IV.

ANCIENT BORDER HIGHWAYS : THE MINCHMOOR (CATRAIL) ROAD, E WHEETH L CAUSEWAY E ANNANDALTH , E FOREST ROAD, THE WEL E ENTERKINL TH PATH D HARRY . AN B ,G . . YR INGLIS, P.S.A.ScoT., F.R.S.G.S.

Owing to their irregular formation, the Border hills offer only a limited numbe f direcro t passes through whic he taken b road n ,ca s and, to understand their bearing on history, it is necessary to recognise that while there are perhaps a dozen open passes between north and south, there is only one really accessible highway between the east and the west—the Tweed valley. The twenty-mile-broad belt of hill and moor- land which, stai'tin t AbbsS t a g, stretches dow e e centrth nth f o e Lowlands almos Carlisleo t t , form sa barrie r across Scotlan f sucdo ha character tha n wintei t r these hill se Lothian th effectuall f of t s cu y from Berwickshire, and Roxburgh from Lanarkshire, as well as from Northumberlan d Cumberlanddan . It is obvious that with a barrier of hills almost 100 miles in length, and only one practicable route through them between east and west, the natural division of the country in early times must have been an easter westera d nan n kingdom apparentls wa t I Roma.e yth n legions who first created an artificial barrier, when they drew their military 204 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 192411 Y . boundary from sea to sea. When we find that with the withdrawal e legionth o f e s walbattereth swa l de countr th down d yan , soon reshaped itself into the eastern kingdom of Northumbria and the western kingdo f Cumbrimo r Strathclyde)(o a realise w , e thaon t a t period these passes must have been the chief means of internal com- munication, while at another these same roads were the highways used for invading an adjoining country. Undoubtedly the fine Roman road which passed up the centre of Northumbria must have greatly strengthened and consolidated that kingdom, just as the Roman road to Carlisle must have aided Cumberland. But we have no similar relics of a road through Strathclyde to Dumbarton, and it is this possible line of road difficulthao s s ti locateo t t . The break-up of these rival kingdoms of Northumbria and Strath- clyde prior to the Norman conquest, and the subsequent pushing south- ward of the Scots, coupled with the moving northward of the English, caused much strife through many centuries before the political boundary was finally drawn between Tweed and Solway, but it left the Scots with Berwickshire, Roxburgh, and Dumfriesshire cut off from the Scottish capital by a range of hills, and easily open to an English invader. Therefor e buildinth e f Roxburggo h Castle, Jedburgh Castle, and Lauder Fort necessitate keepine dth g open oa roaf d over Soutra o thest e places d withouan , t doub e "Malcolmth t s Rode f ancieno " t charters, joinin e Romath g n roa t Newsteada d a y havd ma ha e, military rather than an ecclesiastical origin. Examinatio e leadinth f o ng Border roads show s tha u sn Der o t e Street (the Roman road) there is no trace left of military defence (other than Roman camps) between Edinburgh and Woden Law, where a long breastwork cuts acros roae guardd th s dan t againsi s invaden a t r from the south. On the Wheel Causeway, excepting a low turf dyke near Wheel ,o defensiv n ther e ar e e relicsa mil t eye ;e belowth n i , Liddell valley a grea, t embankment dow e hillsidth n e pointn a o t s effort to bar this road against an invader from Carlisle. On the Teviot valley, between Langholm and Hawick, there do not appear to be any signs of defensive works on the hills about Mosspaul, nor are there any on the Annandale Forest road on the hills' above Moffat, though there is the curious square enclosure at Little Clyde called " the Roma e Welth nln o Camp.Pat t hBu "abov e Durrisdeer a massiv, e square defensive fort, wit a h traverse nortth n ho eside , guards this pass from an invader coming from the south; while on the Enterkin, what appear a riflt overlooke pi eb o t sd guardan s e pasth s s against someone comin. gup But it is on the Minchmoor road that we see the strongest military ANCIENT BORDE5 R20 HIGHWAYS. defences, for, guarding this road from an eastern invader, a strong breastwork, wit e ditce rock th hth somn i hf ,o form et placeou st cu s a long defenc t onl eacn no ey o s carriei he road t th sid f bu ,do e across the river Tweed to Torwoodlee, clearly barring a large force from attempting to make headway into the west. Five miles to the west, along this same road, another strong breastwork, in some places double, defend t thisi s time agains invaden a t r fro weste mthougth d an ,s i t hi unlikely the two have any connection, it shows that formidable barriers were necessar occasiono tw n y o thi n so s route.

THE MINCHMOOR (CATBAIL) ROAD. e rangOth n f hillo e s between Traquai d Selkirkan r e namth , e "Catrail" has been marked on the Ordnance Survey Maps as a track passing alon ridge gth e fro e sammth east e a westo timt s td i t an ei , shown as turning north and south. It is unfortunate that this great mistak s evehavo t inteconnectioha o tw go rn eo e printth r fo , n with each other. The Catrail runs north and south, while the east and west track is the ancient Minchmoor road (fig. 1), which is cut through the embankment of the true Catrail. This ancient road was the main highway between the east and west of Scotland via Peebles, used by Edwar d mentionean s . Itinerary,dI hi n i d s wella s referrea n i o t d e Statth e Document s appointe1505n i swa , n whe keeo ma dt e a npth road through Minchmoor from robbers for eight days each year at Roxburgh Fair (5th August) s antiquitIt . s alsi y o testifiee th y b d existenc bridga f eo e ove e Ettrickth r a lin n e,i t witLindeana t hi o s , far back as 1234, when the Justice Eyre met there; and it is quite possible that this may be part of the wagon road referred to in the Charter s leadinsa g between Kelso Abbe Lesmahagowd yan . Starting from Lindean, every vestige roas disappeareth ha df o e d e grassoth n ys onli hillsidet i y whed e reacan , w e nhillto th h t a p Linglee thae e fincleaw th td r traca roa f o kd cutting througe th h earthen rampar e Catrailth f o e t .th e Lookin e rightse th e w o ,t g earthwork stretching down the hillside till it enters a marshy pond. To the left we can follow it up the hill—a strong defensive barrier actualle rock formind th f t an deepl, o ycu t a gbreast-higyou h trench. Here is evidence of a strong force having been posted to keep an invader fro e easmth t from using thia simila s a road d r an earthwor; k lies acros e e valleTwee th th scontinued f yo an d s almose Galath e w o ,t t have what appears to be a defensive barrier against every possible way of obtaining a passage to Peebles and the west. On the opposite side of the river Ettrick similar earthworks are visible descending from 206 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924

Coldshiels Hill o thas , t thi e scountrysid th par f o t e appear havo t s e been at one time an important battle-ground in connection with this ancient highway. After passing throug e Catrailhth e roath , d continues uphiln a n i l unmistakable linee hard-beateth , n track being very eviden s ona t e proceeds t three-quarterbu ; a mil , whef so eon n crossing the shoulder of Peat Lawbeatee t onlth no ,s y i n track unmistakable t alternativbu , e trail surfacevisible e sth ar quarten A eo . mila f ro e further, just over

;. "SELKIRK

Minchmooe FigTh . 1 . r Bead.

the highest point, are two stony pits on the hillside, so utterly different from anything met with along the track, that one could readily believe tha t thita sancienlookinn e a poine sitar f th e o t w tgta watch-tower, e exposefoth r d position command a greae roa r a th vies dfo f t wo distanc botn ei h directions quarteA . mila f ero furthee remainth n o r s of a cairn stand close to the road on the right. On descending the hill to Red Nick, the, road disappears altogether in the heather. Continuing blindly on, along the southern slope of Three Brethren, an almost obliterated track comin p frou g m Philiphauge ar e s struckhi w d an , now on what appears to be the later road coming from Selkirk, which would come into use after the disappearance of the bridge at Lindean, whe e Linglenth e roas abandoneddwa three Th e. mile f roao s d from ANCIENT BORDER HIGHWAYS. 207 this point onwards to Wallace's Trench are all much of the same character, and it is quite remarkable that the line of road is seen clearly on the hillside, although one is not walking on what appears to be anything like a road. The grass seems to he of a different colour, yet the differ- enc s onli e y visible adistancea t thers fencese A e gateth .ear n t si i , is comparatively easy to pick out the line of road when it disappears. n reachino s i Bu t i gt Brown Knowe thae remarkablth t e feature known as " Wallace's Trench " (fig. 2) becomes visible ; for here is a fine breastwork 4 feet to 6 feet high right across the road. It is quite

Fig. 2. Wallace's Trench.

evident that at some period of history someone has had to defend this highway agains n invadea t f r o e nortfro e westth d mth r hen fo , the trench has apparently rested on a marsh, while the southern end e earthworoth f a secon s kha d protecting breastwor f considerablo k e dimensions on the steep slope of the hill to the left hand. That it was a defence against the west is shown by the presence of a shallow trenc fee5 h6 t back fro e breastwormth s easterit n ko n side, only deep enough to drain away water. It would therefore appear that at some rainy season there was a regular encampment to bar the road. Whether it had any connection with the Catrail (4^ miles away to the east) is har o sayt dt unquestionabl bu , e latteth y r face n eastera d n invader, westera thie son n foe. 208 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924.

Half a mile further on, the old road from Selkirk to Peebles (about 12 feet wide arid only abandoned about 100 years ago) is reached, and e ancienth t roa s coveredi thin grass-growd i ol sp du n turnpike which, e nexfoth r t 3£ mile o Traquairt s , seem o follot s s coursewit . From this point it appears to have followed the valley of the Tweed to Peebles, probabl e soutth n he riveryo th sid f eo . The only other point that one has to mention in connection with this roa tha, dis t 4£ miles beyond Peeble e Romath s i s n cam t Lynepa , set at the junction of two valleys in such a position that its strategical purpos t verno ys e i clear t unquestionablBu . e Romath t lookyi f i ns a s cam t Newsteaa p e hill-barrierth f do stood en , d e whilguaron et a d Lyne cambeed ha n pt i erectelook f i forc a s sy a db e approaching from the north in an attempt to break the barrier by attacking it from the rear. The distance by road from Newstead to Lyne is 30 miles, but it is onl mile4 ydirec2 a n si t line.

THE WHEEL CAUSEWAY. How the Wheel Causeway (fig. 3) came to get this name is not clear, for tracea t f causewano par o y foune f thib an highwayd o t o n sol t di s yi . Its position, indeed, would have been more or less problematical were discovere th r interestinn a fo t f Scottisye o no th t i f o g h p ol Borderdma s draw n 1590i n , which shows Wheel Causewa a stretc s a y f roao h d somewhere nea e ancienth r t churc villagd an h f Wheeleo ,e closth o et boundary between England and Scotland. A further reference to it appear e Bordeth n i sr Papers in 1585, wher s describei t i e mila s ea d within Scotland. In 1917 e latr Adath , eM m Lain f Hawicgo myseld kan f startet dou folloo t course wth f .thieo s road. Startin t Deadwatega r Station, near e headwaterth e Nortth f ho s Tyne e founw , a dgrass-trac k about 9 feet wide at the back of the adjoining farmhouse. Digging into it, nothing but virgin soil was exposed on the English side of the Border; morno e thawas it na beate n quartea track a milhad In .it of er faded away, and the only indication of continuation was at a right angle in the wall erected by the Duke of Northumberland (to define the real boundary between England and Scotland), where the signs of what looked like an opening in the wall having been built up, suggested that the road must have passe nexe th d ttimer e halfthrougFo on .- t a t hi mile Mr Laing and I traversed the gently sloping hillside at the foot of Peel Fell, in parallel lines, in the hope of striking some signs of a track such as was shown on the Ordnance Survey; but nothing what- ever was visible. But as we neared Peel Burn a definite road became ANCIENT BORDER HIGHWAYS. 209 visible (in a position nowhere near that of the road marked on the Ordnance Survey), and passed between a row of pits from 1 foot to fee8 depthn i tfee w froaboud o t fe t an ,mfee a 4 length n 1 ti t . Coming upon these so suddenly on the hillside, one felt a sense of mystery as o theit r origin, heightene e facth t y thab d t they recalle a sligh n i d t degre e piteth s found beside Dere e AnnandalStreetth n s o wela , s a l e Forest road; but as they had no regularity or uniformity, there was every room for conjecture as to their origin. Subsequent inquiry, how- abandonino t ever d le , theorisingy an g r similafo , r pits exist ovee th r adjoining e farlimeston th f Peelm o d an , e formation suggests thee yar cause e sam Avatery th db t e A time remarkabls . i t i , thin ei s thai t i t one place beside the road that they appear on the hillside.

SCALE 3MILES

w

Fig. 3. The Wheel Causeway. A little furthe n ovee o moor th e rroa th r d fades d wa away d an , not again observable until we saw a perfectly definite line of descent to the little burn at the spot called " Bagraw Ford." Here the track was accentuate s havinit y b dg bee o somt n e exten watercoursa t n ei heav jusd ha ty t aboui rains d tan , enough widt tako ht e ecarta Th . burn at this point is quite small, and one can jump across without difficulty. On the ascent of the slight hill on the opposite side, after a few yards, nothine onl th s e visiblroady d th gwa thin f an , o e g that might indicat s coursfaina it e s t wa edepressio e ridg th fee0 5 en to n above, which one might assume as being its probable course. From this point onwards to Wheel there is little depth of soil on the hillside, and there is hard, firm ground at any point along the crown ridgew lo tha o e s thi,n oth o tf s lan roae dth d again faded away, though one naturally shaped a course to the present sheepfold, built out of the stones taken fro village mth churcd ean Wheelf ho . VOL. LVIII. 14 210 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924.

juss i t thit a tI s point tha perfectla t y clear embankment—the only one on the whole road—becomes visible. It stands only 1| foot to 2 feet abov e surroundinth e s quiti gt i e soilt welbu , l defined, carryine th g road acros shile naturaparo th l n f thato d ha tl drainage. This I ,think , is unquestionably the real "Wheel Causeway" referred to in 1585 and 1590, which antiquaries have given as the name of the whole road, but which from the documents referred to seems to have been little more than a definite place beside Wheel village. The situation of this hamlet demands particular attention. Looking northwards e righth tn o ,are . irregular mounds which cove l thaal r t remain e almosth f o st forgotten villag churchd an e . Situatee th n o d backbone of a low ridge just 1000 feet above sea-level, the ground falls away gentl a tin o yyt rivule n eaco t he east th side ,o t Pee: l Fell stretches upwards, 1000 feet of steep, grassy slope j to the west, Dod Fell, about a mile away, is only 500 feet above; to the north, the grassy slope steadily easn risea yn i s gradien e watersheth o t t t Wheelrigheada d , carryin anciene gth t road surroundine almose leveth th f o lo t t g hills. Southwar e grounth d d slopes easily downwar e Liddelth o t dl Water, a little over a mile away. The land is not fertile, and its high altitude renders it useful onl r grazingyfo . Ther s nothini e g her o supporet t any .population, and the village has likely never been more than a place of call. The one remarkable feature is the wonderful spring of almost ice-cold water close by the church, which no doubt has been the primary cause of the existence of the chapel, and the village has risen in proximity. We shall notice again the existence of chapels by the way- side on the Annandale Forest road, as 'well as on the Well Path. continue Aw s thp eu e hill from Whee longa n i l , easy gradienn o t backbony dr ridgee e th th f ,eo afte vera r y short distanc unmistake eth - able signs of a turf dyke cross the road. Whether the embanked area is the remains of something permanent or is merely some old temporary erection is difficult to say, but it has all the appearance of being nothing more than a casual fence. As the road at this point has crossed the 1000-feet level, cultivation has ceased, and the rank grass and boggy soil make the course of the road difficul follow o t s quitti t I e. evident tharoae s beeth t dha o n s frequently washed out that tracks have been made parallel to, crossing, and rejoining it all the way up, and even at present the true course s veri y indefinite n thiI .s localit e s soibee th yha o frequentl ls n y washed awa sectionn yi s acrosthae se roade t th sn ther,ca thae e on t neves ha r beepavingy nan r eve under-structur,no n na heavf eo y stones. The road at every point shows no signs of anything but being a beaten track over virgin soil. ANCIENT BORDE1 R21 HIGHWAYS. e roaAth s d approached Wheelrighea y ove la n opea rt i dn moor. Wheelrighea e righth littls i n t do e more thafee0 nt10 abov roade eth , while Neides Law, hale left a th mils aboufi , n e sameeo th t . This si only importan s showina t g that this passage throug o e hills th hs i s open that it could neither be blocked, nor could an ambush of any kind be laid, and therefore would be safer than the adjoining narrow pass of Note of the Gate. Uthio pt s poin e roas beeth t ha d n followin a well-defineg d ridge, ans reachedha e watersheddth . Looking forward amon e roundegth d grassy hillse look a definiton ,n vai i r s fo n e objective. Ruberslaw— 9 miles away—is e roavisibleth dd seeman , o bent s than di t direction, grouny e ridgedr keepinth e f do th . o Foe gnext th r t mil e roath e d descends across loos es mori bogs r d leso e an s, visionary t aftebu , r crossing a burn it is more distinct on dry ground, and keeps fairly near the ridge for about 2 miles, passing a tumulus a mile on. Most of the way the road is quite well denned, and one has no difficulty in followin t occasionallye ; it g t i ydisappears s easili t ybu , picke. up d On reaching a wall at the neck between Pedens Cleuch and the Wolfhopeburn, the Ordnance Survey make a curious omission. The s followina p e right e th walma roath go e ,t s lshowi d theth n no n bendin ge summi e hilsharpl th t e lef 112th a lo th t f t 4 o to y.feett t bu ; e whehill-slope e lookth on t na s a perfectl, y clead well-markean r d track goes straight up the hillside to the summit at 1124 feet. From this poin e roath t d continue , makinon s r Bonchestergfo , joinine gth main road by Note of the Gate at Cleuchhead. The distance from the Scottish boundary at Dead water to Cleuchhead is just 8 miles. Having thus describe t existi e roa th ds da s to-day cannoe on , t fail remaro t k thae Ordnancth t e Survey have planned this roa thein o d r map as coming up the valley in a straight line, like a first-class high- way, curving round in a fine sweep that would suit a main line of railway. It is this mistake that has led antiquaries astray in regard .to the real purpose of the road, for attention has been' diverted from e perfectlth y plain fact thiy b s s fanciful line. For, whe e reverseon n s the route, and descends from Wheel Church along the obvious line of road down the ridge in a southerly direction, a 'perfectly well-defined e lef o Deadwater,e th t rightt th d o t roao an ,t t dt no goebu , on s without any difficulty whatever one finds oneself walking along a well- worn track, which, unhappily, the surveyors have marked "Catrail" or " Picts work ditch," and so misled a few generations. What we have here in the Wheel Causeway is the main and ancient road from Carlisl o t Berwicke , referree o th twict d n i e Itinerary of Edward the First. The route is a simple one. Going 212 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924. from Carlisl p Liddesdaleu e o y sizstrea n s ,crossean i e f mo d after Kershope Burn, and the Liddell Water is crossed near its source. At Whee e roaridgey th ldr d e ,keep th and n followiny o sb , e liny th gb e Abbotrule, Jedburg s reachei h d without ever crossin a riverg . This, indeed, is practically the only possible way of reaching the at one time important seaport of Berwick from Carlisle, as the Northumbrian Hills lie so much across any route on the English side, and necessitate so many rivers to cross, that, perforce, the Jedburgh route is the only one possible. In this way, Jedburgh Castle and Roxburgh Castle lay associated as the barriers to Berwick, for they guarded this high- way. With these two castles in English hands, soldiers could march freel ye other froth e tow on mo t .n Were bot Scottisn i h h hands advance from either bass futileewa . The true course of the highway known as the Wheel Causeway is therefore the well-defined track coming up from Newcastleton, leading o Wheet le leadin Churchon e o Deadwatet g Th . d Tynedalan r s i e merely a branch road. Wheel Church would therefore appear to have e hasamth d e relatio o Wheet n l Causewa e e chapeth th n s o a ly Annandale Forest road, above Moffat, had to that lonely highway, or e chapeth l nea re Wel Crawforth lo t Pathd ha d. As to the course of the road from the summit of Wheelrig to Jedburgh, one cannot help thinking that the present route by Bon- chester was not the ancient road, for it is neither direct nor likely. t fro e poinBu mth t mentione e necth kt da betwee n Pedens Cleucd han Wolfhopebur e roath n d e cadividesnon follod an , wa perfectl y well- defined track via Lustruther and Abbotrule, which carries one forward without crossin streay gan f sizmEastea o vi e r Fodderle Jedburgho et . Mr Lain .1d , gan however , followe a fain dt quitbu t e definite track to Doorpool, and one could imagine no difficulty in the road continuing on past Abbotrule Church to East Fodderlee, as this makes an almost direct and dry road to Jedburgh. The Wheel Causeway therefore appears to be a portion of the ancient road from Carlisle to Jedburgh. At what date it was abandoned, and the road through the narrow neck by Note of the Gate was brought into use, I am unable to trace, for the earliest authority—Ogilby's Itinerary (1665)-give e routo detaith n f es o u ls between Castleton and Jedburgh, and "Woolie," mentioned in it, is so nea boto t r h roadeithee b y sr tha 160n ma routei e Whee t s 4i th t A . l Church was united with that of Castleton, one would think that the road fell into disuse a little earlier than that period, for with the abandoning of the road or village the necessity for the chapel would cease. ANCIENT BORDER HIGHWAYS. 213

THE ANNANDALE FOREST The golf-course above Moffat is intersected by what is locally called the "Roman Road," and while in the locality no doubt exists about s originit e evidencth , s muci e h more slender tha imagineds ni e Th . late Dr James Macdonald, in the Proceedings of the Society, vol. xxviii., 1894, deals wit e hs showi th thi t i n o sns a roa r n detaili dfa o s , Ordnance Survey, and I propose to describe the road beyond the point at which he stops (fig. 4).

V ,/ ~" ^

\ cX??^R-?*D --' £ -^lz^^S:va.n <- F O *$?;IJ~=~ —Water^f^^^'

Fig. 4. The Annandale Forest Road.

At the road between Beattock and Moffat, passing between Loch- house Tower on the left and the Standing Stones close to it on the right e roath , d take a straighs ridgey dr t e ,cours th passin p u e a g small quarry, and then, between fences, to the golf-course. Here its e hard-beateth linn o e n soi unmistakables i l t continuei d an , s steadily upward y grounddr n o o schapelt , e t wit lefte th ridgbu th h, .n o e Her e Wheee th again n i ls a Causeway, a goo d dan , many othed ol r roads, we find an ancient chapel beside the road, built by the Knights t JohoS f f Jerusalemo n e roaTh d. continues onward, well marked, on the grassy slopes, for a mile and a quarter, then bends to the right, crossing the Holehouse Burn exactly at the point where the moorland e streacoursth f e mo rocketh endd y an sgle n begins. Continuing onwards on a perfectly straight and decided course over Archies Hill, it now keeps on the ridge, uniformly 15 feet wide, and with pits at the side at regular intervals. Local information says that these pits were dug to obtain stones for the repair of the road; but as these pits are Charter of Annandale Forest, Nat. MSS., Scotland, , "volIxx . forbi. i . tha . . dt anyone

througo g 1 aforesaie hth d forest unles straigha y b s t road appointed" (A.D. 1124-53). 214 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924. visible along Dere Street at many points where the ground is hard, this explanation doet seeno sm adequatee tb o t i maye tha B s a .t , these pits are quite regular along this part of the road, but there are none further on, though the circumstances, are quite similar. At the Auldhouseshielmos e presenth s t main road from Edinburg o Moffat h t is crossed, and from this point onwards the Ordnance Survey ceases to be reliable, for the old road can be seen quite unmistakably descendin e otheth bur e n d risinrth o nsidean o p gt gu , crossine th g present road again a quarter of a mile further on. It is here that we can observe that there have been no less than four roads of different dates leaving Erickstane. (a) The first and most ancien ridge th es i t roa have dw e been describin Romae "s gth a n Road." Next came a road (6) through Moffat up the Annaii valley to Erickstane Farm, then straighe steeth pp u t hil o thit l s pointo T . ease this hill, probably about 1770 a lon ) ,g(c gradual ascent from Bridgend mad .excellenn a e d comparativelan t y easy climb d thian s, s followe e constructiowa th y b de presen th f o n t roa ) abou(d d t 1820. All these mee t thia t s point. left e Strikinfine th hile w ,d o th t l ourselvep gu s almost immediately maza n f i roado e s crossing, recrossiiig d obliteratinan , e anotheron g , and it is really impossible to disentangle the one from the other until we find the key a mile further on, where the different types and widths are visible side by side. This enables us to say quite definitely that the first road went in a fairly straight course almost to the top of Erickstane Hill, and then forward almost straight to the edge of e Beeth f Tub o makt closo ,s s a ee understandable e storth th ef o y Highlander who escaped from his guard by -wrapping himself in his plaid rolled an , d ovee edgth r e dow e hillth n . Her e ancienth e t road crosses into Lanarkshire, and continues a hard-beaten track, 8 to 10 feet wide, belo grasse wth y moor vera , y short distanc e easf th o t o et the old road shown in the Ordnance Survey at Tweedcross. The second road zigzagged acros e firsth s t a verroa r yfo d short distance, then kept round the neck to the left of Erickstane Hill, rejoined the first at the Beef Tub, and then continued on a course a little lower dow e hilth nl tha e firsth n t road, mostl fee5 1 y t wide. o unittw ee jusTh t beyon o Tweedshawst d n Tweedcroso o g d .an s A short distance furtheo separate , whertw on e r e oldesth e th , t road strikes over the dry ridge to Tweedhope, while the later road keeps down the side of the hill to the left, fairly close to the present line of road. From this point onwards the present road and the old road cross and recross each other, but the general line down the valley is much as at present. Whether the road was always^ on the west side ANCIENT BORDER HIGHWAYS. 215 or ever passed down the east side is difficult to decide; but the presence of Tweedsmuir Church on the east side of the river, and what seem to be appearances of tracks on that side, make one inclined to leave the matter open, for unquestionably the west side used now has the smaller streams, and would be least troublesome in winter. Reverting to the road at Erickstane, where we leave the present main s roadher i e grea t i eon , thate th problee troa th s i df mo presented . e map th e For" Romanth n s o , " roa o dowdg doet n no s the Tweed valley, but crosses the hills to Elvanfoot and Crawford, and continues down the Clyde valley towards Lanark. One would have expected that examinatio e surfacth f no e would have shown i n which direction the road originally went, and that it would be seen thae roaon td wena branche othes t th t wa Bu straighd r . an n o t o evidencn d an e, Y coul e for a e obtaine th f b d mo n i e fors th dkwa r othero y . wa e froon t mi Fro n antiquary'ma s poin f viewo te westerth , n road presenta s remarkable serie f problemso s , because although ther e nominallar e y two roads shown on the maps, in reality there are three. For one of the a doubl ms i e roa t manda y points, sometimes touchingn o t bu , a slightly differen d beeha n t ti mad a supplementarlevelf i s a es a , y trac r somkfo e purpose t beforBu . e going intdetaile oth f thio s s part s e makrouteu oth f t a e survee poindistrictle , th th o f t d to y an , whic hroaa d migh expectee b t leado dt . Here is a range of hills nearly all 1500 feet high, and occasionally risin 200o gt 0 feet, stretching across Scotland. Ther e onlar e ya smal l numbe l f ordinarpassesal o r d an , y traffic from Dumfriesshire must pass through one or other: Wheelrig, 1350 feet; Note of the Gate, 1238 feet; Limekilnedge, 1160 feet; Mosspaul, 890 feet; Eskdalemuir, 1100 feet; Birkhill, 1120 feet; Erickstane, 1330 feet; Beattock, 1028 feet; Well Path, 1290 feet; Enterkin, 1890 feet; Leadhills, 1530 feetf O . these, Erickstane, Mosspaul, and the Well Path are the only ones that give a fairly direct connection between northern and southern towns. Other passes are to be found without doubt, but none makes such a fairly direc d unmistakablan t e highwa s thesya e thre n whici e e th h traveller can go without fear of losing himself among the hills, and these passes must, from time immemorial, have beee highwayth n s between the north and the south. e entirt I ith s e absenc f thio e s featur f directneso e s that makes this road from Erickstane to the Clyde valley so perplexing, for one traverses it with a sense of complete bewilderment as to its direction feele On s . lost amontime on t e seeme hillsi t b gth A o .t s 216 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924.

making for the north, when it suddenly turns west, and one has to traverse five mile f tracko s , sometime somea d hilside f an on leo -n o s times on another, with no objective in view, crossing into different valleys, instead of keeping to one definite line. Contrast this with the simple route up Annandale, where the road follows an almost straight line to Erickstane, then goes straight forward into the next valley of the Tweed, and continues down it till Broughton is reached! I have e slightesth t no t hesitatio n sayini n g that this westerne th roa o t d Elvanfoot, fro ma topographica l poin f viewn impossiblo ta s i , e main road frod s positioan ,m it e hill th sn nresembleo roao sn f antiquity do . Any traffic to Clydesdale and Lanark would go by the unmistakable easd an y route down Tweedsmuir thed Broughtoy an ,n b Biggad nan r to Lanark. I feel sure that this road is nothing more than a way to e leath d mine t Elvanfooa s d Leadhillan t s from Annandale thid sHa . road primarily mean reaco t t e headwaterhth e Clydeth f t o i swoul, d never have taken suc hwindina g s clearlcourseha t I y. bee aftern na - development, and one feels inclined to think that in the double road— onln whicca y e conjecturhon havy ma ee been designe keepo dt ' apart the traffic going in different directions—some military hand has been time t worr anothea on eo t ka r makinCrawfore th r gfo d district. Reverting to the place where this road turns away from the Tweedsmuir road at Erickstane, the oldest part of the road strikes off in a north-westerly direction, and, on crossing the Lanarkshire boundary, at once becomes a twin road. This does not refer to the fact that the later road run st tha parallebu t, it thi o t sl particular roa doubles di d an , ther actualle ear y three roads sid sidey eb , consistin earliese th f go t road, 8 feet to 10 feet, its twin alongside, 8 feet to 10 feet, and lower down maie th l road feet5 fee1 2 ,1 ,o t tabandone 182n di 1 when Telford's roay db Beattock was completed. At a small stream one mile on, the original road (8 feet wide) keeps up the hill, turning more northward, and descends to Rowantree Grain, where it turns direct westwards, merging shortly into the later mail road, whose separate course, curiously enough, is now almost entirely obliterated, whil e ancienth e t trac stils ki l quite clearly marke hillsidee e remainth th n fact n dr o brokeI a .fo f ,st o werno n t ei bridg seee eb thanquartea n ca milta f o er lower dow e burne nth on , could scarcely believe tha maie th t l roaeved dha r existed t thia s point. At the next burn—about half a mile further on—the roads separate, the mail road keeping lower dow e hillnth , perfectly clearly marked, about 16 feet wide weld an , l engineered anciene Th . t road turns westwards, then north-west to the ridge, and it is upon this section that we see so roado plainltw se sidy th side—th y eb hardeo tw wited e loweran h th f ro wheel marks, the other perhaps 6 inches higher, sometimes more, sometimes ANCIENT BORDER HIGHWAYS. 217 less t neverthelesbu , roado tw s s sid sidy b e distinctnes e Th (fig . .5) f so the two can best be seen on the ridge, where one goes straight through some muddy ground, wherea twie sth n road skirt same sth e spot. Crossing the shoulder of Erickstane Hill, the road crosses into the Fopperbeck valley descendd e buran ,deeply-marke a th n ni o t s d track. s normait t Thino ls si condition s beeha nt causebu , haviny db g become mor r leso e a watercourses eacd an ,h heavy rainstorm ploughs deeper e hill reachinn th fooe o ,f th int o te soilt oburnth e A .gth , thero n s ei sign of a bridge or of anything of that nature of a river-crossing. On

double FigTh . e5 . road—Moffa Elvanfood an t t Road.

the ascent on the other side the road has been completely washed away hundree firsfow th rfe t e placeth d n yardsi s wherd an ,e washout eth s cros tracke sth , thersigo n f pavin no s e i r evego n bounding stones, such s wera ee Roma founth n o dn roadnothins i t i ; g more tha ngrassya , hard-beaten track. The road resumes its course up the shoulder of the hill, wit a secon hplacf e i sign on s dea t sroada branched ,ha d off uphill; t therbu nothins ei connectioy sign o ther d an n follo o g f t an e s o , ear wup n again further on. A mile short of Little Clyde the ancient and the mail roads unite, the latter fee6 1 fee2 t 1 o ,t broad t , wel froy e lasl wa mth marke t e th l dal junction, and from this point onward all trace of the ancient track is presen it covere n i p t u dfeet0 widt 2 f . o h This roa s saii d o t d 218 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, 1924.

continue on down the right bank of the Clyde to Lindsay Tower and Cold Chapel, and its course is quite visible; but here again, at every washout which crosses it, there are no signs of paving or stone founda- tion; nothinfoune b t firm o bu t d s ,gi hard-beaten soil. This course roath df o edow e Clydnth e valle generalls yi y supposed to be one of the main routes in Scotland, but as it joins at Crawford the very, ancient highway that led past Biggar and the Well Path to Durisdeer, Penpont, and on through St John's Town of Dairy to Minigaff Whithornd an s nevei t i , r likel havo yt e been anything more thana connecting road between the two chief highways to Dumfriesshire. There is one question, however, that will naturally be asked, What "e Romaoth f n Camp t Littl"a e Clyde whad e th an , tn o bearin t i s gha e absenc roadth y ver n sa f I excavatioyo ?e n littlca e on than s i t positiv a cursore t aboubu , y it tglanc e shows thao streamtw t n ru s through it. One of fair size, cutting off a small section and leaving it isolated from the rest, suggests that if it has ever been a camp, the choice of. position has been amazing, for the dry ground lies to the south of the camp, and it looks much more like the protecting bank that had once been erected roun dplantatioa keeo nt e cattl pth e fro treese mth , e countryothee parth y th f seee an n o r tb O n handni .n succa t i ,s ha may have been some hastily constructed rampart erected for a tem- porary purpose, and the break in the banks for the entrance and exits of both streams may not have mattered, but there seems to be on the north side sign whaf so t might have bee ntraversa e outsid openinn ea g in the rampart. This, however, may be accidental formation of the ground. Taken as a whole, one feels very sceptical about its having ever been a Roman camp..

THE WELL PATH. One of the great highways of olden times was the direct road from Edinburg Whithoro ht Gallowayn ni remarkabls i t i d mucan ,w f eho ho it is preserved to-day as a hard-beaten track alongside or parallel to the present road. It is a testimony to the heavy traffic along it in, ages past, tha s coursit t e has hardly ever been obliterated even thougs ha t hi r hundredfo e us yearsf f so o bee t .n ou From Edinburg Carlopo ht e w s are constantly observing its course beaten out on the shoulder of the Pentlands; o abov tw roadd d ol e an ,Wesn a s a t Lintoe us s stili n t i nli old bridges, one with an inscription, testify to its importance. As far as Biggar we are either on it or see its course beside us; at Lamington it is the same, and we only leave it when we cross the Clyde at Clydes- bridge on the way to Abington; but it is still the farm-road that runs ANCIENT BORDER HIGHWAYS. 219

besid railwae eth o Colyt d Chapel, beyond whic e Clyd fordehs th ewa d betwee e Kirkton th ther d f Crawfordean no mila , e befor presene th e t village. From this poin t weni t t straighe hil e bacth l th f oveko p u t r Blvanfoot—moro t e tha course non e being still visible hillsideth n o e — l betwee vallee e Powtrai co theth d th e f p an ynth o u n o Welt l l Hild lan Durisdee r Penponty Hillb n o ; , Tynron, Moniaive t John'S o t , s Towf no Dairy d thencan , o Bridgt e f Deeeo ; then fee0 1 fee9 t, o t wideta n i , valley paralle presene th o t l te bac roath f Murray' ko t da s Monument, to Minigaff, near Newton Stewart, where Queen Mary's Bridge spans the little stream. Much of this route is recognisable in the Itineraries

3MILES

Wele Th Fig l. Pat6 .Enterkin d han .

of James IV., and we find along it the record of ancient bridges erected at various dates, many of which have disappeared. The chief point of interest in this road is undoubtedly the short lengt f roao h d know e s Wel"a nth l Path" between Durisdeed an r Elvanfoot (fig. 6), where it crosses the watershed between Lanarkshire Dumfriesshird an t 129a e 0 feet. Starting fro e Durisdeemth r ende th , present grass road makes a well-engineered ascent of the pass by an almost uniform gradient, and has all the appearance of a road made in the eighteenth century—probably one of the many benefactions of the Queensberry or Buccleuch families. At the summit this excellent road ends, and we have here one of those unhappy transactions, where the splendid work accomplished by proprietoe on s renderewa r d almos te adjoinin futilth y b e g landowner s primitivroaleavind it ol n di e gth e state. 220 PROCEEDINGS OF THE SOCIETY, FEBRUARY 11, But it is in the ascent of this pass by the well-engineered road that we see below us on the other side of the valley the ancient track. It ia narros y witwwa h divergin d converginan g g subsidiary tracks, keepin e drth y o groungt e wes th e burntn do th sidcoupl A f . e.o f eo centuries have passed s bee s clearlsinc s lina ha usen it n i s t e i ei t y,ye marked as it was in olden times. Looking across the valley, a curious phenomenon is observable, for the hills on the western side are green clad to the summit, while those on the eastern side are covered with heather. The old road kept to the clean, grassy side; the later road is cut through the heather.

Wele FigTh l. Pat7 . Deed han r Castle. mila s i et i nort t f DurisdeeBu ho r tha mose th t t interesting feature pase o th fexhibiteds i s , viz e grea.th t earthwork know s Deena r Castle .(fig. 7). Here on the open hillside just above the road, and commanding massiva s i , it e rampart, squar shapen ei , wit entrancn ha e openine th o gt north, guarded by a traverse. It is not near the summit of the pass, and one can only conjecture that it has been erected at some period for the purpose of preventing some force from the south passing along this road into Lanarkshire. Lower down the valley and opposite Durisdeer, a Deil's Dyke is marked as leading along the base of the western hill, but the local information is that this was a water-lead in connection with the old castle, opposite the village. e summiA pase th e Weltth th sf o lt Path strikes inte heatheroth - land, and the road goes forward between banks 8 feet to 10 feet wide, ANCIENT BORDER HIGHWAYS. 221 crossing various burns without bridges roae Th d. itself, whic beates hi n har t apparentldbu t pavedyno onls i , y abou fee7 t inche6 t s wide. There is, however, a well-built little bridge with a span of about 6 feet over Cleuce th h Burn, wit hroaa d feetwidt2 1 .f ho From this point onward the old road crosses and recrosses the present main road, but none of d wheebridgee ol th e anciente l ar sth track d sucn an , o s h portions a s e visiblar e see mfee 4 e roa inche6 th t dd itselsan f abou fee7 t inche6 t s broad. Wherever it has had a washout no signs of paving are visible; merels i t i tracya k with stones trodde. nin It is to be noticed that the road passes Durisdeer Church, Kirkton beside Crawford, and Cold Chapel, a mile beyond Abington.

.THE ENTERKIN. Sinc openine eth roae th df g o betwee n Leadhill Sanquhad san e th y rb Menock Pass, the Enterkin (fig. 6) has fallen into disuse, and at the present moment the eruptions of gravel on the sloping hillside are slowly covering up what at one time was an important highway in connection with the Lead Mines. Though little known to the general public, it has attained a fame far greater than almost any other valley, owing to the charming account of the Glen by Dr John Brown, in connection with the rescue of the Covenanting prisoners in 1684. The path starts hig nea p watershehe u th r d fro e bleamth k village of Leadhills wit s cleanhit , tidy houses passed an ,e Smeltin th s g Mill. Although at points partly washed away, it is a firm, clearly-defined beaten gravel road 10 feet to 12 feet wide winding along the Moss, and overlooked by Green Lowther. In a mile and a half it becomes a grassy track, quite unmistakable on the hillside. At no point does the road coincide with the county or parish boundary which runs close beside it, and this in itself shows that the road has no antiquity, d road t ol beei t wouli n d , fona ha r d have e th bee s a nf o mad e us e boundary. The level t whicsa roae hth d proceeds mak realise on e e tha t musi t t have been almost constantly in the mist, for, starting from the 1300-foot level at Leadhills, it climbs to 1750 feet at the county boundary, and at the summit below Lowther Hill it has reached 1890 feet. Here on a narrow col of grass-covered hillside overlooking the winding Menock Pass roae ;th d suddenly bends over inte Enterkino th looe w k d dowan , n a steep, V-shaped, grass-clad valley, blocked at the end by a steep precipitous but green-covered hill—the " Stey Gail"—which seems to lie across the defile. The road can be seen steadily descending in an almost straight line, till it reaches the level of the burn at the foot a mile away. PROCEEDING E SOCIETYTH F O S , FEBRUARY 13, 1924.

Th t e vervalle e sideno th y e bure f steepsyo th ar s abount i fee 0 bu , t30 t below on the left, at the top of the pass, arid if the grass is wet a slip migh seriouse b t . The road at this point is 10 feet to 12 feet wide, but, further down, a hundred year neglecf so permittes ha t gravee shelvine dth th f lo g hillside graduall slido yt e ove scree—half-wae highwaye th rth t a d ,an y down—the pat s barelhi yfooa t differena wide d an , t point s scarceli s y passable, though one can easily scramble over the loose gravel. A mile down, on reaching almost the level of the stream which runs down the valley, a f grounrocko t ybi s encounteredi d d betweean , o rockstw n , through whic roae hth d passes s neveha t ,i r been more tha fee0 fee1 n8 t o widet t ; it is here that tradition places the site of the rescue of the Covenanting prisoners. On reachin bure gth nlittla e lower down, -the road emerges froe mth narrow Enterkie parth f o t goed nan s forward alon e flagth t bottof mo th°e valley for nearly a mile, badly marked, and evidently crossing the burn several times, but on reaching Glen Valentine it strikes up the hillside agaia stee y nb p gradien e hilld theree ridgth an th , f o eo -t t after is a well-marked broad road again. Continuing along the ridge between the Enterkin and the Dalveen Pass, it soon merges into a modern road, and evidently has continued on towards Morton on the Dumfrieso t y wa . Reviewing the appearance of the road and looking to its summit being 1890 feet above sea-level, its use in winter would be very limited, whils s constanit t t liabilit o misd raiyt an nt preclude t i fros m ever having been an ancient highway in general use. Without doubt it has been a specially made road to carry the lead from the mines to Dumfries, s beina coas e e th g oag o trt s town looit y n shipmentr ko ma s fo e w d an , coincident with some effort to assist the disposal of the lead during one of the periods of activity, such as 1512. In a list of distances dated 1646, the road appears to have been the main highway between Douglas and Dumfries. There is one historical note I should like to make in connection with this roa regar n dincideni e th rescue e Covenanterso dth t th f f o to e t A . summie pase justh d th san f t o twher t turnei s intEnterkie oth a s ni shallow trenc breastworr ho k commandin passe gth . Although nowhere near the spot where tradition assigns the rescue of the Covenanters to have take remarkabla n n placei s i t i , e situatio defencer nfo . Situatet da ascentmile-lone x si th f n i o ,e d whegon en horse e e nth th s woul blowe db n spenn thasuddeo y me (s ant te an dth n active movement woul almose db t impossible) s place i e rifl t th ,tha o pi eds t wouli t invisible db e froe mth road until quite close, and retreat could easily be made in several ANCIENT BORDER HIGHWAYS. 223 directions. The steepness of the hillside precluded any divergence from the road in the pass up which the soldiers would toil, and it would be almost impossibl spuo et spenra t horse agains barriere th t , while there mila kiny r coveo ean n dfo f bacs o r o whicwa kt e soldierhth s could retire. Looking to the position of the trench and breastwork, which fits in wit descriptioe hth actuae th f no l events wit r greatehfa r probability than the position of the traditional site, it would be interesting to find if by any chance tradition has fixed on the wrong spot for the scene of the rescue. It is at the foot of the valley where the road passes between the rocks that tradition has stated the rescue took place; but one finds difficulty in reconcilin positioe gth n chosene horseth r sfo , would the frese nb o t h star e climbth t , retreat under coves quit wa re soldierse th eas r yfo , whil e rescuine th tha f o t g party above would have been impossible, as they had no cover at their back. Indeed, at this point it would have been possibl seno et ddetachmena t acros burne sth , and, ascend- ing higher on the opposite hillside, render the rescuers' position unten- e valleth f suco o yn hp to manoeuvrin e th able t A s possible. gwa d an , r thifo s s i reaso t i n e stor thath e facte wort yf th ar to s h examining carefully.