or the Continuity of British Archaeology
A ' DEW-POND is an impervious or watertight hollow
artificially excavated on a hill-top, deriving its main supply of water from
dew, fog, and mist. It is obvious that every pond must receive a greater or less
amount of rainfall, but a real dew-pond does not depend chiefly on that source.
Neither is it largely fed by surface drainage, least of all by springs.
The name dew-pond is in common use, and is favoured by writers on the subject ;
nevertheless, it is partly a misnomer. Dew is an inadequate factor in keeping
the ponds full, it is more or less universal throughout a given area, and it
affects lowland ponds as well as those of the uplands. Correct notions on this
point are desirable, especially if we accept the modern theory of the formation
Dr. John Aitken, in an address delivered before the Royal Society of Edinburgh
in 1885, described a series of elaborate and careful experiments made by him in
order to ascertain the origin and nature of dew. His conclusions were partially
opposed, or at the least supplementary, to those of Wells, whose theory of dew,1
propounded in 1814, had been accepted by all subsequent writers. Dr. Aitken
claimed, with some reason, to have demonstrated that the greater part of a
deposit of dew is formed of night vapour that has just risen from the ground,
and has been trapped and condensed by grass and other objects in its course
up-wards.2 This theory, supported by many careful experiments, is, at any rate,
worthy of careful consideration, for, if correct, it would imply that a portion
of the water of a dew-pond is derived from the pond itself by previous
evaporation. To this extent the popular idea of the source of supply of
dew-ponds is plainly erroneous.
Dew, as we shall shortly see, adds its quota to the supply of the pond, but if
this moisture first rises as vapour—a vapour which from accidental
circumstances, such as the advent of a slight breeze, may not be laid down as
dew—a more comprehensive name must be sought for the ponds. Dew is believed to
be deposited without the aid of a cloudy intermediary, the vapour from which it
is condensed existing in an invisible state. Again, dew and fog may occur
together, and when there is fog hanging over the hills there is generally dew on
the grass. The heaviest dew, however, would not deposit from one to two inches
of water on the grass, an amount which has been recorded as the rise in level of
a dew-pond in a single night.
Mr. G. Dines, as quoted by Mr. E. A. Martin, concluded, as the result of
observations, that the average annual deposit of dew on the surface of the earth
actually falls short of 1.5 inches (see footnote to Table of References, Chapter
XIV). Moreover, high downs are much exposed to wind, and dew is with difficulty
formed under windy conditions, because the moving air does not remain long
enough in one place to allow condensation to take place.
In Wiltshire, I have observed that the peasants talk of a ' mist-pond '. An old
labourer of eighty, who, until five or six years ago, lived near Worms Heath,
Surrey, applied the same term to the pond there. ' Mist-pond ' is also reported
from Kent.3 Mr. F. J. Bennett has suggested ' mist-and-rain-pond ', which has
the merit of preciseness, but the demerit of length. Even then, ' mist ' must be
held to include ' fog ' also. In any event, while we continue to use the name
dew-pond, we should remember its draw-backs.
As to geographical distribution, the dew-ponds occur most generally on the Chalk
Downs of the South and South-East of England. They abound in Sussex and Hants,
and are not uncommon in Berks. and Wiltshire. The example illustrated (Fig. 27)
is from Surrey. No dew-ponds seem to have been reported from Hertfordshire, but
structures by some considered quite analogous, have been noticed on other
geological formations in Warwick and Suffolk. Records have been also garnered
from the South of Portugal.4 Marshall, in 1798, noticed Yorkshire ponds which
had some likeness to our Southern ones.5 The Lincolnshire wold farmer of to-day
secures reserves for a dry season by the construction of a huge tank of
galvanized iron. Much further information and more extensive records concerning
distribution are desirable, but meanwhile it is essential to remember,
especially when we discuss the question of age, that absence of dew-ponds in one
district and presence in another tells neither way. Other places, other modes ;
and this diversity may be as true of water-supply as the variety of contemporary
dwellings, customs, and occupations.
The facts about the efficiency, of these ponds may be simply stated. Dew-ponds
preserve a supply of water on the hills when the ponds in the valley have
failed. A dew-pond near Inkpen Beacon (Hants), 900 feet above the sea-level, is
never dry, though it waters a large flock of sheep.6 Another pond, situated near
the celebrated Chanctonbury Ring, on the Sussex Downs, contained after the dry
season of 1899 several thousand gallons of water, whilst the ponds on the plain
below ran dry.'
Gilbert White, who was much interested in this subject, states that in May 1775,
when ponds in the vale, which were of considerable size, dried up, the smaller
ones on the hill-tops above Selborne were little affected. One pond, at an
elevation of 300 feet above White's house, was only 30 feet in diameter, and at
no time more than 3 feet deep in the middle ; yet it was never ' known to fail
', even during severe droughts. Yet that pond constantly afforded drink for
three or four hundred sheep, besides twenty head of large cattle.8
To give a recent instance : four or five years ago, after a spell of dry
weather, the strange sight could be witnessed, in a Southern county, of horses
and carts climbing a dry chalk hill to fetch supplies of water.9 Mr. E. A.
Martin has humorously conjectured that such a proceeding may be the origin of
the immortal Jack and Jill,10 and there is no impossibility in the idea. Again,
a small ' sedge ' of herons has been seen to desert the parched lowlands and
live for some weeks near an upland dew-pond, finding therein frogs, snails, and
aquatic insects. Thus the words of Rudyard Kipling are scientifically correct
Only the dew-pond on the height,
Unfed, that never fails
if a little poetical licence be allowed for unfed ' (not
visibly fed). At first thought, the replenishing of a pond during a drought, and
the actual diminution which sometimes occurs during a moderately rainy season,
are startling and paradoxical facts.
Before going further, it must be stated that many dew-ponds are known to be of
modern date, and it is with the construction of these that we shall next deal.
The method of formation has been well described by Messrs. A. J. and G. Hubbard
in their interesting book, ' Neolithic Dew-Ponds and Cattle-Ways.'
We are told that there still exists at least one wandering gang of men who make
the construction of dew-ponds their special work, just as in Mediaeval times
there were itinerant bands of masons and bell-founders. The writers inform us
that at Alfriston, Sussex, a family has been proficient as dew-pond makers for
three or four generations. The pond-makers select a dry soil, remote from the
smallest rivulet, which would be detrimental to the highest success.11
The site having been selected, earth is scooped out, much in excess of the
requirements. A coating of dry straw is laid over the bottom. Then follows a
layer of well-chosen, finely-puddled clay, and over this, again, stones are
closely strewn. Lastly, the margins are effectively protected by well-beaten
clay.12 The writers lay stress upon the dry straw as an essential part of the
structure. Independent inquiry, however, shows that the use of straw is far from
being general. I have even heard its employment strenuously denied, but, as with
other sweeping statements, some qualification must be made. Where straw is
used—if it be still used, it doubtless assists the puddling process, by
toughening the clayey floor, and by permitting of expansion and contraction
without any cracking of the material. Unhappily for theories, the premiss is not
always clear. Messrs. Hubbard make the straw to be the bottom layer ; other
writers place it as a middle bed.
Mr. Clement Reid, who has made a careful study of the dew-ponds of South-Eastern
England, says that he does not think that they are so scientifically made as is
generally reported. Some dew-ponds do dry up during droughts. Farmers are
continually making new ones, and sometimes, but only by accident, hit upon a
satisfactory site.13 Mr. E. A. Martin has recorded that the famous Selborne
pond, already mentioned, has been empty within late years.14
Mr. H. P. Slade, in a short treatise published in 1877, gave a number of
practical hints on the construction of dew-ponds, which, however, for reasons to
be afterwards given, he preferred to call Artificial Rain-Ponds '. A drawing,
illustrating the author's recommendations, is given in the pamphlet. Mr. Slade
insists that the slopes of the pond should be non-absorbent, else four-tenths of
the supply will be lost.15 All the weeds are not to be destroyed, as some are
required to keep the pond pure. Incidentally, we are told that sheep, if left to
themselves, prefer the water of dew-ponds to that of the colder springs16 A
well-made pond, according to Mr. Slade, remains effective for a long period ;
one particular pond, made in 1836, lasted and was serviceable until 1876, when
it had to be repaired17
By all authorities it is agreed that worms, and the trampling of sheep and oxen,
are the great enemies of dew-ponds, since these agencies tend to disintegrate
the puddled floor. Richard Jefferies, describing a dew-pond—' a broad, circular,
pan-like depression, partially filled with water '—says that when it was made a
layer of soot was put down to repel worms and grubs.18
Much to our loss, White of Selborne gives no account of the making of a
dew-pond, but Marshall, who wrote a little later (1798), speaks of the Drinking
Ponds ' of the Surrey Downs and describes their construction. Each bason ', as
he calls it, was lined with a coat of chalk, six or eight inches thick, beaten
down with rammers. Next, a grout or batter, com-posed of pounded chalk and hot
lime, was evenly spread over the hollow, to a thickness of half an inch. Two or
three other coats were then usually applied, to ' glaze ' the bottom and prevent
the ravages of worms. Marshall adds that these sheep ponds should be railed in
so as to keep off cattle and horses.19 The method just described differs largely
from that given by Messrs. Hubbard, but a survival, or recurrence, of this mode
of construction is seen in the cemented or concrete-faced ponds now becoming
common. Mr. C. J. Cornish records a dew-pond with a concrete bottom, situated on
the Brightstone Downs in the Isle of Wight, which has never run dry for thirty
years. Edward Lisle (d. 1722) recommends that artificial ponds should have four
equal slopes, and should be ' covered with a gravel or a mortar earth, four or
five inches on the tops, which, cattle treading in, will cement with the clay,
and bind . . .' This idea of the results of treading is erroneous, but Lisle
also advised that the pond should have ' shade ' to prevent the cracking of the
Among shepherds it is a very general maxim that the pond should have one or two
overhanging trees to aid in condensing the fog. Mr. Clement Reid also says that
the pond should be shaded by a stunted thorn or oak, or a holly bush. Failing
these, the South bank should be cut sufficiently deep to screen off much of the
Another maxim is, that when a dew-pond is constructed near the spur of a hill it
should have the slope towards the spur, so that a part of the rain-wash may be
intercepted. This device would, of course, put the pond outside the strict
definition, though Mr. Baldwin Latham has asserted that there is no dew-pond
which does not receive a certain amount of surface drainage.21 On the other
hand, the Rev. J. Clutterbuck, in his prize essay on ' Water-Supply ', states
that the highest elevations are usually selected so that there shall be no
accession of surface water.22 This advice was anticipated by Thomas Davis in his
volume on the agriculture of Wiltshire (1811). The supplementary water, in any
case, must not be allowed to make a forcible inrush. Mr. Anthony Collett,
writing evidently from personal knowledge, says that even a rivulet entering the
pond renders it useless by destroying the virtues of the clay crust and the
coating of straw.23
Just noticing, as we pass, the allusion to straw, one may observe that the
supply from rain-flow would be wash ', that is, mud and water, and this, in
time, would tend to choke up the hollow. The best dew-ponds are said to be those
of the Sussex Downs, near the sea ; these ponds get the full benefit of
sea-fogs, and the first draught of the moist South-West winds.
Let us next see if there is unanimity concerning the source of the bulk of the
water of these ponds. In other words, are they Mist-ponds or Rain-ponds ? Mr.
Slade, in the little treatise before quoted, boldly avers that Rain is the main,
if not the whole source of the pond's supply ' 24 And again, Heavy rains
increase the value of the rainfall '—a safe statement, surely—' while minor
neither increase nor diminish it.' He appends numerous tables of observations
upon which his conclusions were based, as well as some graphic representations
of the fluctuating supply. In view of more recent experiments, one is driven to
believe that there must have been some serious error which vitiated most of Mr.
Slade's experiments. Indeed, at the end of his tractate he seems to reveal a
little misgiving on his own part, for in a footnote this remark occurs, The
annual condensation [presumably by the water of the pond] will be found to bear
but a small ratio to the annual rainfall.' 25 Such a statement cannot be
accepted as axiomatic. It does admit the agency of fog, but depreciates, on
inconclusive evidence, as will be shown, its total effect. It should also be
added that Mr. Slade's experiments appear to have been concerned chiefly, if not
entirely, with one pond.
The more recent experiments, to which allusion has just been made, were carried
out on the suggestion of the Rev. J. G. Cornish, at Lockinge, on the Berkshire
Downs. When a heavy dew was predicted by the local shepherds, a notched stick
was thrust into the pond overnight, care being taken that the notch was level
with the surface of the water. The stick was removed next morning, and a fresh
notch cut, to show the water-mark. On the night of January 18, 1901, a rise of
1½ inches was recorded, on the following night 2 inches, and a little later,
January 24, there was an increase of another inch. During one particular fog,
occurring even in May, the level rose 1½ inches. Five nights of winter fog gave
a total accession of 8 inches. During five days of heavy spring dew, with no
accompaniment of fog, occurring in April and May, an increase in the same pond
of 3½ inches only was registered. This amount, if truly produced by dew alone,
is immensely greater than estimates would generally warrant. Though one of the
dews was heavy, it seemed clear that the dew was a less important agent than
mist.26 It appears to be also a just inference that there must have been
atmospheric movement to transfer the vapour to the dew-pond area, and since the
old theory taught that a calm air is essential to the deposition of dew,
Aitken's teaching about rising vapour seems here to get support.
The results obtained by Mr. Cornish, though at first sight startling, will not
altogether astonish those who have had occasion to spend several hours in a
fog-laden atmosphere on a moderately high hill. Tennyson had some insight
touching this matter, though, with pardonable freedom, he ascribes the moisture
to dew alone :-
Calm and deep peace on this high wold,
And on these dews that drench the furze,
And all the silvery gossamers
That twinkle into green and gold."
Mr. J. H. Shorthouse, in ' John Inglesant ', brings out this
phenomenon very aptly, writing as he does obviously from experience. And Gilbert
White; to revert once more to the sage of Selborne, noticed that shepherds, and
fishermen, and ' persons that are much abroad, and travel early and late ', have
good reason to know these ' prodigious fogs '. Even in the hottest parts of
summer, ' the surfaces of things are drenched by those swimming vapours, though,
to the senses, all the while, little moisture seems to fall.' 28
Mr. Clutterbuck gives calculations for ascertaining the volume of the unseen
supplies of water, but as he assumes an average daily amount to be consumed by
each member of the flock, his conclusions are necessarily faulty, and of vagrant
It will, however, be worth while to step outside the strict limits of our
investigation, and ask what explanations of the principle of dew-ponds have been
put forward. This done, we must return to our real inquiry, the age and origin
of the ponds.
Dr. Stephen Hales, we are informed by White, settled the matter for himself
promptly and decisively by asserting that ' more than a double quantity of dew
falls on a surface of water than there does on an equal surface of moist earth '
29 This dictum avoids the real difficulty, but even as it stands it cannot be
accepted, since the rate of cooling, and consequently the amount of
condensation, depends largely on local circumstances 30
Messrs. Hubbard's elucidation of what they call the ' thermodynamics ' of a
dew-pond invokes the agency of the straw-lining already mentioned. During a warm
summer's day the earth stores up a considerable amount of heat, ' while the
pond, protected from this heat by the non-conductivity of the straw, is at the
same time chilled by the process of evaporation from the puddled clay.'
Consequently, during the warm night the moisture of the comparatively warm air
is condensed on the surface of the colder clay. ' As the condensation during the
night is in excess of the evaporation during the day, the pond becomes, night by
night, gradually filled.' 31
As I understand this explanation, it could only apply to the first filling of
the pond. Afterwards there would be no ' cold clay ' exposed, save at the
margin. The writers proceed thus : The dew-pond will cease to attract the dew if
the layer of straw should get wet, as it then becomes of the same temperature as
the surrounding earth, and ceases to act as a non-conductor of heat. This,
practically, occurs if a spring is allowed to flow into the pond, or if the
layer of clay (technically called the " crust ") is pierced.' 32
Surely the straw must ultimately become wet under any conditions, be the clay
rammed never so tightly. Springs may be neglected, as being unusual. Would it
not be possible for the straw to get sodden by hydrostatic pressure from below ?
We remember, too, that some authorities make the straw the middle layer, not the
bottom, as in Messrs. Hubbard's plan. Moreover, as a fact, the tread of animals
penetrates the clay crust, and the pond, though impaired in efficiency, yet
continues to serve for years. What, too, of those ponds, evidently the majority,
in which straw is not used ? Neither Marshall nor Jefferies mentions straw ;
inquiry to-day traces it only doubtfully in most districts. Theories based on
the non-conductivity of this material could therefore, at most, give only a
partial solution ; where the straw is absent, they are worthless. Dry straw,
indeed, as already noted, is said by Mr. Collett to be employed by the Berkshire
craftsmen when lining the pond at the present day.33 Clutterbuck also mentions
straw, though he describes its purpose differently. Speaking of the making of a
pond, he remarks, ' As the portions are finished, they are protected from the
action of the sun and atmosphere by a covering of straw.' 34 This seems to refer
to a temporary loose bed of straw placed over the crust until nightfall arrives
and the pond is to start its history.
Clutterbuck tells us that when all is finished water is introduced by artificial
means. If obtainable, he advises piled-up snow as the readiest and least
expensive method of securing a start ; wattled hurdles should be erected to
encourage, during a snowstorm, the requisite drifting.35 A century ago, Davis
recorded this filling with snow ; the after-supply, he asserted, is kept up by
clouds and dew.
In the neighbourhood of Calcutta ice is obtained by placing porous earthenware
pans of water in an earth cavity lined with rice straw. The straw seems here to
have some value.
Another theory supposes that fog and mist collect around minute particles of
dust, and these tiny spherules are more readily attracted by a sheet of water
than by any other surface 36
A third hypothesis, propounded in 1900 by Mr. A. Marshall, teaches that there is
a difference of electric potential, and therefore an attraction is set up
between the innumerable particles of fog-water and the summit of the hill on
which the pond is situated. Mr. Marshall suspended two porcelain basins, alike
in all respects, by silk threads over a spot on the South Downs. An upright
sheet of copper was placed in each basin. One basin was allowed to remain
insulated by the silk thread, but the other was earth-connected by wire. During
the night of April 1, 1899, when a thick mist prevailed, the bathe the insulated
' screen ' received 155 c.c. of water, the other, with the earth-connected
screen, collected 18.0 c.c. The insulated apparatus tended to get an electrical
charge from the particles of moisture, therefore the attraction was less than in
the apparatus earth-connected. The result was the more remarkable, as the silk
soon became saturated, rendering the insulation imperfect.37 I quote Mr.
Marshall's account, adding the simple comment that it seems to furnish a
probable, though only partial, explanation of the phenomena.
Professor L. C. Miall gives an explanation which is here summarized.
Moisture-laden winds, rising from lower levels, become chilled by expansion, and
the hitherto invisible moisture is thrown down, either as clouds which gather
around the hill-top, or as dew. Thus there is an excess of condensation over
evaporation owing to height. This is the first factor. Then the pond itself acts
as a powerful ally in the following manner. The surface-layer of the water cools
by radiation, becomes denser, and sinks. Continual replacement of the
surface-layer by the rise of the warmer portions is set up, and this continues
until the water is cooler than the surrounding rocks and earth. At this stage
the pond becomes an efficient condenser of vapour. Professor Miall believes that
a part, and probably a large part, of the pond's supply is in this way derived
from the invisible moisture of the atmosphere.38
Some writers have argued that the presence of a tree counts for nothing, because
the increase of water is not confined to ponds so shaded. We note, however, that
the tree is planted as a supplementary agent only. Mr. Reid speaks of ponds
which, on a hot summer's day, contain only a few inches of water, insufficient,
seemingly, to last a week. Towards evening, or at early dawn, a sea-mist drifts
in, and there is a continuous, steady fall of moisture from the smooth leaves of
the overhanging tree. The dew-pond at Marden Park, as shown in Fig. 27, is well
furnished in this respect. In town, also, one sometimes sees the pavement under
an avenue of planes or poplars quite wet from the incessant drip of moisture,
whilst elsewhere the dust lies like dry powder.39
To test the ' distilling ' power of trees, Mr. C. J. Cornish placed two vessels
out of doors on a dewy evening. One was put under a cherry-tree in full leaf,
and the other on stone flags away from trees. The latter vessel would be
affected by dew and by the more limited condensation only, the former received
the drippings from the cherry-tree. In the morning the vessel under the tree was
found to contain twice as much water as the other.40 In the Indian jungle, amid
dews and mists, water patters from the trees like rain. Briefly, the
predilection of the flock-master for a co-operating tree seems to be justified
by results, and we may conclude that trees enhance the efficacy of the pond.
A few additional considerations are worth submitting. Wells cites experiments
performed by Six, showing that there is a fall in temperature of 1° F. for an
ascent of 250-300 feet." Closely connected with this fact, no doubt, is the
increase of rainfall in elevated regions, amounting to 3 or 4 per cent. for
every 100 feet of rise 42 Again, as noted by Miall, the evaporation on high
grounds, owing to the diminished temperature, is less than on low lands." There
are times, too, as Baldwin Latham observes, when the water in a pond at great
elevation seems to be in direct communication with the clouds, though the sky
around is quite clear."
The theoretical aspect of dew-ponds has no* been treated with a fullness which
may be pardoned in view of the scanty and little-known references to the subject
in our scientific literature. We must now return to the high road of our mental
We have seen that some of the dew-ponds are modern constructions. We now ask,
Are all dew-ponds of recent date ? Mr. Slade, referring to examples in his own
district of Wallingford, asserts that dew-ponds were not known there till about
the year 1836.45 Thomas Davis notices them in Wiltshire thirty-five years
earlier. Mr. Clutter-buck states that in one known instance the increase in the
number of dew-ponds has been a matter of definite arrangement between landlord
and tenant.46 This, of course, merely proves that dew-ponds are made at the
The reverse evidence is stronger. A keen Wiltshire archaeologist, able to
estimate the value of that befogged folk-memory which can always recall ' the
beginning of things ', tells me that he considers the ponds on the Wiltshire
Downs to be far older than the date 1836. Mr. Collett speaks with a like voice ;
the Wiltshire ponds, he says, outrun all memory and tradition. In the eighteenth
century they were used to water the huge flocks of sheep that were driven over
the downs of Wiltshire, Hants, and Dorset on their way to the great fortnightly
fairs.47 The tracks along which the flocks went may still be traced in our '
Drives ' and ' Drove Roads ' of the Southern counties. Vancouver speaks of
dew-ponds in Hampshire in 1813. They were constructed at great labour and
expense to catch the downfal (sic) waters ', retentive clay and flint facings
Gilbert White, a careful writer, gives no hint that the dew-ponds were an
innovation of his day. White was born in 1720, so we might reasonably surmise
that dew-ponds were in existence during the first half of the eighteenth
century. Indeed, we have a corroborative instance to hand. It is given in a
curious work written by Dr. Christopher Packe, of which the following is a part
only of the title : ' Explanation of a new Philosophico-Chorographical Chart of
East Kent ' (1743). He describes a large pond, situated on the top of Collier's
Hill, near Brabourne, in Kent, ' an hundred perpendicular feet ' above the
springs below, with which it does not communicate ' except in a very flush time
of water '. At such periods it overflows by an artificial cut in the lip of the
pond. The pond is never empty, even after a dry spring and summer. Packe is
puzzled by this phenomenon, but he gets very near the truth, for he believes
that the pond is supplied by the ' condensation of elevated vapours '.49 Whether
the pond is a natural or an artificial one is not quite clear, although
artificial provision has been made for an overflow. I find that the top of
Collier's Hill belongs to the Hythe Beds of the Lower Greensand, underneath is
the outcrop of Atherfield Clay, while the foundation is of Weald Clay. The
springs referred to presumably emerge at the junction of the two first-named
groups of strata. The perennial character of the pond's supply is said to be
Beyond the early eighteenth century, records do not seem to be available.
Folk-memory is wanting, or is, at any rate, bemused by educated influence. If
the dew-ponds actually existed at any period previous to the ' wool-growing '
centuries—the fifteenth and sixteenth—their introduction perhaps followed a
disconnecting gap. Something of this kind, we have seen, appears to have
occurred in connexion with linchets and chalk-wells. It is true that
agricultural methods change slowly, and there is a possibility that a few of our
extant dew-ponds are survivals from early times, and have been again and again
repaired. Bearing in mind the shifting of the old population from the hills to
the vales as the woodlands were cleared, I should nevertheless doubt whether the
present ponds antedate the period of the Mediaeval sheep-masters. Valley
pastures are usually well watered, but when the wolds and downs again came into
use as sheep-walks, ponds would be essential accessories.
Yet there is a likelihood that dew-ponds of some kind were known to our
ancestors of the Neolithic and the Bronze Ages. First, there is forced upon us
an a priori probability, amounting almost to conviction, that excavations of
this nature must have been known and utilized. Water was needed. Depressions are
discoverable in the ground near old settlements, and are not always of the
The primitive settlers of the hill-tops, who have bequeathed to us their camps,
barrows, and cultivation terraces, had, above all things, an imperative need of
a constant water-supply. To state an example : Within an area of 160 square
miles on the South Downs, Mr. A. H. Ahicroft has counted sixteen prehistoric
forts, belonging to the Neolithic and Bronze periods ; the number was doubtless
once greater. In the absence of direct and positive evidence, we are justified
in pursuing the exhaustive or eliminative method of inquiry.
We will consider the case of Winklebury Camp, near Basingstoke. Its age is
rather uncertain, but there is a long barrow on its boundary, and the relics
already yielded seem to indicate a late Neolithic origin. This opinion has
lately been expressed, as a result of careful exploration, by Dr. C. H. Read, of
the British Museum 51
Years ago, when General Pitt-Rivers was examining Winklebury, he was struck by
the absence of any visible signs of a water-supply. The nearest place where
water can be obtained in modern times is at the foot of the hill, 850 yards to
the West of the camp.52 A spring on the plain below is traditionally said to
have been used in the Middle Ages. Could such sources have been satisfactory in
Neolithic times ?
We remark, first, that the interior division of the camp was considered by
Pitt-Rivers to have been reserved for cattle. The theory that such forts as
Winklebury were refuges for men and their herds in times of strife may be indeed
taken as rightly well-trusted. But it must not be overlooked for an instant that
ancient herdsmen would require water for their flock at all times, in peace or
in war. Water must therefore have been found for oxen and sheep as well as for
Setting aside the need of adequate supplies for grazing animals, let us admit
that a permanent investment of a refuge-fort was an unlikely event. We still
have to meet the fact that neither man nor beast can live many days without
water, much less could warriors so deprived keep up physical strength for
defence. Nor would the quantity of water required be small.
Did the inhabitants carry into the camp a supply sufficient for two or three
days ? This is doubtful ; first, a hostile inrush would permit little
preparation, and secondly, primitive skin bottles and earthenware pans would
hardly meet the demand. No ; rather is it a modern question which says, Whence
shall we bring water? Primitive man went where the water existed, or could be
readily obtained. We may be sure that not even under the severest stress would
he live on the hill-tops unless he could easily satisfy this first necessity.
Did wells, yet undiscovered, exist within the entrenchments ? We know that
Normans, Saxons, and Romans were in the habit of sinking wells. Yet few Roman
examples have been discovered on the Chalk Downs. I think that the reason is
obvious. The requisite depth of bore would be almost prohibitive of attempts at
this kind of work.
In his long and painstaking explorations at Cranborne Chase, General Pitt-Rivers
discovered no wells on the high ground near the crest of the hills at Rotherly
(Wilts.). On the lower levels, at Woodcuts, over the Dorset border, the spade
exposed two wells, one of which was proved to be of Roman date by the finding of
the remains of a Roman bucket at the bottom. This well was 188 feet in depth.53
Holes or ' put-logs ', 12 inches apart, were cut into the sides of the shaft."
The shallower well, driven to a depth of 136 feet only, showed no sign that
water had been reached.
A modern Woodcuts well, sunk from a level 77 feet below the site of the ancient
borings, yielded no water until the rock had been drilled to a depth of 211
feet. And to-day, the only sources of water in the Chase, wherever there are no
wells, are dew-ponds and the roofs of buildings. General Pitt-Rivers thought it
probable that at Winklebury there may exist wells hitherto unrevealed to the
investigator, but, on the whole, he inclined to settle the matter by postulating
a much higher water-level in ancient times.
Should wells ever be discovered at Winklebury, they will most likely be of
Romano-British date. The question is not, however, Did the Romans and
Romano-Britons bore wells ? but, Did the earlier folk, equipped with rough
spades of bronze and flint and mattocks of deer-horn, essay such a task ? Until
we conceive this probable, and can assume that the tribesmen knew that water
existed at great depths—until, in short, some Neolithic wells are produced in
evidence—we may well hesitate to speak affirmatively. If the dene-holes should
be cited as parallel constructions, we would answer that the true dene-hole is
seldom even conjecturally Neolithic, or even of Bronze Age. While dour compels
one to admit that the tasks of sinking andour a dene-hole and a well
respectively are somewhat fairly equated, yet so long as a simple solution
offers itself, we need not search for a complex one.
As to the ancient water-level of the chalk, was it sufficiently high to furnish
springs close to the besieged camp ? By most writers it is agreed that the
British climate of the Stone and Bronze periods was much more humid than at
present. The felling of primeval forests, the introduction of artificial
drainage, the rise in the mean winter temperature, have, since those days,
diminished our annual rainfall. We have to remember the positive testimony of
Pytheas concerning the dampness of our atmosphere in early times. Geologists
believe that the intermittent bournes ' of the Chalk once ran more frequently.
Unfortunately, adequate data on these points are not procurable, but the main
argument stands. General Pitt-Rivers supposed that the well-level of the Chalk
was from 30 to 50 feet higher during the Roman period, so that springs would
burst out again at greater elevations.55
Remembering that the water-level of the Chalk, especially in the neighbourhood
of large towns, is sinking year by year, owing to both natural and artificial
causes,56 the existence of a higher level in Roman times can scarcely be
contested. Whether as great a difference as, say 40 feet, may be assumed, is
debatable. Again, a conclusion based on the comparison of one modern well with
one ancient well, without close knowledge of the dip of the Chalk, of possible
faults, of the presence or absence of swallows' and large fissures, is not
unassailable. But in the two cases given, it is not a difference in level of 40
feet, but of 100 feet, which has to be accounted for. (Woodcuts modern well, 211
feet + 77 feet 288 feet ; Roman well, 188 feet.) Taking the greater limit
actually asked for by Pitt-Rivers, namely 50 feet, such a rise in the
water-level would not materially affect camps constructed at such elevations at
Winklebury and Maiden Castle. Fifty feet will scarcely meet the difficulty, for
Mr. Clutterbuck states that the water-level of the Chalk varies from 100 to 400
feet beneath that of the ponds,57 whose elevation generally corresponds with
that of the camps. In the present state of our knowledge, the well theory must
be deemed untenable. On the other hand, if we consider that General Pitt-Rivers's
reasoning is valid, it will, while removing the necessity for dew-ponds, lend
one more testimony to the belief in the existence of ancient hill-top
Once more ; a spring or bourne generally leaves some trace in the form of a
gully or dry water-course, and this might actually be intensified after long
sub-aerial denudation. Some traces of this kind should remain, because turfed
ground is excellently adapted for the preservation of such features.
Not to lose sight of the chief question, it must be repeated that the camps are
largely pre-Roman, and we have to inquire how the inmates secured one of the
prime necessities of life. The query becomes more pertinent if we leave
Winklebury, and consider a ' contour fort ', like that of Cissbury, where we
know that there was a permanent settlement of men engaged in mining and working
flint. Not only was Cissbury, with a periphery of 1 miles, a settled camp, but
it was doubtless occupied continuously from day to day, for we cannot suppose
that the inhabitants climbed up the hill every morning and down in the evening,
and that they carried with them a daily supply of water. Yet this must be
conceded, if we reject the presence of wells or springs, unless indeed there is
some fourth explanation forthcoming.
There is such an explanation—the possibility that dew-ponds, or hollows of a
kindred nature, were utilized. The evolution of the dew-pond, first suggested, I
believe, by Mr. F. J. Bennett, is worth considering.58 The ditches of the
earthwork camps would be constantly trodden down, first by the excavators, and
afterwards by the members of the garrison and their impounded cattle. We have
only to suppose a surface of Clay-with-flints, or any other retentive clay, and
it is evident the fosses would soon contain water. Additional defence would thus
be provided, and, during warfare, perhaps a supply of drinking water, in
addition. Mists and dews would soon have free play, and the greater rainfall
would keep up the reserve. In the course of time, the trenches might be closely
rammed of set purpose. Finally, dew-ponds, imperfectly water-tight, no doubt,
might be designed within or near the fort.
The accidental puddling of trenches came under the writer's notice in April
1905, when examining the series of round barrows on Povington Heath, Dorset. The
weather had been dry for many days previously, and the thirsty heath, save for
an occasional bog, was much parched. Where, however, the trenches encircling the
barrows had a slightly clayey bottom, the water was held up. The trampling of
moorland sheep and cattle may have helped in the puddling, though in the modern
artificial dew-pond, where specially constructed layers or crusts are concerned,
such action would have a reverse effect.
There have also probably always existed, here and there, a few natural
dew-ponds, formed by depressions in the Clay-with-flints, or in Eocene clays.
It may be objected that prehistoric villagers could not have understood the
principle of dew-ponds. The objection is sound, but we reply, ' Neither would
they understand, except experimentally, why a clay-lined earth-cavity, or a
crude pan, would retain the water in which primitive folk are accustomed to
parboil their meat.' Puddling was resorted to in each case. Experience was the
only light, as it is to the average flock-master of the twentieth century. One
repeats, though it ought not to be necessary to say it at all, that practice
does not always involve a knowledge of theory.
The theory of the development of dew-ponds above sketched may or may not be
correct. But there is a coincident clue. Richard Jefferies, who shall be cited,
not as possessing authority, but as one who had shrewd native insight in matters
affecting agricultural economy, observed that two modern dew-ponds are situated
near isolated camps in Wiltshire. He was led to infer that similar ponds could
have furnished the only available supply to the ancient communities.59 It
is necessary to state that in this and other counties modern ponds are sometimes
found near old forts, but these instances of juxtaposition do not exhaust all
We may now reasonably ask whether some of the depressions in the vicinity of
ancient earthworks were not dew-ponds. The hollows may be dry, or they may be
choked with rank vegetation. They may, again, still be subject to swampiness, as
in the depression which I have elsewhere reported from Henley Wood, Chelsham
(Surrey).60 An oval hollow, quite dry, occurs a little distance away. Mr.
Bennett records a small dew-pond at Ladle Hill Camp, Berkshire, which is
probably of recent construction, but he also mentions several hollows, seen by
himself and Major Cooper King at Perborough Camp (Berks.), which may represent
ancient ponds.61 A pond situated towards the South of Chanctonbury Ring, Sussex,
in a direct line between that earthwork and Cissbury, is believed to be ancient,
though it has probably been cleared more than once. Two modern ponds are to be
seen in the neighbourhood. At the Eastern angle of the fort on Ditchling Beacon
(800-50 feet) there is a dew-pond, probably of ancient date. There is also, at
the North-west, what the historian calls a rain-water pond ', now dry,
considered to be coeval with the earthwork, because the ramparts have been
modified in order to induce supplies (Fig. 28).62 Another example may also be
traced hard by. Other dew-ponds, of unknown age, lie to the West.
Mr. T. W. Shore noticed dew-ponds near Hampshire camps of the Bronze Age, and
believed that this mode of obtaining water was known to the fortress-builders.
He found that the large camps were near wide, open, treeless areas, supposed to
have been unafforested from the earliest times. The hill-fortresses, he said,
have, at the present day, either within or near them, dew-ponds or cloud-ponds
', or they possess remains which attest the former existence of these
accessories. The smaller forts, found in the forests, are supplied by streams or
springs. The only instances of wells, two in number, discovered by Mr. Shore,
occurred in small camps of this nature. He also particularizes St. Catherine's
Hill, near Winchester, 327 feet in elevation, and requiring, according to
General Pitt-Rivers's calculations, 3,300 men for its defence. This number does
not take into account women and children. The water-supply, Mr. Shore thought,
was afforded by a permanent dew-pond which lies at no great distance, and which
to-day waters a large flock of sheep.63
Further examples can be given. Visitors to Maiden Castle, near Dorchester, may
see, within the actual fort, what was probably an ancient pond. Messrs. Hubbard
have described a dried-up dew-pond in St. Martinsell Camp, near Marl-borough."
They give other instances, some of which must be received under reserve, until
we know whether or not they are of modern construction.
Messrs. Hubbard also lay down several conclusions of vast importance should
these ever be firmly established. Some dew-ponds, the writers assert, are
fortified in a manner similar to that of Neolithic settlements. Signs of a
dwelling, of a kind like those found within the camp, are not infrequently
discovered in close proximity to the pond, and this kind of residence the
writers believe to have been a guard-house, where perhaps the cattle were
counted as they passed in and out of the rings. Some Neolithic earthworks seem
to have been built expressly to be in communication with the ponds.65
The only instance yet noted where the pond is within the camp is the
already-cited case of the Dorsetshire Maiden Castle. It is suggested that, in
the majority of cases, such a position would be shunned because of the trampling
At least one failure on the part of our primitive fathers is recorded, for at
Ogbury Camp (Wilts.) an unsuitable site was selected, with the result that the
pond never held water. Cattle-tracks run down to it, to reappear on the farther
side, as if the herds passed right through the hollow on their way to the river
below." The writers believe that the usual plan was to dig a pond in the very
middle of cattle-tracks-that is, tracks formed by the tread of cattle, as
distinguished from artificial cattle-ways—so as to intercept all paths but one.
There was thus no alternative for the cattle but to drink at the dew-pond on the
hills, and so keep away from the wolf-haunted lowlands. To this day the old
danger from wolves is recalled by a peculiar instinct found among the half-wild
cattle of Chillingham. The cows of this herd are said still to retain the habit
of concealing their calves.67
I have tried to present Messrs. Hubbard's theories with all possible fairness,
though I think that the writers go much farther than the data warrant. Before
these speculations can be heartily accepted, more must be known. Mean-while the
ideas offered to us stimulate thought and encourage a search for further
Facts and opinions may now be gathered and allowed to converge. That many
dew-ponds are of recent date admits of no question. Generally, the excavations
have a recorded history of two centuries only. Probably, however, they existed
from the fifteenth century onwards. They would not be needed in Norman and Saxon
times. The old mound-builders required a good water-supply, and there is no
available proof that the water-level of the Chalk was sufficiently high to feed
the down-tops ; a definite verdict on this matter must nevertheless be
suspended. Celtic farmers and their Neolithic predecessors cannot be proved on
present testimony to have been well-sinkers. Lastly, hollows of the dew-pond
kind would have served the purposes of the dwellers in the forts, though the
ponds could not have been dug on scientific principles, being partially
rain-fed, and occasionally failing of their purpose. And, indeed, all ponds
situated at great heights must be to some extent fed by showers of rain and
If we accept the prehistoric dew-pond, we need not rush to the extreme of
thinking it the only possible mode of obtaining water. On a hill of moderate
elevation, with a favourable concurrence of geological conditions, intermittent
bournes or lavants ' might serve very well. Advantage might be taken of natural
or artificial runnels whenever there was a downpour of rain, no unusual event.
Temporary camps or earthworks constructed for special purposes might be
intentionally planned near running streams. Thus, ancient pits, presumably for
habitation, once existed near a constant spring at Coombe Hill, Croydon,68
though no trace of either huts or spring is now to be found. Another most
important consideration is that away from the Chalk Downs there were often
alternative sources of supply. Hence, as with dwellings and settlements, there
may have been contemporaneous diversities.
The desertion of the upland camps and settlements led to new requirements and
new systems. The Teutons who tilled the valleys needed the dew-ponds little, if
at all. Springs and rivulets sufficed, otherwise wells were bored. When the
down-tops were once more required as sheep-walks, dew-ponds would again be a
necessity. Guided by experiment and observation, the Mediaeval husbandmen and
shepherds trod safely the path along which the ancients, with more or less
success, patiently groped. The folk-memory of ponds was, for the time,
rejuvenated. Once more tradition failed, and once more it was taken up in the
Actual continuity of dew-pond construction has not been proved, perhaps cannot
be proved. The same may be said of the very antiquity of the ponds. The
archaeologist would be pleased were the feat possible, but it is entertainment
enough to give reasons showing that the ancient and modern methods of the
husbandman and the shepherd probably approximate.
REFERENCES AND BIBLIOGRAPHY
*1. For old theory, see W. C. Wells, ' Essay on Dew,' 1818 (preface dated
1814). Also good summary, T. H. Huxley, ' Physiography,' 1885, p. 54.
*2. Trans. Roy. Soc. Edinburgh, xxxiii (1886), pp. 9-64. * Summary in Chambers's
Cyclopaedia, 1904, iii, p. 782.
3. Eng. Dial. Diet., under Mist-pond'.
*4. Rept. Brit. Assoc., 1900, pp. 585 et seqq. ; Nature, lxiii (1900), pp. 20-1.
*5. W. Marshall, Rural Econ. of the Southern Counties,' 1798, ii, p. 398.
Wiltshire dew-ponds are noticed in Wilts. Archaeol. and Nat. Hist. Mag., xvii,
1878, pp. 299-300.
*6. Spectator, Nov. 16,1901 (lxxxvii, pp. 755-6).
7. Nature, lxii, 1900, p. 495.
*8. G. White, Selborne ' (Harting's edition, 1880), pp. 225-6. Gideon Mantell,
in 'Geol. of S.E. of England,' 1833, p. 9, n., quotes White, and mentions the
natural stocking of ponds with animal and vegetable life.
9. Spectator, lxxxvii, p. 990.
10. E. A. Martin, in Knowledge, N. S., iii, 1906, p. 481.
*11. A. J. and G. Hubbard, Neolithic Dew-Ponds and Cattle-Ways,' 1905, pp. 2-4
(speak of wandering band). Stationary workers referred to in S.E. Naturalist
(1906), pp. xxxix, xl, 20.
12. Neolithic Dew-Ponds,' p. 3. *13. Nature, lxiii, 1900, pp. 20-1. 14.
Knowledge, N. S., iii, 1906, p. 481.
*15. H. P. Slade, A Short Practical Treatise on Dew-Ponds,' 1877, p. 30. 16.
Ibid., p. 12. 17. Ibid., p. 6.
*18. R. Jefferies, Wild Life in a Southern County' (new edition, 1889), pp.
*19. Rural Econ. of S. Counties,' ii, p. 398. Cf. E. Lisle, ' Observa-tions in
Husbandry' (ed. by T. Lisle, 1757), ii, p. 406.
*20. C. Reid, in Miall and Fowler's edition of White's Selborne,' 1901, p. 166,
21. Proc. Geol. Assoc., x, p. 377. Cf. Cobbett, Rural Rides,' i. p. 20.
*22. J. Clutterbuck, in Jour. Roy. Agric. Soc., i (2nd Ser., 1865), p. 273. Cf.
T. Davis, Wiltshire Agric.,' 1811, p. 13.
*23. A. Collett, in Macmillan's Mag., N.S., i, 1906, p. 217.
24. Slade, Treatise on Dew-Ponds,' p. 20. 25. Ibid., p. 31. *26. C. J. Cornish,
The Naturalist on the Thames,' 1902, pp. 129-30.
A capital summary given in the Spectator, lxxxvii, 1901, pp. 775-6. 27. In
Memoriam,' c. xi, v. 2.
*28. White's Selborne ', Harting's edition, 1880, p. 226. The damp, foggy
atmosphere of the Downs is well described in J. H. Shorthouse's John Inglesant
', 1887, p. 20.
29. Selborne,' p. 226.
*30. White's Selborne ', annotated by Miall and Fowler, 1901, p. 164. Rept.
Brit. Assoc. (1900), pp. 585 et seqq.
*31. ' Neolithic Dew-Ponds,' pp. 2, 3. 32. Ibid., pp. 3-4. 33. Macmillan's Mag.,
N. S., i, 1906, p. 216.
*34. Jour. Roy. Agric. Soc., 2nd Ser., i (1865), pp. 281-7.
35. Ibid. Cf. Davis, ' Wiltshire Agric.,' p. 13, and account in Wood- ward's '
Geol. of Eng. and Wales,' p. 421. (The Calcutta custom is recorded in Nature, v,
*36. Spectator, lxxxvii, 1901, p. 798.
*37. Nature, lxii, 1900, p. 495.
*38. White's Selborne ' (Miall and Fowler), pp. 165-6.
*39. Ibid., p. 166.
40. ' The Naturalist on the Thames,' pp. 130-1.
*41. W. C. Wells, ' Essay on Dew,' p. 225.
*42. Ganot, ' Physics' (12th edition, Atkinson, 1886), p. 961. 43. Ibid., p. 316
; White's Selborne' (Miall and Fowler), p. 165. *44. Proc. Geol. Assoc., x, p.
377. 45. Slade, ' Treatise,' p. 6. 46. Jour. Roy. Agric. Soc., 2nd Ser., i, pp.
*47. Macmillan's Mag., 1906, pp. 216-17.
*48. C. Vancouver, Agric. Survey of Hampshire,' 1813, i, pp. 47-8. * Cf. W.
Cobbett, Rural Rides' (1886), i, p. 20.
*49. C. Packe, Philosophico-Chorographical Chart of East Kent,' 1743, pp. 87-8.
50. Murray, ' Handbook to Kent,' 5th edition, 1892, p. 37.
*51. C. F. Dowsett, Winklebury Camp,' 1904, pp. 4, 7 ; also Times, Nov. 26,
*52. ' Excavations in Cranborne Chase,' ii, p. 237.
*53. Ibid., i, pp. 10, 27-8. 54. Ibid., ii, p. 237. 55. Ibid., i, p. 28.
*56. ' Neolithic Man in N.-E. Surrey,' pp. 25-6.
*57. Jour. Roy. Agric. Soc., i, N. S., p. 273.
*58. Trans. Newbury Field Club, iii, pp. 257-8.
*59. R. Jefferies, ' Wild Life in a Southern County,' 1889, pp. 25-8.
*60. ' Neolithic Man in N.-E. Surrey,' pp. 83-4.
61. Trans. Newbury Field Club, iii, pp. 257-8.
*62. Vict. Hist. Sussex,' 1905, i, pp. 460-1. The Sussex ponds are dealt with in
the abstract of a paper by A. H. Allcroft, Morley College Magazine (London),
xvi, 1906, pp. 40-3.
*63. Jour. Anthrop. Inst., 1889, xviii, pp. 336, 338, &c.
64. Cornhill Mag., No. 119, N. S., 1906, p. 617.
*65. ' Neolithic Dew-Ponds,' p. 17.
66. Ibid., pp. 60-1. 67. Ibid., p. 59. 68. King, Munimenta Antigua,' i, pp.
Note.-Since this chapter was written, I have seen Mr. E. A. Martin's excellent
articles in Knowledge, iv, N.S., 1907, pp. 108-9, 126-8, 197. Some slight
additions have been in consequence incorporated, but in the main Mr. Martin's
authorities are the same as those here quoted. The reader may, however, do well
to compare Mr. Martin's summary with that of the text.
PAGE 299. Puddling ponds. I have found an early reference to this practice in J.
Houghton's Collection for the Improvement of Industry and Trade ', 1693, iii
(No. 68), where he shows how to make ponds hold water. If Clay be well ram'd [he
is speaking of light soils] and upon that a good pitching of Stones be laid,
Water may be preserv'd to great purpose.'
Walter Johnson - Folk Memory - Publisher Clarendon
Press Oxford, 1908