. Still there are entries which
make against any theory which would give to the hide but 30, 40 or 48
acres. In Hertfordshire we read of 'a hide and a half and 26
acres[1585].' In the same county we read of 'a half virgate and 10
acres,' and this seems to tell of a hide of at least 88 acres[1586]. In
Gloucestershire we read of a manor of one hide and are told that 'in
this hide, when it is ploughed, there are but (_non sunt nisi_) 64 acres
of land,' whence we may draw the inference that such an acreage was
unusually small[1587]. We pass from Mercia into Wessex. In
Somersetshire we read of 'three virgates and a half and 5 acres[1588],'
in Dorset of 'three virgates and a half and 7 acres[1589],' in Somerset
of 'one and a half virgates and 8 acres[1590].'
[Acreage of the fiscal carucate.]
To prove that the fiscal carucate was composed of 120 (fiscal) acres is
by no means easy. If, however, we have sojourned for a while in Essex
and then cross the border, we can hardly doubt that in East Anglia the
carucate bears to the acres the relation that is borne by those hides
among which we have been living. Norfolk and Suffolk are carucated
counties, but while in the other carucated counties it is usual to
express the smaller quantities of land in terms of the bovate (8 bovates
making one carucate) and to say nothing of acres, in East Anglia, on the
other hand, it is uncommon to mention the bovate--in Suffolk we may even
find the virgate[1591]--and men reckon by carucates, half-carucates and
acres. We allow the description of Suffolk to fall open where it pleases
and observe a hundred consecutive cases in which a plot of land (as
distinguished from meadow) is spoken of as containing a certain number
of acres. In 22 cases out of the hundred that number is 60, in 8 it, is
30, in 7 it is 20, in 5 it is 40, in 5 it is 15; no other number occurs
more than 4 times, and yet the numbers that appear range from 100 to 2.
We have tried the same experiment on two hundred cases in Norfolk; in 28
cases the number of acres was 30, in 16 cases it was 60, in 13 it was
40, in 13 it was 16, in 12 it was 20, in 10 it was 80, in 9 it was 15,
though the numbers ranged from 1 to 405. Surely the explanation of this
must be that 60 acres are half a carucate, that 30 acres are a quarter,
that 40 acres are a third, 20 a sixth, 15 an eighth. We have made many
similar experiments and always with a similar result; wherever we open
the book we find plots of 60 acres and of 30 acres in rich abundance. We
use another test. When land is described by the formula '_x_ carucatae
et _z_ acrae,' what values are assigned to _z_? We find 40 very
commonly, 42, 45, 50, 60 (but this is rare, for it is easier to say
'_x_-1/2 carucates' than '_x_ carucates and 60 acres') 68, 69, 80 (at
least four times), 81, and 100[1592]. On the one hand, then, we have a
good deal of evidence that the carucate contains more than 80 acres,
some evidence that it contains more than 100 acres, and some that it
does not contain many more, for no case have we seen in which _z_
exceeds 100. Perhaps in Norfolk the figure 16 occurs rather more
frequently than our theory would expect, but 16 is two-fifteenths of
120, and the figures 32 and 64 occur but rarely. Also it must be
confessed that in Derbyshire we hear of 'eleven bovates and a half and
eight acres,' also of 'twelve bovates and a half and eight acres[1593].'
These entries, to use an argument which we have formerly used in our own
favour, seem to imply that half a bovate is more than eight acres and
would therefore give us a carucate of at least 144. We can only answer
that, though men do not habitually use clumsy modes of reckoning, they
do this occasionally[1594].
[Acreage of the fiscal sulung.]
Of the Kentish sulung very little can be discovered from Domesday.
Apparently it was divided into 4 yokes (_iuga_)[1595] and the yoke was
probably divided into 4 virgates. We have indeed one statement
connecting acres with sulungs which some have thought of great
importance. 'In the common land of St. Martin [i.e. the land which
belongs to the _communitas_ of the canons of St. Martin] are 400 acres
and a half which make two sulungs and a half[1596].' Thence, a small
quantity being neglected, the inference has been drawn that the Kentish
sulung was composed of 160 acres, while some would read '400 acres and a
half' to mean 450 acres and would so get 180 acres for the sulung[1597].
But the entry deals with one particular case and it connects real acres
with rateable units:--the canons have 400-1/2 or more probably 450
acres, which are rated at 2-1/2 sulungs. If we passed to another
estate, we might find a different relation between the fiscal and the
real units. Kent was egregiously undertaxed and as a general rule its
fiscal sulung will have many real acres. Turning to the cases in which
the geldability of land is expressed in terms of sulungs and acres, or
yokes and acres, we can gather no more than that the sulung is greater
than 60 acres, so much greater that '3 sulungs less 60 acres[1598]' is a
natural phrase, and that the half-sulung is greater than 40[1599] and
than 42 acres[1600]. We may suspect that the Exchequer was reckoning 120
(fiscal) acres to the sulung but can not say that this is proved.
[Kemble's theory.]
And now we must glance at certain theories opposed to that which has
been here stated. Kemble contends that the hide contained 30 or 33 Saxon
which were equal to 40 Norman acres, and that the hide of Domesday Book
contains 40 Norman acres[1601]. Now in so far as this doctrine deals
with the time before the Conquest, we will postpone our judgment upon
it. So far as it deals with the Domesday hide, it is supported by two
arguments. One of these is to the effect that England has not room for
all the hides that are attributed to it if the hide had many more than
30 or 40 acres; this argument also we will for a while defer. The
other[1602] is based on a single passage in the Exeter Domesday relating
to the manor of Poleham. That entry seems to involve an equation which
can only be solved if 1 virgate = 10 acres. William of Mohun has a manor
which in the time of King Edward paid geld for 10 hides; he has in
demesne 4 H., 1 V., 6 A. and the villeins have 5-1/2 H., 4 A.[1603] Now
three or four such entries would certainly set the matter at rest; but a
single entry can not. By way of answer it will be enough to say that the
very next entry seems to imply an equation of precisely the same form,
but one that is plainly absurd. This same William has a manor called
Ham; it paid geld for 5 hides; there were 3 H., 8 A. in demesne and the
villains had 2 H. less 12 A. Shall we draw the conclusion that 5 H. = 5
H. - 4 A.? The truth we suspect to be that here, as in Middlesex,
geldable units and actual areal units have already begun to perplex each
other. Both Poleham and Ham are what we call 'over-rated' manors. It is
known that Poleham contains 10 hides and Ham 5 hides, but, when we come
to look for the acres that will make up the due tale of hides, we can
not find them; for let King William's officers have never so clear a
terminology of their own, the country folk will not for ever be
distinguishing between 'acres _ad geldum_' and 'acres _ad arandum_' But
be the explanation what it may, we repeat that the one equation that
Kemble could find to support his argument is found in the closest
company with an equation which when similarly treated produces a
nonsensical result. This is all the direct evidence that he has produced
from Domesday Book in favour of the hide of 40 acres. Robertson, while
holding that the hide of Mercia contained 120 acres, adopted Kemble's
opinion that the hide of Wessex contained 40 without producing any
witness from Domesday save only the passage about Poleham[1604]. Eyton
reckons 48 'gheld acres' to the 'gheld hide,' but he leaves us utterly
at a loss to tell how he came by this computation[1605].
[The ploughland and the plough.]
Another theory we must examine. It is ingenious and, were it true, would
throw much light on a dark corner. It starts from the facts disclosed by
the survey of the East Riding of Yorkshire[1606]. In that district, it
is said, the number of carucates for geld that there are in any manor
(this number we will call _a_) is usually either equal to, or just twice
the number (which we call _b_) of the 'lands for one plough,' or, as we
say, teamlands. Further, it can be shown from maps and other modern
evidences that the manors in which _a = b_ were manors with two common
fields, in other words, were 'two-course manors,' while those in which
_a = 2b_ were manors with three common fields, in other words were
'three-course manors.' The suggested explanation is that while the
teamland or 'land for one plough' means the amount of land that one
plough will till in the course of a year, the 'carucate for geld' is the
amount of land which one plough tills in one field in the course of a
year. Manor _X_, let us suppose, is a two-course manor; the whole amount
of land which a plough will till there in a year will lie in one field;
therefore in this case _a = b_. Manor _Y_ is a three-course manor; in a
given year a plough will there till a certain quantity of land, but half
its work will have been done in one field, half in another; therefore in
this case _a = 2b_.
[The Yorkshire carucates.]
Now we must own to doubting the possibility of deciding with any
certainty from comparatively modern evidence which (if any) of the
Yorkshire vills were under a system of three-course culture in the
eleventh century. In the year 1086 many of them were lying and for long
years had lain waste either in whole or in part. Thus the first group of
examples that is put before us as the foundation for a theory consists
of 15 manors the sum of whose carucates for geld is 91-1/4 while the sum
of the teamlands is 91-3/4. What was the state of these manors in 1086?
Three of them were absolutely waste. The recorded population on the
others consisted of four priests, one sokemean, eighty-four villeins and
twenty-six bordiers; the number of existing teams was 35-1/2; the total
_valet_ of the whole fifteen estates was £7. 1_s._, though they had been
worth £72 in King Edward's day[1607]. It is obvious enough that very
little land is really being ploughed, and surely it is a most perilous
inference that, when culture comes back to these deserted villages, the
old state of things will be reproduced, so that we shall be able to
decide which of them had three and which had two fields in the days
before the devastation. Further, we can not think that, even for the
East Riding of Yorkshire, the figures show as much regularity as has
been attributed to them. In the first place, there are admittedly many
cases in which neither of the two equations of which we have spoken (_a
= b_ or _a = 2b_) is precisely true. We can only say that they are
approximately true. Then there are other cases--too many, as we think,
to be treated as exceptional--in which _a_ bears to _b_ some very simple
ratio which is neither 1:1 nor yet 2:1; it is 3:2, or 4:3, or 5:3.
[Relation between teamlands and fiscal carucates.]