PALISADEE TH D SETTLEMENT HAREHOPET SA , PEEBLESSHIRE. EXCAVATIONS, 1960

. FEACHEMW . 'byR , M.A., M.SC., F.S.A., F.S.A.SCOT.

. I INTRODUCTION THESE structures are situated at a distance of 700 yds. NE. of Harehope farm-house heigha t a gentl e f 110d o t th . O.D an . 0slopft n eE o . Crailzif eo e Hille para th , f o t ridge forming the watershed between the Flemington and Meldon Burns, 4 miles NW. of Peebles. An,earthwork of generally similar appearance lies 1,100 yds. to the SSW., and another the same distance to the NE. (fig. i). The open, well- drained land in the vicinity of these monuments is alt under grass pasture, in con- tras mano t t y comparable hillsides within view upon which heather predominates. settlemente Th s were chose excavatior nfo n durin preparatioe gth Inventore th f no y of Peeblesshir Royae th y leb Commissio Anciene th n no t Monument f Scotlano s d because they are typical examples of a class of earthwork of unknown affinities that has been recorded throughout southern and eastern Scotland. Typically, a structure of this class is a sub-oval enclosure - often with a suggestion of the subrectangular - forme shalloa y db w ditc eithen ho rearthe w sid lo f whic o ea ns i h banke Th . banks meet roune entranc e ditce headth th th dt a f h, occasionallyo s or e o tw , entrances. In greater Tweed-dale alone 23 such structures survive (fig. a).1 While three times as many earthworks of unknown date and purpose are in addition recorded in the same area, none conform closely enough with the above specification to warrant their inclusion here. This is not to say, of course, that some of them may t belonsame no th eo gt periodevidenco n t bu , e wil available b l e until excavations have been carried out. The Commissioners decided that the excavation of the Harehope earthwork might help to throw some light on this widespread type of structure, and the work was don Juln Augusd ei yan t indebte1960m a Commissionere I th . o t d o t d an s members of the staff for support and assistance. All the recording and drawing wer Feachem. eW undertake. R s , Mr whos y nb e wor kacknowledgI e with particular gratitude. Permissio excavato nt havo t d e edailan y readils accessite th ewa o t sy given by the proprietor, Mr A. Boyd, and a grant towards the expenses involved, and tools with which to work, were generously provided by the Society of Anti- quaries of Scotland. e excavationTh s wil e describeb l e followinth n i d g e variousectionth d san , features wil discussee b l interpreted dan Section di n III. 1 These earthworks average th , e internal dimension 16followsy s a b 0 . e ft.ft - ar ,0 In :whicf s20 o e har ventory of Roxburghshire, number , 15953 , , 1606s52 , 165, 323, 647, 865(4), 865(6), 865(7), 874, 943, 948, 950, 1000; Inventory of Selkirkshire, 140, 143, 147; in Peeblesshire, Quarter Wood Side, Harehope, Lyne Burn and Nether Stewarton whicr fo ,Inventorye hse forthcoming Berwickshire,n i ; Over Hartside. THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE 175 LOCATION OF HAREHOPE

HAREHOPE .OTHEOTHER IRON AGE 0 SETTLEMENTS & FORTS —-.' IOOO FT. CONTOU R

FIG. i. The National Grid Reference of the site is NT 203448

EXCAVATIOE TH . II N The appearance of the settlements can be seen from the plan (fig. 3). It is clear thasuccessivo tw t e structural phase representede sar , each forme ditca y db h with a bank on either side but one considerably smaller than the other. While their entrances almost coincide defencee th , s diverg eithen eo r sid thif eo s point until they become completely separated alss i ot I .clea r tha smallee th t r structur latere th s ei , as its outer bank overrides the inner one of the larger. When the two structures are separated, they appear as in fig. 4. 176 PROCEEDING E SOCIETYTH F O S ,

THE i" CHEVIOT "' A •iS:': 2676 . i

/:.---. • '> V'Vx .-••/': X , : ,0 •-:.:.:---:/..--- RWF

IO 2O CONTOUR. MILES 1200 FEET 366 METRES 15 2O 25 3O 35 4O KILOMETRES

FIG. 2. Earthworks of Harehope Type in Greater Tweed-dale

A section taken on the surface from E. to W. from, the entrance right across the t oveinterio defencee ou opposite th r d th an rn o s e side showe . oveft 5 r 2 drisa f eo the whole site. This line was used as a base from which two systems of lo-ft. squares with 2-ft. baulks were laid out. One of these enveloped the S. side of the entrance othee places th wa rd dan well withi interioe hopne th th en ri tha t trace internaf so l structures, none of which were visible on the ground, might be discovered. This latter system, called describe s i Sit , eA d first.

(i) Site A (fig. 5) When the turf was removed the roots of the grass were found to extend to the natural crushed rock surface some 6 in. below. Three contiguous sherds of medieval pottery (fig. 8, i) came from the grass-roots. When the turf had been removed, various settings of stones could be seen as well as areas of concentrations of stones and boulders laid flat which could be described as coarse paving. It soon became clear tha settinge th t stonef so s were chock-stone palisada n si e trench thad e an ,tth coarse paving represente attempn da mako t t enaturae parth f o t l rocky 'floor' less uneven. In the event, three separate successive structures, all in contact with each THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE 177

HAREHOPE ^..-.•..,,,.,11,,i, ^^^MMi.M.,.,rf,,^/|i}{v

•* :*'''^'x'i%

PLAN KMF RWF 3OO FEET 9O METRES

FIG. 3

other, were disentangled fro palisadee mth post-holespavinge 2 2 th , t d i an ,d an , wil conveniene b l describo t t e ordee theth discovery f mn o ri . House 3 (fig. 6). A circular timber-framed house 36 ft. in diameter, the founda- tion whicf so h were formepost-holex si y db s (number continuoua s 1d 15o 0t an ) s palisade thre e latese trenchth th f es o t structureswa , five eTh . post-holes numbered 10 to 14 were all much the same size, varying between 6 and 8 in. in depth below the uneven natural surface and about 9 in. in diameter outside the chock-stones. Wher lattee eth r were stil placn i l e their disposition suggested tha poste th t s they originally held may have been about 6 in. in diameter. The only fragments of charcoal found on the site came from post-hole 13. The sixth post-hole, number 15, outee th rn i trency la f Housho , describee2 d dimensions belowit d an , s wert eno accurately estimated.

THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE 179

a~5 "? HAREHOPE nlV*s.. ., *,''P" 4tkh ?< SITA E 4fc.'Al-a 3 * i|r""" * ^ ^"V* i\~'J^^'*A?~+*\ mW,' • " »^_ -'

«J^va? i. «gi.

^n^J-I-.-T^C; - •.•.':•-•'. :: :TT'>^ •'••• '•' '\'

FEET _5_____6 KMF METRES RWF

FIG. 5 The enclosing palisade trench measured about i ft. in width and about 8 in. in depth. A good many chock-stones were still in place, and, particularly in the NW. arc, it was noted that flat stones had been rammed down nrmly along the inner thef i hel d t trencspacee s yha a sidreguladth no a tha d t resf e e eo hdan ou th tar t r row of posts. This indicated that the wooden outer wall of the house had been composed of a screen of contiguous split timbers, rather than of a series of posts connecte wels wa l , denned wattlesy d arc b whol e . entrancee W Th .Th f .e o th n i , . arc, howeverS sarc d fadin. an founmissings e E . b e wa N ,o g dt th e graduallth , y l8o PROCEEDINGS OF THE SOCIETY, 1959-60 away onto an eroded slope. It is difficult to suppose that this was due to natural erosion, and the explanation may lie in the reported trench 'dug across the site' by . ACuriO . e abou year5 2 t s ago.1 The only other features attributable to this house are the little trench T which form traversa s . withift entrance9 e post-hole nth th d an ,e numbereds i 16t I . possible that another post-hole, shown on fig. 6 as a dotted ring, may have existed t thi f 16o Bu s .. positioN . confuses ft e fillin 3 nwa th pavind y an gb d g blocking post-hole ii (see House i below)existenc s t proveit no d s an ,d e wa beyon d doubt. HAREHOPE

HOUSE ONE HOUSE. TWO FEEO 2 T DIAMETE 4 FEE2 T DIAMETER R HOUSE THREE FEE6 3 O 3 T DIAMETE4OFEE O 2 R T O I O 10 METRES

FIG. 6 Nevertheles probabls i t si e that trenc , post-holhT postulatee th ed 16an d post-hole supported and formed parts of a structural feature designed as an interior entrance hall - there was no external porch. The rough paving already described was con- centrate mose th r aislte dparfo th e n betweei t pose woodee nth tth rind gnan outer wall. No hearth was definitely recognisable. Four odd post-holes, numbers 17 to 20, were found within the house. A remarkably worn part of the conical upper stona fla f teo bee-hive rotary quer pavine founs th npoina wa n (figt o gda ) 9 t. 6 ft. S. of the S. end of trench T. This house is discussed in Section III below. trence th f e HouseSSEo h c Th carryin (figar 2 .. 6) . outee gth r wal Housf o l , e3 pavinsoms d it an f eo g stones, were foun overlio t d parn lengtheo i tw t shallof so w curving trenches or depressions cut into the rocky surface. The trenches varied in depthe froOn widtm n i abou. o t abou in d . h6 in an t fro . . 3 1tft io 0 in t m . ft i lengtn i . ft h 0 . (lik 3 whicE palisad e forme c e th ar th h n n fadeda i e t trencdou f ho House 3) after running half way round the circumference of a circle 20 ft. in dia- meter. The other trench formed an arc about 15 ft. in length opposite the former coursa n o ed abou an insid. ft 3 t. eThiit s trencpostA . -hE e alsth o n fadei t dou 1 Information from Mr Angus Graham, F.S.A. THE PALISADED SETTLEMENT HAREHOPET A S , PEEBLESSHIRE l8l hol diameten i dept en . i ft i(7). d in han ,8 r outsid chock-stonese eth situates wa , d similaf longee o th ) 9 f o rrotherd o proportiond trenchan tw en 8 ( . sd N an , e th s t a . sectorinteriore SE th f e otheo se th A th . n i rd an . NW wer e th e n foundi e on , settin f halgo dozea f n large fla r fro. sector centre N fa t mth t stonee f th no ,eo n si the house, was probably a hearth. These remains were recognise thoss dsmala a f eo l ring-groovhouse th f eo e type first describe e Hayhopth t da e Knowe settlement, comparable on e th sizn ei o t e marke plan.e dhouse th 'Xn Th 1o 'e is discusse Section di I belownII . House i (fig e remain. 6)Th . f thio s s structure consisted entirel largx si f eyo post-holes, numbers I to VI. In all cases but IV they were entirely obscured by the rough paving in the aisle of House 3, while V was in addition, in the first place, overlai shortee th y nb r trenc Housf ho . Whee2 n excavated, these holes were found contaio t n both several large, wedge-shaped chock-stone r moro e e on larg d ean s boulders. Emptied, they measure numerican di l order 16, 162 2 , 15d ,an 182 2 , averagn inchea d depthn si ha e d diamete. Thein an , 0 2 y f foro r m. circla ft 4 e2 in diameter. They had been so rammed full of stones and boulders that it was concluded that this must have been time donth et ea whe postse nth r theio ,- re r mains, were pulled out or had completely rotted away, leaving large, gaping pits. This must have been done whe npavine th Houss built d f Houswa g o an 2 e; 3 e eventually overlay both the original fillings and the trench of House 2 where this overlay the filled post-hole V. Three of the unattached posts - 17, 18, 19 - fall within the ring formed by post- hole VIo t therevidenco t n sbu I ; s ei shoo et w whether they belonge Houso dt eI or to House 3, or to neither. This house is discussed in Section III below. Other Post-holes. Apart from the four unattached post-holes 17, 18, 19 and 20, othero tw s came withi excavatee nth d are f Sitclos y a. o These la eA , e 22 ,d 21an together near the brink of the eroded slope. The former was 10 in. deep and as many in diameter, and its filling contained the only piece of prehistoric pottery to diametern i . foune in b 9 . dee dd in . 7 pe (figlattean , 2)s 8 Th .. rwa (ii) The South Side of the Entrance (fig. 7). x squaresSi ,d witeacan h. y 1h b ft 0baulk . 1. wid0ft ft 2 se between them, were lait alon dou base-line gth visible coveo th et l e al r. sid remainS f eo e th f so entrancee th werd an , e extended where necessaryconsecutivo tw e Th . e structural phases which can be seen on the plan (fig. 3) were revealed in this excavation, and t wili conveniene b l describo t earliee eth r first. e largePhaseTh . rI enclosure (fig ) whic4 . h constitute firse dth t cultural phase was mutilated near the entrance not only by the construction of Phase II but by later ploughing. It can be assumed that originally the banks were of reasonably entrance uniforth f o md . eS an distanc siza e l roundr eal th fo o t t f 15 eo . ;bu 0 ft 300 ft. to the N. of it the outer bank of Phase I has been almost obliterated, only the slightest intermittent traces still remaining. In addition, the sections of the banks P.S.A.S., LXXXIH (1948-9), 47, fig. 2. 1 HAREHOPE ENTRANCE

--:--'•, .v'

' ITO

KMF RWF

Fio. 7 E PALISADETH D SETTLEMENT T HAREHOPEA S . PEEBLESSHIRE 183 which flank the entrance were removed and shovelled into the heads of the shallow ditch by the builders of Phase II, to clear the approach to their gateway. e excavatioTh n revealed tha continuoua t s palisade trenc alonn hra crese gth t innee oth f r rampar Phase th f eI o defencest . Thie s . sidth appeare f W eo e th t da . r befor12 ft fo curvo place t . th . e t N eNE edyina whern t cuttinra gou e d eth gan rampar seatewhics n i t bee d wa awag t dhha i n du y int ditche vere oth Th y. lowest few inche thif so s bank coul tracee db littlda e furthe t begai s d a rroun o nt en e dth of the ditch, but beyond that point no vestige of it remained within the limits of the excavation. No sign of the gateway was found, but an estimate of its probable position is shown in fig. 4. n contrasI Phase e remain . th earliee II th o t tf o sr structure , e latethosth f reo one were found to be comparatively well preserved and quite undisturbed except decayy b . Here, once again, the settlement was found to have been enclosed by palisades moundw lo se n eithen i t o s shalloe r th sid f eo w ditch n thiI . s case, howevere th , arrangement of the palisades and other features at the gateway could be fully under- stood. The palisade on the outer bank turned out a little as it approached the entrance from the W., and then curved back through N. and NW. to terminate . anothe in post-hole a 6 n . i ft r 2 post-hol f o . p Aftega ea r forme terminatioe dth f no innee th r palisade, whic curvthencn o t h ra . ed eNE roun an . d NW throug d an . hN so run off W. along the inner mound. The narrow gap between the two termini gave access from the entrance passage to the semicircular space formed by the curve outee oth f r palisade. Within this space were four post-holes cornere th t a f st o se , a rectangle measuring 7 ft. from E. to W. by 4 ft. transversely. The two southern post-holes were joine palisada y db e trench. The gate itself, represented by two large post-holes 9 ft. apart, one on either side of the roadway, was joined to the palisades on the outer mounds by a fence frop whicgate-poste u mth n semicirculae hra th o t s r curves e roadwaTh . s ywa metalled with fine cobbles. Whe removes ture nth fwa d fro bane mth k entrance flankin . sidth S f eo e gth e passage t ,onc a ther s e ewa reveale a gread t quantit f stoneyo f moderato s e size. Most of these were dislodged, and many had fallen down onto the road; but at one innecurvplacee e th th t rn ea ,o mout passagee th f ho numbea , stonef ro s were found positionn i , formin cappinga earthee th o gt n bank. From thi followt si s that origin- ally the rest of the stones formed such a capping to the rest of the bank, at least as gatee fence th th t .fas ea a r (iii) Finds) 9 d (figsan 8 . Pottery. shere founs On d wa post-hol n di e 21 (fig , 2)t represent8 .I . se parth f to wallarga f o le vessel, possibl flower-pof yo t shape slighe Th . t groove f alono e gon longee th r broken edge structurals si paste Th reasonabls .e i y grite smootth sd han of moderate size, while the surface of the outer side is encrusted with sooty deposits. At present, littl esaie b mor dn aboueca t this sherd than thacharacteristis i t i t f co the local native pottery in use in the pre-Roman Iron Age. 184 PROCEEDING SE SOCIETY OTH F , 1959-60 Three contiguous sherd latf so e medieval pottery were grass-roote founth n di s in Site A (fig. 8, i). Stone.e verTh y wor ne uppe parth f ro t stona fla f teo beehive rotary quern (fig ) mus9 . t originally have appeare restores da illustratione th then s di wa n t I . very closely comparable to one found at the Iron Age Settlement at Huckhoe, Northumberland, exampler 1fo . A chipped stone dis ca quarte incn sino a t diamete n i hf .p o ru thic d kan r (fig. 8, 3) lay on the surface of the bank, under the cobble layer, just outside the N. semicirculae th f o c ar r terminatio outee th f rno palisade. e naturaTh l rocky 'floorcomposes f Sitwa o ' eA d eithe exposef o r d living rock

O INCHES CENTIMETRES RWF FIG. 8 or of chips and coarse mud of the same shaly substance. A few broken pot-boilers were found whe aree cleareds nth awa pebblealsd o oan ,tw flattenef so d spherical shape, each measuring about 2 in. along the longer axis and i J in. along the shorter. apparentls wa e On y e othenaturath rd fabricatedan l . Whil s unlikeli t ei y that such objects would occur naturall t thaya t place, evidenco thern s i e shoo t e w whether they were selecte seriouy an kepd r dan sfo t purpose wite us h ,r sucfo s ha the sling.2 III. INTERPRETATION AND DISCUSSION Since the autumn of 1946, field-work and excavations carried out both in Scotland and in the North of England have led, amongst other results, to a great increase in numbee recordeth e th f ro d remain earlf so y timber structures. While preparine gth

Arch. Ad., 4 ser., xxxvu (1959), 270 and fig. 14, 2. 1 cf. P.S.A.S., LXXXIV (1949-50), 12figd 2an . 6,6. 2 -A- ifc^iil

INCHES o i IO 15 20 3O CENTIMETRES RWF

FIG. 9 l86 PROCEEDING SOCIETYE TH F O S , 1959-60 programme of field-work to be done by the Royal Commission on the Ancient and Historical Monument f Scotlano s d durin e 194gth 7 season r KennetD , h Stee- ob r serve n air-photographo d e Nationath f o s l Survey indications whic e thoughh t migh e surfact th wel e b le trace f palisado s e trenches formin ge earlies parth f o t t defence settlemene th f so Hayhopn o t e Knowe, Roxburghshire. This interpretation was prove correce b smala o dt y b t l trial excavation durin summee gth f 1947o r ; d onc e possibilitan th e e surfacth f yo e recognitio f palisadno e trenche d beesha n discovered and established, the Commission's officers noted nine other structures of the same sort within the county.1 Meanwhile two other palisaded enclosures were recorded in Roxburghshire, both being found by chance during the excavation of later structures that had obliterated them - a method which until now was the onlavailable.e yon 2 Since that time many more such structures have been found, particularly in County Durham, Northumberland, Peeblesshire and Dumfriesshire. The enclosing palisades take various forms, and the round timber-framed houses tha foune ar t d within man thef yo m also sho wgreaa t variet constructionn yi . Prior to excavation, there was no indication that the Harehope earthworks, too, were palisaded structures. They were representative of a type not previously ex- amined, and appeared on the surface as a simple two-phase structure, each phase consistin enclosurn a f go e forme shalloa y db w ditchw flankelo a eithey n b do p rli mound. The lack of any visible ruins in the interior suggested that either any structure it might ever have contained had been of timber, and so had completely disappeared r non o ,eved eha r existed. Excavations showed tha enclosure th t e did, in fact, contain the foundations of circular timber-framed houses; and that the low mounds held palisades or fences. Earthworks of which the surface remains appear closele b o t y comparabl Harehope thoso et th f eo e examples occur widely demons a , - principle strateth slighe t th Bu figf n di eto . ban.2 k which, though never itsela f serious obstacle, become whee son t ncarriei palisadea s havy mucma ea , h wider distribution s ofteha nt I bee. n observed r instancefo , , tha e 'defencesth t f somo ' e structure sizpositioe d sth ean whicf no h puts them int hill-fore oth t class r thao , f o t large settlements equalle ar , y inconsiderable mann i d y;an such case reasonabls i t si y certain that neither the removal of stones nor any other form of mutilation can be blame r theidfo r weak condition e sameTh f course. o , , applie mane th o yst other as yet undated earthworks the surface appearance of which differs in some degree from those of the Harehope class. possibles i t I , therefore, that thi stime techniquon e t widela s ewa y usedt A . Harehope, a very few of what are now known to be the chock-stones of the main palisades protrude throug e ture moundsb w hth fy lo nea e creste ma th t th r i ; f o s that occasionsn o , , such stones have been wrongly attribute revetment.a o dt During excavation hill-fore th t sBonchestea n o t r Hill, Roxburghshire, sectioa n

throug outee hth r defences (Cutting VII) included wha interpretes twa rebuila s d3 a t rampart, the second phase of which was faced both within and without with stone 1 P.S.A.S., LXXXII t theI no (1948-9) no listetw e d ,werTh 6 . 4f e Shoulder Hill (Inventory . 670No d )an Steer Rig (No. 1044). 2 Hownam Rings, P.S.A.S., LXXXII (1947-8) Cappuckd f.0 an , 20 , , Inventory . 803No . 3 P.S.A.S., LXXXIV (1949-50), 113 f. THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE 187 block topped san d witrowo stakeshf tw so thin I . s case, however widely-placee th , d stakes toppin gstone-facea d rampar widtn ti . 14ft h bear relationshio littln r eo o pt principle th e practise t Harehopeda t corresponbu , d more closel well-knowa o yt n arrangement found exampler fo , t Hollingbury.a , 1 Defences of settlements or small villages consisting principally of close-set twin palisades are not now so rare in the record as they were when the defences of Hayhope Knowe were first recognised. For instance, an example almost identical with Hay- hope Know s beeeha n discovere Morton do n Mains Hil Nithsdalen i l , Dumfries- shire, while a smaller example crowns Potholm Hill in Ewesdale, in the same county. t Huckhoe2 A , already cited same th , e phenomenon occurs, whil alss i t oei 3 known in small homesteads such as Gray Coat, in Teviotdale, Roxburghshire,4 and t Brandoa n Moor, County Durham. nonn I 5 f theseo e examples, howevere th s i , palisad emoun plantew lo eithe n a do n di r shallosida f eo palisadee wth ditcl al h- s are, on the contrary, set in trenches dug into the rocky subsoil. With this reservation, however, it may not be unreasonable to infer a comparatively close relationship between the two types. The more elaborate gateway at Harehope may suggest a later date tha simplee nth r style adopte t Hayhopda e Knowe. Amon e earthworkth g s generally simila n appearanci r o Harehopt e e on s i e which, if correctly identified, may throw further light on the broad dating of the group t TamshieA . l Rig, Roxburghshire, earthworn a 6 ramparto tw f ko s witha median ditch, enclosing a suboval area measuring about 220 ft. by 190 ft. inter- nally overlais i , smala y db l walled settlemen rounf to d stone house typa f so e dated to the second century A.D. Such a succession is described as a common phenomenon and it may be inferred that the earlier structure belongs to the pre-Roman Iron Age. The design and, presumably, the execution of the defences at Harehope show a certain degree of technical ability and imply a substantial measure of corporate effort indees a , founs di d elsewher t comparablea e places t HayhopA . e Knower fo , example, no fewer than 1,600 posts, each about 8 in. in diameter and six or seven fee lengthn i t , must have been require e twith nr fencdfo e alone d timbean , n i r quantitie thif so s lease ordeth tt a rwoul d have been necessar eacr yfo h phast ea Harehope. Here, in addition, the hooked ends of the outer palisade contain settings of four post-holes which probably forme framewore dth littlf ko e tower space th es- betwee settingso eithen o tw e greae o roade r non th ,to th sid hav o t s f ti , eo e per- mitte erectioe dth gantra f no y spannin roade whole e scalgth th f Th . eo structure is still small, even thoug e worth h s elaborateki , whil e finth ee cobbline th f go entrance road suggest degrea s f orderlineseo t alwayno s s recognised r creditedo , , in such native structures. A tentative reconstruction of part of the second phase work t Harehopsa shows ei fign ni palisade . 10Th . e trenches wer intg e eodu th made ground of the low bank, and this may have something to do with the fact that no distinct post-holes were distinguishable in their filling. On the other hand, they may have held a split-timber fence. 1 Ant. , xiJ. i (1932), if. unpublishedt 3 Playe s na .

Plan as yet unpublished. Inventory, 994.

4 53 Arch. Ael., 4 ser. xxxvii (1959), 224, and forthcoming excavation report. 8 Inventory, 943. : ; !l|H, b\i== I'illftm: : i ii. ,tf '?£. ^^^--^^^- ^^^Si£ P^£^-^^^r-^5fc^ ^ = 5£r: HAREHOP E- CONJECTURA L RECONSTRUCTION

FIG. 10 THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE l8g It may be remarked that while Hayhope Knowe and many other palisaded settlements were built in the commanding position proper to a hill-fort, the great majorit Harehope-type th f yo e earthwork indefensibln o e sli e gentlee slopeth d san small minorit inconsiderabln yo e ridgesformee almoss th i f t i I .rs a t clas havy sma e bee worke nth f pioneeo s r settlers whil lattee eth r were thos f communitieeo - ex s panding in settled country. The three houses found in Site A include two of recognisable types, but the earlies less i se eason t parallelo yt t WesA . t Plean, Stirlingshire, r SteeD r recorded a house which consisted of a ring of twelv1 e post-holes but with a central post, whereas House t Harehopa i centrao n s eha l post. However thin i , s casgreae eth t sizf eo the post-holes suggests that the roof could be supported on them alone, without the need of a central one. It may be concluded that House i could be of a Late Bronze Age/Early Irotype neAg placee whicb n dh ca secon e earlth n yi d e halth f fo first millennium B.C. thad t an ,coul i t d have bee wholinitiale e parna th th r f eo o t , possibly unenclosed, occupation of the site. Hous relates i typ e e2 th eo dt firs t recognise described dan t Hayhopda e Knowe, sincd an e recognised from surface trace t mana s y other place Greaten si r Tweed- dale. The trenches are intermittent and the ground between the various sections is undisturbed. Likewise conford post-holee an recognisabl o th ,n w mo fe t e sar r eo recurrent plan. That such houses should be so comparatively common, however, implies that they were considered to be serviceable, and although reconstruction is difficult, the picture achieved in fig. 11 of the definitive paper cited above serves as well as any. A house which appears to be a somewhat advanced version of this type was Toftse founth n d,i Stanwick Norte th n hi , Ridin f Yorkshire.go 2 Here, aftee th r absence of a unitary scheme among the post-holes had been noted, it was observed that 'light timber anchore mixee th n ddi surfac boulder-claf eo unlikels yi leavo yt e a complete imprint'. It is probable, then, that House 2 belongs to a second occupation of the site - Harehope phasth f o eI e defences type Th e. occur severan si l hill-fort settled an s - ment vicinitypossible s i th t i n si d e,an tha closta e paralle thio lt s phas occupatiof eo n . slop settlemene f WhitNE exisy th o e e n ma i th te n Knoweo t , Roxburghshire,3 wher t leasea t 11 house thif so s typ enclosee ear vern di y slight mounds flankinga median ditch. Hous conforme3 s closel t exactlno typa f y i o yt e first recorde homesteae th t da d t Wesa t Plean. Whil earliee eth r house ther consideres Latf ewa o e eb Bronzo dt e Age style, its successor was found to be of the pre-Roman Iron Age. It was, of course, impossibl gues o et lon w Late gsho th e Bronz style f house-buildineAg eo g carried on through the first millennium B.C., and at what point the later house was built, but it was suggested that no gap may have intervened. The second house was pre- sumed to have lasted possibly until the arrival of the Romans in the middle of the 8o's of the first century A.D., but no longer.

P.S.A.S.. LXXXIX (1955-6) f7 . 22 , 1 Wheeler Stanwicke Th , Fortifications (1954) PId , fig8 ,.an VII.3 . G I Inventory, 159. 2 3 PROCEEDING E SOCIETYTH F O S , 1959-60 Whil Wese wale th eth f t o l Plean later house consiste t intervalpostf a do t sse s of 3 ft. in a palisade trench, that of House 3 at Harehope was formed by continuous rows of flat timbers. Both houses measure 36 ft. in diameter. The only poin plano t whica ttw se diffehth r substantiall entrancee th t a t Wess yi A . t Plean the trench of the outer wall of the house turns out on either side of the entrance to form a porch, while at House 3 there was some kind of an internal entrance-hall, littla y denne eb innee wooded th t ren da n partition (Trench T)thisn I . , House3 recall housee sth t Scotstarvitsa , Fife, which, although large mord ran e elaboratn ei 1 design, had internal entrance-halls. House 3, then, represents a variety combining featuree somth f e o Wesf so t Plea , somnII Scotstarvif eo som d s onew it an t f ene so own t probablI . y datestheso d s ea , others, fro undeterminen ma dlatee timth rn ei pre-Romae yearth f so n e Iroth n y represen y b Age .severaf p ma o u t e t .I on pu tl builders of the Phase II defences. The quern-stone probably belongs to this final perio occupationf do . evern I y casproblee eth datinf mo timbee gth r structures recorde pastoran di l North Britain is very great, owing to the invariable scarcity of datable small finds. excavatione Th t Braidwood,a s 2 Hayhope Knowe, Scotstarvit, West Plean, Glen- achan Rig and Harehope have between them produced hardly any objects of 3 undisputed date nond an thes,f eo positionn ei s where their evidence woul fulf o le db value. Lacking such necessary aids, the several authors have been forced to use judicious mixtures of negative evidence and inference to produce tentative assess- ment probabilityf so placen I . s where palisaded structures have been found beneath later walls or earthworks, such as Hownam Rings or Huckhoe, the sequence, at least, is certain t lac f datablbu ;ko e material, e extremaddeth o dt e shallownese th f o s soil betwee grass-roote nth undisturbee th d an s d natural roc r earthko oncs ha , e again prevented close dating of the earlier structures. It is, however, most probably safe to assume that, in perhaps most cases if not all, the succession of walls or earth- work palisadeo st havy sma e been carrie t withoudou breaa t occupationn ki ; this must surely have been the case at such places as Hayhope Knowe and Huckhoe. The same is probably demonstrated by the palisaded hill-fort on the summit of White Hill western i , n Peeblesshire,4 wher enclosurn ea e measurin lengtn i . ft h0 g40 within two widely-spaced palisades is in .part surrounded by an incomplete system of three rampart innermose sth whicf o t obliterates hha d short stretche outee th f rso palisade.5 It must be observed, however, that as yet comparatively little excavation has been done at such places. Large numbers efforts and settlements defended by walls, ramparts or palisades exist in which there are the surface remains of timber-framed

1 P.S.A.S., LXXXII (1947-8), 24 . 1f 2 P.S.A.S., LXXIII (1948-9), . if 3 P.S.A.S., xcn (1958—9), 15 f. 4 Inventory, forthcoming. A considerable number of incomplete forts and other enclosures formed by walls or ramparts and ditcherecordedw no 5 s sincompleti n a . e b Wha y ma te palisaded enclosure, however, summioccure th n so t of Yeavering Bell, in Northumberland. The summit is surrounded by an irregular ditch too large to be the surface indicatio palisada f no e trenc t equallhbu y difficul explaio t t othey termn ni an f r o sconvention . Steer suggestes D rha conversation di n tha t mighti t possibly represen palisada t e trench thaabandones twa d fillebefors wa d t ewiti h posts hill-fore Th . whicn i t liet hi formes si stona y db e wall containd an , s 'platform' houses. THE PALISADED SETTLEMENTS AT HAREHOPE, PEEBLESSHIRE igi housetype anotherr on e o f so t wilI .necessare lb carro yt t excavationyou t somsa e of these to determine whether walls or ramparts were original, or whether they succeeded palisades. But, even so, the likelihood of discovering useful small finds sucn i h place remotes si t wilI .necessare b l carro yt t large-scalyou e excavations at some such site as Traprain Law before the whole picture can begin to appear very clearly. To sum up, it is suggested that the earliest occupation of the Harehope site may have Late beeth en i Bronzr Earl o e y e Ag Irotransitiona e nAg l period, some time after the middle of the first millennium B.C. : that Phase I of the palisades, and Hous , followee2 d afte unknown a r n lengt f timeho , probablf yo towardd en e sth the first millennium B.C.: and that Phase II, the last occupation, which includes House 3, followed immediately after this, possibly even as late as the beginning of e firsth t century A.D.

SocietyThe is indebted H.M.to Treasuryfor meetingillustrationsthe cost the of thisfor paper