|
Introduction
The upland valleys of Caernarvonshire contain many traces of early habitation and the purpose and dating of the rectangular huts in the county has been a problem of especial interest to the Royal Commission on Ancient Monuments in Wales and Monmouthshire in preparing an Inventory of the Ancient Monuments. 1 As part of its research program in the county the Commission excavated in 1953 one of the larger houses, Cefn-y-fan in Eifionydd. 2 It was not clear how a house of this nature compared in date and internal arrangements to the less imposing ' platform ' type of medieval structure to which Gresham has since drawn attention. 3 Therefore, in October 1961 a hut group of this latter type was excavated by members of the R.C.A.M. with three workmen employed for three weeks. 4
General Location of the Site (Figure 1, Plate A.)
The Aber valley is well known for its impressive falls where the Afon Goch plunges over a sill of igneous rock into a broad marshy area where it is joined by two tributaries ; the enlarged stream, Afon Rhaiadr Fawr, rapidly descends a further 11/2 miles to reach the sea at Aber, a village five miles east of Bangor. On the eastern side of the valley 570 yards below the Falls is the hut group chosen for excavation. 5 The group stands on a slight terrace at 600 feet above sea level some 280 yards from the stream ; behind the huts the ground rises steeply to the plateau Meuryn which is itself an outlying spur of Llwydmor Bach which flanks the Falls with a scree slope. The Ordovician shales thus exposed have provided an ample supply of stone for the hut group, but the most obvious visible remains are of a later fold ; the dimensions of this fold can be seen from Figure 2. Its walls still stand between 4 and 5 feet high.
Around and underlying this fold are a group of nine buildings approached from the north by a narrow track (along the 600 foot contour). Five buildings are along the slope and four are across it, three of these being cut deeply into the hillside at the upper end and only slightly raised at the lower end. The general outline of the group was clear before excavation started and there were sufficient surface indications for four of the buildings to be marked on the Ordnance Survey 6 in. sheet. The walls were standing to a height of two or more feet and in many cases only the bracken needed to be removed to reveal the foundation courses and the floors.
Details of the Excavation (See Figure 2 Plans, Figure 3 Sections)
The area excavated is shown on the site plan (Figure 2). The buildings will be described from north to south and their sequence and purpose will be discussed in the following sections.
The approach track, 12 feet wide, was apparent for 100 yards or so north of the site crossing a wet flush as a revetted terrace and on reaching the hut group it broadened into a yard. Furthest north were two buildings F and G. F was initially a shed 8 feet square internally, its lower side roughly paved with stone slabs. Adjacent to it on the south was an oval structure G, 10 feet long and of maximum width 5 feet, its floor similarly paved. Both these buildings lay along the slope west of the track, their upper walls forming its revetment. The two buildings had been altered : F was extended southwards to make a rectangular shed 18 feet long, and the projecting end of G became a feeding bin for a shed H, 17 feet long by 8 feet, with an entrance on the west and a roughly cobbled central walk. In both structures the loosely packed medium-sized stone of the secondary work contrasted to the well-laid lower courses of the earlier buildings. Immediately below the bracken in F was a thin layer of dark turfy soil and in this were five jug fragments of late 18th early 19th century date.
East of the track was a rectangular building E, 28 feet long by 8 feet with no apparent entrance. The lower side was paved with long thin slabs except for a patch of cobbling ; on the upper side the clean yellow gravelly soil lay just below the turf. The walls, standing at the most three courses high, were poorly coursed of boulders without any packing in the gaps. This long shed was secondary to the platform building D which adjoined its end wall on the south.
This platform building D (Plate B ; Section, Figure 3) was 21 feet long by 10 feet, deeply cut into the hillside but only built up 12 inches at its lower end. The entrance on the south wall had as door posts fallen boulders 2 feet square ; the threshold was paved with small stones carefully set in the gravel while just inside the building the paving was of whitish stone, mainly quartzite. West of the door the hut had been considerably robbed to build the fold leaving only the lowest courses, but above the door the character of the walling was quite distinctive. Large boulders of triangular shape formed the base course and the gaps were tightly packed with small stones ; above this smaller boulders were used, also with good packing. The upper end wall stood five courses high, protected from robbing by tumbled rubble, but the side walls had been robbed. The lower end wall showed similarly careful construction with large stones revetting the built-up floor of medium-brown silty soil. The foundation trench was here cut into the gravel subsoil and the stones packed level horizontally.
The floor, of building D was partially paved ; the lower end either had flat slabs or else boulders split length ways to provide flat surfaces, cobbles being packed in the gaps. This was not the original arrangement since near the north west corner a post hole was found beneath the paving. Search for other post holes was unsuccessful. Across the hut dividing the upper third from the lower part was a line of stones set on edge in a shallow trench. This line also distinguished the clean yellow gravel floor of the upper part from the medium-brown soil cover of the paved area. Since the tumble of the upper end wall did not extend westwards beyond this line of stones, an internal division may be indicated rather than a feeding walk.
The upper end wall was protected by a hood which showed as a slight rise under the bracken. On excavation this feature proved to be a 15 inch wide line of small stones against which the yellowy brown sandy hill-wash had accumulated. This hood curved downhill to north of building D, but the placing of a later field wall above this hut and C had obscured its outline to the south. Indeed it was doubtful whether D and C had two separate hoods or shared a large one, or whether C had a hood at all. A trench to find the hood above C only revealed the footings of the field wall, quite different from the hood of D.
Twenty feet south of building D and at a lower level was second platform building C, 17 feet wide. The later sheepfold h destroyed its lower part 6 but it was at least 17 feet long and may have been as much as 37 feet long since a trench outside the south wall showed a continuous line of foundations. The breadth this building seemed to have been a source of weakness and the south east corner had been rebuilt. The original wall was of a single thickness revetting the hillside and was constructed of medium-sized stone roughly coursed ; the rebuilding was only slightly superior work and eight courses stood 5 feet high. Underneath a considerable spill of tumbled stone and soil was a 3 inch layer of grey-brown so suggesting occupation and beneath this an iron-stained gravel floor hard-packed and containing one fragment of 18th century brown glazed pottery trodden well into it. There had been some attempt to pave the floor with small slates but this was only in the rebuilt south east corner. The lower part, as far as could be excavated was paved with small stone (Section, Figure 3).
West of C and in alignment with it along the southern side of the yard was a third platform building B (Plate C). This was 11 fee wide and almost certainly 22 feet long because the substantial fold across the upper end had a deep laid foundation probably using the original hut wall up to ground level. The entrance was midway along the north wall and in front of the doorway was a carefully laid semicircle of cobbles. Within the building the lower part was paved with well chosen stone slabs while the remainder was cobbled. The platform was sunk only slightly below the surface of the yard but the lower end was built up to 24 inches with dark brown soil (Section, Figure 3). The walling stood to a maximum of 3 feet high with four or five courses of small round boulders. The door posts had presumably been robbed by the builders of the fold. Added to the south wall of building B was a small rectangular annexe J, 18 feet long by 8 feet, its axis placed along the slope. The walls were of the poorest quality and at one point incorporated part of a quern, probably found in one of the Iron Age hut circles near the stream. The position of the entrance was uncertain and no trace was found of a laid floor, although a large oak tree had thoroughly disturbed the interior.
Separated from building B by a cobbled passage 8 feet wide was building A, 48 feet long by 11 feet with its axis down the slope (Plate A). There was no attempt to secure a level floor throughout the hut, but instead a series of four steps gave five floor levals of equal area (Section, Figure 3). In two cases these broad steps were held by a line of large cobbles. The floor was light brown gravelly soil with occasional stony patches and areas of streaky grey clay, possibly the caulking of the walls washed out or fallen turf forming the upper part of the walls. The only entrance was in the north wall just above the lowest step and the door posts were of boulders 3 feet high (Plate D). The end walls were carefully constructed with small stones packed between large boulders. The side walls were poor in the extreme although the foundation course was bedded in the gravelly subsoil and stepped gradually up the hillside. As in the other buildings, walls of a single face revetted the hillside and free standing walls were only erected in the lower part of the hut. The lower end wall of A stood on a platform of yellow brown soil only 10 inches high. There was no sign of a drain or of any structural division such as stalls. The side walls were examined unsuccessfully for evidence of cruck roof construction.
The Date and Sequence of the Hut Group
Although the hut group showed signs of rebuilding and alteration, the only evidence for the periods of occupation was provided by pottery. One worn fragment of unglazed red ware lay on the floor at the southern end of building E ; it was probably thrown out from building D or had been lying among the built-up of its platform. This was the only fragment to which a medieval date might be given. The remainder of the pottery, 7 mostly from the Buckley kilns, placed the main period of occupation securely within the 18th century and probably in its latter half; other fragments, notably a posset pot of black vitrified ware, suggested that the hut group had fallen into ruin by the early 19th century. Apart from the quern incorporated in the walling of annexe J, the only finds were a slate whetstone, one iron nail and a perforated copper disc, one inch in diameter.
Less certain evidence of date was provided by the character of the walling in D ; a hut of similar construction on Dean Moor in South Devon was occupied in the period c. 1250 to 1350.8 The stratification on the site at Aber was of little help. The absence of well-defined occupation layers made interpretation of the function of the various structures difficult.
Three periods of occupation can be isolated. Initially the platform building D took advantage of a slight knoll. It was well built of dry stone walling with a turf or rush thatched roof supported by a post or posts at the lower end. Its door faced south away from the prevailing wind and across to a shed or barn represented by building C. This shed C may originally have been narrower but its rebuilding had masked the sequence of events. This house and barn may well represent the full extent of the original hafod, and comparison may be made with Meuryn nearby, Ardda in Dolgarrog parish and Dean Moor in South Devon. 9 A medieval date is quite probable but cannot be proved.
The second period, possibly early m the 18th century, brought greater refinement to the hut group. The rebuilding of C and the erection of B are probably contemporary with the paved floor in D. Further extension of the settlement followed in the mid 18th century ; it now merited the local name ' Pentre Saith Aelwyd ' The Settlement of the Seven Hearths. 10 Sheds were placed along the slope : they were either attached to existing buildings (the poorly built annexes E and J) or were separate (the square shed F and the oval one G). Another shed A stretched down the slope. All these buildings are similar in character and suggest a milking establishment with some agriculture practiced. The clean paved floor of B could belong to a dairy rather than a dwelling house and D be the only occupied hut, though from the greyish soil at the upper end of C some fires must have been lit there, perhaps in braziers. A is best interpreted as a fodder store while the paved floor of G suggests not so much a pigsty as a hay or flax drying room. 11
The final stage was the decay of the hafod early in the 19th century. Occasional squatter settlement replaced a seasonally occupied homestead with a definite function in the economy of the neighbourhood. F and G were rebuilt; the extended F is recognisable as a shepherd's hut and H as a stable. So long as C retained its roof it remained a cattle shelter, but eventually the roof must have been removed and the walls allowed to collapse enabling the fold to be built across both byre and yard.
Discussion of the site
It now remains to relate the hut group to the land use in the Aber valley and also to comment upon the house type excavated.
A large amount of research has already been undertaken in East Caernarvonshire. Dr. Elfyn Hughes in environmental studies 12 and Professor Jones-Pierce in documentary elucidation 13 have concentrated upon Arllechwedd Isaf. However the pattern of land use exposed in detail on the eastern side of Bwlch y Ddeufaen is equally valid for the valleys to the west of this pass. The Aber valley was well populated in the Iron Age as is shown by the number of hut circles and enclosures both below and above the Falls. 14 The more fertile Nant Anafon also has a succession of hut circles extending up the flanks of Maes y Gaer and alongside the trackway which under the Roman occupation linked Conovium with Segontiwn. It is not yet known how far this settlement pattern was modified by the establishment in Aber of a manor of the princes of Gwynedd. The grouping of houses around the llys, the bond vill at Wig and the out settlement or garth in the vicinity of Hafod Celyn can be paralleled in the princes' maerdref at Aberffraw.15 The choice of Aber for the head of the cwmwd of Arllechwedd Uchaf16 was dictated as much by the fertile coastal plain broadening at the mouth of two valleys as by the importance of the river crossing and the low tide passage to Anglesey. The first indication of the use of the upland valleys is the mention of four vaccaries (or cattle farms) at Aber in 1303 and the more detailed Record of Caernarvon names the individual hafotries of Meuryn and Nanteracadrat.17 An identification of these hafodau with actual hut remains was made by the R.C.A.M. but questioned by Gresham. 18 The excavation hoped to establish that the long hut group was medieval and possibly to be identified as Nanteracadrat. It would be stretching the evidence from the excavation too far to say that such an identification could be made. The position of the hut group is, however, one likely to be chosen for a hafod or summer cattle farm. The next mention of this valley is probably in 1547 when two pastures (ffriddoedd) called Meryn and Nanterhiet are illegally occupied.19 In 1716 a Baron Hill rent roll mentions Nant and states how many head of cattle were pastured in the valley.20 The valley Nant appears as one of the townships or parcels distinguished in the parish terriers of 1752 (out of 7 parcels) and 1776 (out of 5 parcels). Between 1799 and 1852 the parish registers record the farmers and labourers living at Nant y Rhaiadr, but this must be the house lower down the valley, marked on the Ordnance Survey drawings of c. 1820 and on John Ingelby's water colour of c. 1840.21
The use of the valley for cattle pasturage rather than for agriculture is evidenced by the visible surface indications. Few areas in the valley were suitable for cultivation, a factor limiting the choice of site for medieval settlement whether permanent or temporary. In many cases in Arllechwedd the round huts of the Iron Age are overlaid by later rectangular huts whose occupants tilled the fields cleared and enclosed by the earlier settlers. In the parish of Aber alone this sequence can be seen on Ffridd Ddu, on the slopes below Maes y Gaer, in the Anafon valley near Hafod Celyn and on the plateau Meuryn Isaf. In each case round huts and long huts can equally well be associated with the fields. In the less fertile Aber valley only two examples of terraced fields and banks are visible. One is near the farmhouse Nant (R.C.A.M. Cacrns. Inv., Vol. I, No. 48) ; the other (No. 39) is a series of five banks up and down the slope adjoining the excavated hut group. Of the round huts one group (No. 40) seems to be most closely associated with the cultivation terraces, although four other isolated hut circles are in the vicinity (Nos. 41-44). The hut and enclosure nearest the Falls (No. 41, Fig. 31) has a later long hut overlying it and another single long hut (No. 45), locally called Ysgubor Goch, is 40 yards away. These isolated long huts set on level ground suggest no more than store sheds but the excavated group closely resembles a hafod in the form described by Pennant or William Williams in the late 18th century.22
In "The Old Summer Pastures " Sayce23 has fully described the life at the hafod. He stressed the simplicity of the conditions in the often temporary buildings of wood or wattle. These huts were barely furnished with beds of hay or heather, utensils of wood, bags made of skin, baskets of coil and ropes of withies. All these items, if discarded, would not be easily preserved in an acidic soil. Furthermore all the household goods were moved down the valley to the hendref for the winter. It is not therefore surprising if the archaeological evidence reveals a settlement with no material possessions. Tending the cattle and goats was the principal occupation but there would be some clearance of land for cultivation. There are indications of this in the Aber valley.
Three distinct vegetation belts are found in the area between the sheep wall below Meuryn and the marshy stream bed. 24 The upper slopes rising steeply behind the excavated hut group (between 600 and 900 foot contour) are heavily clothed in bracken with some blackthorn and rowan. The gentler lower slopes (Plate A) are light woodland, mainly sessile oak with fully grown ash and birch and few bracken patches. The soil is deeper, the ground drier and more heathlike. This belt is limited on the west by the present track to the Falls and the field banks within it show that the ground has previously been cultivated (as it was in the 1939-45 War). The third belt lies below the present track to the Falls and here alder woodland is only suitable for rough pasturage. Immediately west of the hut group a tract of 220 yards north south by 130 yards is clear of trees and this clearance to give a Molinia meadow may represent a cultivation of the marshy stream side by the occupants of the hafod perhaps in the later Middle Ages. To the hay crop produced on the cleared alder swamp must be added the crops raised on the terraced fields won from the oak-birch scrub. Flax, rye, bristle pointed oats (bleu geirch) and six rowed barley (haidd garw) could all have been grown on these upland fields. The pasturage would be grazed from the beginning of May to the end of September and then the crops would be harvested and the herds moved closer to the permanent settlement or hendref.
By the early 19th century the hafodau were falling into disuse. George Kay in 1794 wrote " upon minute enquiry I found no such custom existed." 25 The practice was instead for the milkmaids to journey twice daily into the hills. They would not need to penetrate the valley further than the Falls, since on the marshy Waun was the common pasture of the small holders' cattle. The reason for this contraction in cattle grazing is plain from the Tithe Award map of 1848 which shows all the upland ffriddoedd as unenclosed sheep pasture. The sheep wall built in the decade 1810—1820 left the sheep free to graze to the heights of Foel Fras, Mocl Wnion and Drosgl. The extremely rapid increase in the sheep population of East Caernarvonshire late in the 18th century was the direct cause of the decay of the hafod.
The type of building found by this excavation is quite different from the ' long house ' of South Wales. There man and beast shared the same roof and, despite the elevated position of many Glamorgan platform houses, a permanent settlement was established.26 Here in Caernarvonshire the practice of summer pasturing required less shelter for the cattle. Consequently the pattern which emerges for this district is of a platform dwelling set facing one or two sheds across a yard. This type of plan can be associated with a hafod, as in this excavation and on Dean Moor, or with a permanent upland settlement, as at Ardda or Bodafon Mountain ;27 in both cases some buildings are of platform construction. There is no reason to think that this hut group was at any time a permanent settlement, although Gresham has noted some platform houses in Eifionydd which were originally permanent farmsteads and later degenerated into hafodau 28 Perhaps future excavation in this county may be directed at establishing the thirteenth-fourteenth century origin for those platform houses whose lowland situation indicates a permanent, and not a seasonal, occupation. It is precisely these lowland sites which are more often threatened by ploughing or building and which might give evidence of date being comparatively richer than upland hut groups.
Conclusion
This excavated group in the Aber Valley can be identified as a hafod similar to many other upland settlements. Its origin may he in the 13th or 14th century, but the buildings associated with the earliest occupation of the site can not be closely dated. The main period of occupation is in the 18th century while the decay of the hafod system causes its abandonment in the early 19th century.
L. A. S. butler.
*I should like to thank the owners of the land, The Forestry Commission, and the tenants, The University College of North Wales, Bangor, for giving permission for the excavation ; Mrs. M. Powys and Miss K. M. Cartwright, who gave valuable voluntary assistance in unpleasant weather conditions ; Mr. T. J. Owen of Pen-y-Bryn, who readily provided information on local history and Mr. K. J. Barton, who gave advice on the dating on the pottery. I am indebted to Professor R. Alun Roberts and Dr. Elfyn Hughes, who have helped me upon the ecology of the valley. The tools and other materials for excavation were kindly lent by the Ministry of Works.
1 R.C.A.M. Caerns. Inv., Vol. I, p. xxxiii—xxxvi and Fig. 7 ; Vol. II, p. xxx and Fig. 7 ; Vol. Ill (forthcoming), an introductory section will discuss long huts.
2 Trans. Caerns. Hist Soc., XV (1954), p. 1-7.
3 Archaeologia Cambrensis, CIII (1954), p. 18-53.
4 Although most of the staff helped in various ways, Messrs. W. E. Jr?18 anc* ** Smith were closely associated with the labour of excavation and D. B. Hague took the photographs.
5 National Grid Reference SH 66807057, O.S. 6 inch sheet 7 S.E. Mr. C. A. Gresham was the first to recognise the medieval character of the site.
6. No excavation was undertaken within the folds; the smaller fold was given a temporary roof and this proved very necessary.
7. Fragments of eight different vessels were found in the following positions the earlier group is mid-late 18th century. Hut A : lower 1/5th resting on sandy floor level close to West end wall. Fragment of mug; red fabric ; high olive brown glaze on both surfaces; Buckley. Passage A/B : on cobbling in medium brown soil Red fabric; dull brown glaze on both surfaces. Red fabric with light maroon glaze on both surfaces. Hut B : in lower 1/3rd above paved floor in brackeny soil close to S.W. corner Jug fragment; red fabric; maroon glaze on both surfaces, lightly fluted horizontally; Buckley, 18th cent. Hut C: centre of upper half, pressed deeply into gravel floor— Jug rim; red fabric; dark brown glaze on exterior with thin yellow band beneath rim (broken into 20 fragments). The later group is late l8th early 19th century. Hut C : close to S. wall, dropped among loose rubble when building was already ruined Posset pot, handle broken ; black vitrified ware with dull brown blotches, late 18th century. Hut F: on paved floor near W wall in dark brown turfy soil—Base of jug, red fabric; high black glaze interior; dull black glaze on exterior. Buckley, late 18th-early 19th century. Cream fabric, tortoiseshell glaze on both surfaces.
8. Medieval Archaeology, II, p. 141—157, especially Plate XVIII A.
9. Meuryn : Caerns Inv., I, No. 63 ; Ardda :ibid, Nos. 207—210.
10 This name may be a description of the hut group with its seven roofed buildings; it may indicate that the work of the hafod was shared among seven hearths or households; or it may be a shepherds' nickname suggesting that there were seven corners out of the wind where a fire could be lit. There were no hearths in the buildings but traces of charcoal in three corners, two outside the buildings, showed where camp fires had been placed.
11 Similar structures for drying corn or flax are to be seen at Cwm Fynhadog Uchaf, Dolwyddelan (O.S. 6 inch 23 S.W.) and on the slopes of Penmaenmawr (Caerns. Inv., I, No. 263).
12 Trans. Caerns. Hist. Soc., 11 (1940), p. I—25.
13 Trans. Hon. Soc. Cymmrodorion, 1942, p. 158—188, especially pp.
14 Caerns. Inv, Vol. I notes 12 separate files within the valley Nos. 13,16, 40-44, 48, 50 & 51.
15 Institute of British Geographers, Transactions and Papers, 1953, pp. 51-72, especially p.66-9.
16 J. E. Lloyd, History of Wales, I, p. 236 ; II, p. 686, 765.
17 Bulletin of Board of Celtic Studies, VII, p. 152 ; XVI, p. 116. Record of Caernarvon (Record Commission), p. 138-40 " The ffrithe of Nantmawan, Nanheskele and Ycras with havotries of Meuryn Nanteracadrat and all other appurtenances." (Survey of 1350-1 repeating a grant issued (April 1338).
18 Caerns. Inv., I, No. 39 for Nanteracadrat; No. 46 for Meuryn. Arch. Camb., CVI (1957), p.127 suggests No. 63 for Meuryn and the hut- group which is the subject of this article for Nanteracadrat.
19 Records of Court of Augmentations, p.
20. Trans. Caerns. Hist. Soc., XIV (1953), p. 89.
21 Material in National Library of Wales, Aberystwyth.
22 Pennant, Tours in Wales (1778), II, p. 325-6. William Williams, Survey of Caerns. (1806) MS 82ic, N.L.W. Aberystwyth ; also History of Llandegai, Llanllechid and Aber (c. 1808), MS. 8390 gives full information on land use at the turn of the century.
23 Montgomeryshire Collections, 1956, p.117—145 ; 1957, p. 37-86. A study of the place names hafod, hafoty, lluest and hendre is in Mont. Coll., 1959> p. 13-20 ;1960, p. 177-187.
24. This section on vegetation is partially based on two surveys : ' An account of the vegetation of Southern Anglesey and Northern Caernarvonshire ' by Miss W. H. Wortham (Unpublished M.A. Thesis in Department of Botany, U.C.N.W., Bangor) and ' A Botanical Survey of the Aber Valley 1957-9 ' by Miss Catherine Richards (Unpublished Field Report at Nature Conservancy, Bangor.)
25 General View of the Agriculture of North Wales, p.16.
26 Gelligaer Common: Arch. Camb., 1937, p. 247—268 ; 1939, p. 163—199. Neath Uplands Trans. Cardiff Naturalists Society, LXXXIII, p. 9—18.
27 Trans. Anglesey Ant. Soc., 1955, p. 12-26. Arch. Camb., CIII (1954), p. 43. This excavated hut-group does occupy a site characteristic of many ' platform houses ' in that it is set partly into a steep slope, partly on a rocky knoll, to preserve for cultivation as much good land on the gentler slope as possible. The similarity between the curving field banks below the hut-group in Aber and the funnel shaped pens sometimes encountered in Eifionydd is only superficial
|
