HERIOT'S DYKE by ANGUS GRAHAM, M.A., F.S.A., F.S.A.SCOT.

TRADITIOE TH . I N HERIOT'S DYKE markes 6-ina ,e th .n dO.So . map, lineaa s 1i r earthwork runnin. gE and W. for rather more than a mile across Moor (squares 7048, 7148). The Ancient Monuments Commissioners, who surveyed in 1908, held it to be considerably longer, as they carried it, with interruptions, from Haerfaulds (574500), near Blythe, to the Greenlaw-Duns highway (A 6105) at a point (721484) near the demolished house of Flourish walls2 (not marked on the 7th series of the i-in.

Mileo t s 5 o

FIG. i. Approximate position of Heriot's Dyke (dotted line) as show Chalmersn no Nortf o p h ma Britai' Caledonia,n i . 1807ed , enclosee Th . d57 oblond i, an betwee 6 g5 represent. npp e th s area illustrate fign di .2

•"'Berwick; upon Tweed « O.S. map) distanca , somf eo e nine miles literarA . y tradition,3 however, creditt si with a length of something like twenty-three miles, placing its eastern end on the Whiteadder threr Wateo o e miletw r s abov latter'e eth s junction wit Tweee hth n i d- one version at Hutton4 - while there is even a suggestion that it ran on to Berwick-on- Tweed (fig. i). In view of this conflict in the record, and also of a tendency towards scepticism regarding such theories which has resulted from recent work on the principal Scottish earthworks, real and supposed,5 it seems desirable to re-examine all wha e evidencee se th to t conclusion d ,an sDyke'e emergth o t ss etrua e characted an r purpose. Scepticism is encouraged, in the case of Heriot's Dyke, by three principal causes.

Provisionae th f o W N l 1edition4 Shee7 6-ine T N tTh .. sheets5 referre5 , thin i NE so d t 4 pape 5 e ar r (the whole), 56 SW, 64 NW, 64 NE, 65 SW, 74 NW, 74 NE, 84 NW, 85 SE, 95 SW, all in loo-km. square XVI, XVI, , SW ISW ISE V X , SW V oldeXI e , th rn i NW edition r o V ; XI , XII s, NT sheetSE INE I sVI i-ine th XXI , .XXI, n I seriesXXII, NE . INW I INE ,INW XX , NW I XX , NW X XVIIX , X, XNE ISW 7th edition same th , e groun coveres di sheey d b (Dunbar 3 t6 smala d l)an par sheef to (Berwick-on-Tweed)4 t6 . 2 Berwickshire, pp. 95 f., 117, 123. The spelling 'Flourishwells' appears to be a misprint. principae 3Th l reference Heriot'o st s followss Dyka e ear : Berwickshire, xxxvi f.5 ,,9 117, 123; Chalmers, , CaledoniaG. (1807), i, 162, 243; 11, 211 Berwickshired ; Ol Gibson n A , R. , Town, ff.8 ;24 H.B.N.C., , 114v ; vi, n, 17; x, 311; xvin, 115; xxv, 201; xxvi, 360 ff.; xxix, 74; N.S.A., H (Berwickshire), 43 n., 72; P.S.A.S., ix (1870-2) f.9 ; 46 ,L V (1920-1), 240, 253; Slot. Acct., xiv (1795), 51. 2n tH.B.^.C., xvin, 115. 5 Especially Roxburghshire, App. D, on the Catrail; Selkirkshire, No. 182, on the 'Military Road', and App. A, on the Picts' Work Ditch; P.S.A.S., LXXXIII (1948-9), 174 ff., on the Deil's Dyke in Galloway; LXXXVIII (1954-6), 13Deil'e th 7 fF.n so , Dyk Dumfriesshirn ei Ayrshired ean . 227 228 PROCEEDINGS OF THE SOCIETY, ig6l-62 The first of these is the smallness of the base on which the edifice of tradition rests. trus i t eI tha bees gooa t ha ndk expendedeain time f d Dyke o l on eth an t en a do another, but, when the various reports are compared, the conclusion is plain that almost the whole story derives from a certain John Spottiswoode of Spottiswoode, whose theor a 'long f yo ' Heriot's Dyk s firsewa t printe n Chalmersdi ' Caledonia.1 Spottiswoode is there said to have written a manuscript account of the work about middle eighteente th th f eo h century, tracin g'ramparta fossed an , , fro mBritisa h strength, called Haerfaulds, hillmilea o n tw ,o s north-wes Spottiswoodef o t , through- out the country, to the vicinity of Berwick-on-Tweed. In that age, it was, in various places, very discernible know s people name th wa th Herrit's-dike.f o d ny et o e b an , ' The 'survey of this curious remain' was communicated to Chalmers by Spottis- woode's soninternad an , l evidence suggests that Chalmer probabld sha y seee nth manuscript himself2; but nothing is known as to the source of Spottiswoode's in- formation ther d facn i an , s tei some reaso suspeco nt simply t thama ye h thav e been repeating a piece of local lore, seeing that the minister of Greenlaw, writing twelve years before Caledonia was published, reproduces a 'tradition' which resembles Spottiswoode closely. The only other potentially important point emerging from earliee th r record thats si 1834n i , Dyk'tracte e sais th th , df evisiblwa e o 3'stilb o t l e ,f o . N villagnortr e thougfa th f hw o e h ho doe t appearsno theore Th . y 4 a wholsa e must therefor hele eb d eithe refleco rt t folk-memory, whic unlikels hi o yt be accurate in details even if well founded on fact, or on the reported opinions of a single perso recogniseo n f no d standin antiquaryn a s ga , who, moreover, livea t da notoriously uncritical period. It is easy to imagine, for example, how Spottiswoode might have been influenced by Alexander Gordon's account of the Catrail, which

publishes wa 172n di pictured 6an d that monumen kina s barbariaf a tdo n limes some 5 fifty miles in length. seconA d reaso cautior nfo acceptinn i g Spottiswoode' intrinsin s theorow s it cys i thinness. Whatever may have been visible in the middle of the eighteenth century, and Chalmers' use of the words 'in various places, very discernible' implies that the work was not even then 'discernible' as a continuous unit, since the coming of the Ordnance Survey, at any rate, the gaps have far outrun the remains of walling or earthwork. Details will be found in Part II, but, to anticipate results, it may be said here that, in the twenty-three miles of the Dyke's supposed length, all the actual structur seee eb thann todatca y extend aggregatn a o st lesf eo s tha half a nmil a d ,ean thad thif an o t s more tha nmila e constitute self-containee son d unit. These doubts are reinforced by the evidence of Roy's map of . As this was made durin perioe gth d 1747-55 t musi , closele b t y contemporary with Spottis- woode's paper consequentld an , conditione yth s tha t recordi t s should presumably correspond with those with familiar s whicwa e hh . However meano n y b s t suggest,i s tha t groune line mos th f Spottiswoode'th eo f o n t do s dykthes ewa n unimproved moorland; cultivated land is differentiated by the conventional indication of rigs, . 18071Ed , i, 243, with footnot . ef 2 ibid., footnot. eg 3 Stat. Acct., xiv (1795), 512 n. This note places the western end of the work at Boon, not at Haerfaulds, but this is probably due to confusion with the 'Black Dyke' recorded in Berwickshire, No. 246. 4 N.S.A., ii (Berwickshire) . 72 , 6 Itinerarium Septentrionale, 10. 2f HERIOT'S DYKE 229 and on this showing much of the area would seem to have already been under the plough. The main exception is the moorland lying between Hallyburton and Marchmont. particularls i t I 1 y important her noto et large eth e amoun arablf o t e ground shown SE. of the Blackadder Water, in the stretch downstream from Green- law, as this tends to undermine Spottiswoode's value as an authority for the former extension of the Dyke all through this region. t anotheYe r sourc suspiciof eo n frequene existth n si t failur observerf eo diso st - tinguish between ditch-and-bank earthwork roadsd trus i ol t ed I . thasan t theso etw classe remainf so sometimey sma puzzlingle sb y alike, especially wher ditce eth f ho earthworn a k happen havo st e been usetraca consequentls d da kan y shows markf so wheels, but experience should eliminate doubt in the majority of cases.2 However, literature th e constantle showar o s tw thae y th tconfused, sometime extenn a o t s t

which suggests that the write3 r in question failed to realise that any difference existed between them. In the present case it is impossible to feel confident that Spottiswoode 4 could have mad distinctione eth ever o , n tha necessite th t doinr yfo woulo gs facn di t have occurred to him. A fina ltraditioe iteth mn i n which call consideratior sfo name th s nei 'Heriot's Dyke'. The form 'Heriot', actually 'Herriot', is first used in the New Statistical Account,5 writte t publishe no n 183n i t 4bu d untiadoptes e wa th l t i 1845y db d an ; Ordnance Survey, though only afte revisioe th r n carrie monumen189e n i th t 8- d ou t being shown on the first edition of the 6-in. map simply by a fine line, to which no name was attached. The Survey's authority for the name cannot now be traced, as the Name Bookrevisioe th f so n were destroyed durin secone gth d World War.e Th

form appearing in the Statistical Account of Scotland,1 in Chalmers8 ' version of Spottis- woode's report, and in the R.C.A.M. articles is 'Herrit' or 'Harrit', and it may 8 9 consequently be asked whether 'Herrit' may not really be something distinct from the well-known surname Heriot, and, if so, perhaps contain a hint as to the earth- work's true background evidencee th l Al . , however, point oppositn a n si e direction.

Thus, although the earliest recorded form of the surname, which dates from the 0

period 1164-74, 1S tne trisyllable 'Heriet' disyllablee th , s 'Heret 'Heroted an ' 1 ' like- wise appear as early as the sixteenth century,11 while the equation of 'Elliot' and

1 The principal cultivated areas seem to be as follows, when allowance is made for the deficiencies of Roy's topography: Between Blyth woodland ean d flankin Brunte gth a Burn; round abou policiee th t f Spottisso - woode House; E. of Raecleugh, through square 6151; all over what is now Westruther village; some further to the E., perhaps between Westruther Mains and Cammerlaws; on the left bank of the Blackadder Water N. of Hallyburton; probabl squarn yi Fangrise eth f 6948o . t ,W Burn ; between Marchmon Blackaddee th d an t r Water thencd an , e pretty generally eastwards througgroun w Mersee lo th e f do hth . Useful criteria are (i) that roads adapt themselves to gradients while linear earthworks do not; contrast, for example2 , Roxburghshire, (ol7 figd60 . roa Outen do r Cock Law d figan .) 615 (the Catrail): (ii) thatn o , reachin ga steep-side d gully, linear earthworks stop, usually just belo lipe wth , while roads descena o dt crossing-place, obliquel twista n o yr yo course . 3 Typical examples of such confusion are noted in Roxburghshire, Nos. 119, 129, p. 483; Selkirkshire, Nos. 95, 104, 116, p. 126.

e.g., P.S.A.S., xxxn (1897-8), 62 ff. 4 s Vol. ii (Berwickshire). n 3 4 , I am indebted for this and other information regarding the Ordnance Survey's record of the monument to M6r A. L. F. Rivet, M.A., F.S.A., of the Survey's Archaeology Division.

Vol. xiv (1795), 51. 2n Caledonia, loc. cit. Berwickshire, loc. cit.

7 8 9

0 Black1 F.. G , The Surnames of Scotland, s.v. " ibid. 230 PROCEEDINGS OF THE SOCIETY, igGl-62 'Eliot' can be cited as an exact parallel.1 Again, 'Herrit' was very lately, and pre- sumably still is, in vernacular use for 'Heriot'2; and finally 'Herrit's Dyke' does not suggest to a place-name expert derivation from any probable descriptive expression.3 It therefore seems safe to take 'Heriot' at its face value, as the family surname, what- ever it may actually imply. REMAINE TH . II S A review of the remains that can actually be seen on the ground may best be carefud fule basean th l n do l statemen f Spottiswoode'o t s theory publishe 192n di 8 by the late Mr J. H. Craw, F.s.A.scoT., which embodies his own observations made 4 in the field at the time. The notes that follow accordingly deal with the items set papers hi ou n i t, taken seriall. E yo t fro . mW

•mtt» Earthwork ««-• Watercourse suitably aligned to serve as a boundary marlo. Possible earlu areas of s- i!!!!!!- woodland Sir- moss

FIG. 2. Map illustrating possible boundary-lines of Catrail type in Lauderdale and the valley of the Blackadder Water. Circled numeral sseriae refeth o lt r number earthworkf so 'Note th Grossn n eo si - dykes . 239(p ' ) i. Haerfaulds. The first piece of structure noted by Craw is a linear work running eastward yds0 32 . r fro north-easterse fo mth fore Haerfauldsf th o t f o c nar 5 (574500). At its point of departure, the S. side of the N. entrance, the work, which is here a boulder-buil basee thic. ft th t t 4 walk,hig. a overrided ft lh4 an forte ruine th sth - f so same th et a tim wald elan curves toward alloo t produce e f lin s b i N.w e it s o est th a , d anti-clockwis lattere northerhere d th th f eAn o .y Craeb c nar w faile notico t d ea significant fact, namely tha rickla t f boulderseo yds6 2 , . long, extend. s SW fro e mth arc of the fort to the lip of the deep and steep-sided, ravine-like valley in which runs 1I am-indebte . Beattythir W dfo s r poinM , C.B.E.o t t , LL.D. . Parkindebtem W a r I 2,informatio e M M.A. th o dr t ,fo n that Heriot'sHospital football tea knows mwa n ni Hawic 'Herrit'ss ka vere th y n receni ' t past. indebtem a 3I . thiNicolaisenr H dfo s. informatioF . W r , DR.PHIL.M o nt , B.LITT. * H.B.N.C., xxvi, 359 ft.

Berwickshire, . 218 . No 117p , , where, however lengte wore th ,th gives f ki ho n incorrectl abous ya 0 25 t yds. 6 Fig . .11 p 106n 7o show relationshie sth dykeford f po an t . HERIOT S DYKE 23! the Blythe Water, and to realise that the linear work thus ends not on the fort itself adjacente th n d vero an t y, bu important, natural feature s unnecessari t I . o t y suppose tha remaine th twale th lf sthao t cros interioe forsth e th t f playero pary dan t in this process, as they do not connect either with the rickle or with the main work; their ends, saythao t s ,i t touc fort-wale hth l respectivel lattee d th f yds.an ro 8 y. 2 S 50 yds. N. of the former. After leaving the entrance of the fort, the linear work quickly degenerates into a rickle of boulders spread to a width of some 8 ft.; and about 90 yds. out this turns int ospreaa d turf ban whicn ki h occasional boulders appear bouldere Th . s wero en doubt derived from the ruins of the fort, and fewer of them were used as the distance from their source increased. The bank, in turn, fades off into a low lynchet, with a field uppes tracit n ko r side bounding rig-cultivation stild an ;l furtheline th s e i n o r marked by little more than a slight difference in the vegetation, associated with the barest vestige of the bank and a few boulders at wide intervals. All traces are lost just short of the stone dyke bounding the Blythe enclosures, at a point 35 yds. SW. of gata e (578500), thoug yard lass w it bane tfe hn sth i k again becomes visible hert i s ,a e contains a good number of boulders. In the enclosed ground within the dyke nothing whateve surviveds ha r . Apart fro msligha t angl yds0 t fro e6 forte ou .m th , made grounde th f woro e necessare th , li fairls k i e th y y straightyb curves ;it lone d sar gan gentle diffed an , r altogether fro smale msudded th an l n irregularities that result from careless gang-work. The 'trench' recorded by Craw does not exist.

Apart from the fact that this work is certainly late1 r than the date when Haer- faulds became ruinous earlies i d an r, thablocko ntw rig-cultivatiof so n which air- photographs show very clearl separatee lackingb s i o ye , t evidenc it ag y s d.b it f eo Considerable remain occupationf so t recorded adjoinine ye see th e t b n nno o ,n ca g, ground; these includ t onleno y huts, cairn smald san l enclosure lengtht also sbu o tw s of turf-and-boulder ban t differinkno g greatl appearancn yi undew no re e froon e mth discussion. Non f themeo , however actuan i s i , l contac e place-namt Th wit . hit e tell nothings su 'gres a , y enclosure' suitabla s i 2 naturad ean l appellatio fora r tnfo built of bare boulders. A plain analogy exists, however, with a number of linear earthwork Cheviotse th n si , whic similarle har y connected with clearly defined land- marks.3 2. Blythe. Craw's next item is a bank of large boulders and turf - not, as he records it,4 a ditch and bank - which survives in a corner of unimproved ground 400 yds. NNW. of Blythe, at a point just E. of a fank (583499). It is now only 115 yds. long, both ends being cut off by enclosed arable ground, and up to 7 ft. wide by i ft. high s generaIt . l appearance doe t diffeno s r greatly fro me Haerfauld th tha f o t s section, and, as its alignment would be in keeping with a westward course directed towards the latter round the upper part of the small valley of the Spotts Burn, it 1 H.B.N.C., xxvi, 362. 2 Fro harE mO , 'hoary', 'grey' (Ekwall ,Concise e E.Th , Oxford Dictionary of English Place-names, s.v. 'Har'). Ekwall remarks that the theory that OE har 'had developed the meaning "boundary", "boundary-defining" ... is not well founded'.

For forts and homesteads used in this way see Roxburghshire, Nos. 58 (i), Cauldshiels Hill; 188, Blake- billend3 ; 456 and 494, Lantoncraigs; 681 and 775, Cheviot Burn; 682 and 777, Fundhope Rig. 4 H.B.M.C., xxvi, 362. 232 PROCEEDINGS OF THE SOCIETY, 1961-62 seems probable enough that the two once formed parts of a longer whole, as was believe R.C.A.Me th y db othed an . r earlier observers regards A . possibls sit e con- nection with the fragment on the Brunta Burn (infra), nothing more can be said than that both, and also the Haerfaulds section, are all more or less in alignment; cultiva- tiorazeintervenins e nth surfacha e l dth al f eo g groundavailable th d an , e air-photo- graphs show nothing. 3. The Brunta Burn. Craw next states1 that a mound is visible on both banks of Brunte th a Burn Bruntaburpoinf a o yds0 t . a , 40 N t. n Mill. t altogetheThino s si r accurate, though there are some remains of earthwork at the place in question (597500) lefe th t bur e bann th O f nditc.a k o h wit banko htw s runs dow shorte nth , steep slope that separate e arablth s e field above fro e narrowmth , marshy strip flankin burne gth ; while opposit right-bane ecrese th thisth f o t a , k slope sligha , t hollow, which might represent the section of a vanished ditch, can be seen under- lying the foundations of a dry-stone field-dyke. The left-bank work is 14 ft. wide over-all, the ditch being 4 ft. wide and up to 2 ft. deep from the top of the southern bank, which is the bulkier of the two. Craw assumes that these remains were once joine witp dBlythu e hth e fragment, whic distans hi t about 145 0evidenco yds.n t ,bu e for this is forthcoming for the reasons given above. The form of construction seen here does not correspond with the boulder-studded banks at Blythe and Haerfaulds, the difference being evidently one of actual design and not merely the accidental resul locaa f o t l lac boulders.f ko Again woodlane th e f i , th n do markey Ro y db 2 right ban thif ko s burn existe earlien di r times boundary-lina , e could have been nru throug blazingy b t hi , entanglement, etc.,3 more conveniently tha buildiny nb n ga earthwork. No useful inference can be drawn from such uncertain data. . Craig4 Plantation. impossibls i t I e toda checyo t k Craw's statement4 tha ditchta - and-bank work was 'intermittently traceable for over 100 yds.' in Craig Plantation, 'running fro soute mth h gatewaa sid f eo Spottiswoodo yt ee 'gateway fort'th s a , ' canno identifiee areb te th undes a i d d an densra e cro younf po g conifers. Despits eit thinness, however, this record may just possibly be founded on fact, as it is supported by an earlier one derived from a resident at Spottiswoode5; while the piece of hearsay that Craw adds, to the effect the earthwork 'was said to have' passed along the fort's northern rampart, recall mannee th s whicn i r lineae hth r dyke coalesces wite hth wal f Haerfauldo l s (supra). consequentls i t I meano n y yb s impossible that Spottis- woode fort6 provided t 602502a , eastern a , n termina somr fo l e kin lineaf do r work; though whether, if so, this was or was not continuous with the fragment on the Brunta Burn must remain an open question, It must also be remembered that the Spottiswood vera e d barony ha lon s gy ha Robera life s a , Spotteswode d t mens ei - tione n 1296,di thad an 't consequently large number dykef so s defining fieldd an s holdings may well have been built hereabouts in the course of the centuries. It would, 1 H.B.N.C., xxvi, 362. The change from earthwork to masonry noted on Broad Law (Roxburghshire, No. 394, p. 189) does not

provid2 e a parallel, as the masonry is really a revetment, on a steep slope, rather than a free-standing structure. 3 On which see Roxburghshire, 483, and other references there quoted (n. i). 4 H.B.N.C., xxvi, 362. 6 P.S.A.S., ix (1870-2). f 9 46 , • Berwickshire, . 287mentiono N No . , however supposemade s i , th f eo d linea rR.C.A.Me worth y kb : 7 Cal. of Docts., n (1272-1307), p. 203. 3 23 HERIOT'S DYKE therefore unsafe b , accepo et recore th t thif do s facwors it et k a value supportins a , g John Spottiswoode's theory. 5. Westruther. Craw did not himself observe anything at Westruther to be associ- ated with Heriot's neverthelese Dykeh t bu , s suggest Eastsidey b sf n o tha ra . t N ,i t Jordonlaw Moss, and close to the village on the N. In this last point he presumably followed the New Statistical Account, according to which the 'tract' of the Dyke was still visible ther I834-n ei 1 Nothin seee b nn neag ca village th r e today other thana revetted bank, wit hdrainaga e ditch outsid , whiceit clearls hi y designe encloso dt e the plots of land behind the houses, and which possesses no feature suggestive of a running earthwork whild always i an ; t ei s possible tha remaine th t sucf so hstruca - havtury ema e vanished since 1834 considerationo , tw ther e ear s which diminish confidenc recorde firse th th t n eplacei n I . retirea , d shepher residenw dno Westn i t - ruther, whose father and grandfather were likewise shepherds in the parish and whose family connection wit place hth e must consequentl baco yg k approximatele th o yt Statisticalw Ne e timth f Account,eo state neved thaha e r th hear Heriot'f do s Dyks ea existing near the village, and in fact only knew of it on Greenlaw Moor (infra). In secone th d place obsolete th , e word 'tract'Statisticalw Ne writee ,e useth th f y o drb Account, seem approximato t s meaninn ei boto gt h 'trace d 'track',an ' mighd an 2 t consequently apply just as well to the traces of a disused road as to the remains of a dilapidated earthwork thus i t sI .impossibl confidene b o t e t tharecore th a t f do 'tract' at Westruther refers to an earthwork at all, particularly in view of the frequent failur observerf eo distinguiso st h between earthwork roads.d san same Th 3e suspicion must detract fro value mth whaf eo t would otherwise have ranke corroborativs da e evidence dating fro 1870e mth 'longcasa e 3 r .*eTh fo ' Heriot's Dyke thus receiveo sn real support from any record referring to Westruther. 6. Cammerlaws. nexe Th t poin Craw'n ti placee b s n lis t 648499da ca t , where eth farm-track from Harelaw to Wedderlie crosses a small burn running in a shallow gully remaine Th . s seen her Cray e b not e war , however, thosearthworn a f eo t allka ; he was misled by a group of three hollow tracks, evidently belonging to a roadway older tha existine nth g track, which mak econspicuoua ssoute marth hn k o sid f eo gullye th . This piec supposef eo d evidence must accordingl writtee yb n off. 7. Haliburton Hill. Craw state o goet n so 5 tha Dyke th t e reappear fore th f m o n si . flan E heathe e Haliburtotracf a e ko th th n ki n o r n Hill thad t an extend,i t e th o st Fangrist Burne R.C.A.MTh . . likewise describe a 'grass s a t yi s line amone gth heather', with no mound remaining.6 Both these observations are correct but the inference drawn from them is faulty; what exists in this area (squares 6848, 6948) is simplroadd ol n ,y a identifiabl flattee th n reo groun grase th sy db markin courss git e steee th pn o slop d ean tha Fangrise t th fall o t s t Burcurvina y nb g hollow tracf ko typical form. Ther questioo n s eearthworn i a f no k having ever existe thin do s line. 8. Greenlaw Moor. East of the Fangrist Burn there begins the stretch of ditch-and- bank earthwork tha identifies ti Heriot's da s O.S e Dykth .n emapso descriptioe Th . n

1 Vol. n (Berwickshire), 72. This statement is repeated in H.B.N.C., xvm, 115. 2 O.E.D., s.v. 'Tract', especially §§ 10 and n. a On this see p. 229 above. 4 P.S.A.S., ix (1870-3) . f 9 46 , 6 H.B.N.C., . xxvif 2 36 , " Berwickshire,. 96 . p 234 PROCEEDINGS OF THE SOCIETY, ig6l-62 , ohoweverfit , tha R.C.A.M. e gives i tth y nb somewhas i t misleading mucs a , f ho 1 it seems to refer not to the Dyke but to a hollow roadway and some associated lengths of field-dyke, which are met with at and near the Dyke's western end. To make this distinction clear t musi , explainee b t d thawestere th t n portio earthworke th f no , fro ma poin t jus e conspicuoutth Wf o . s stop-but n abandonea f o t d rifle-range (708484), approximates to the line of a small watercourse, the Snaw Burn, which, though now apparently dried up, was once considerable enough for farm-tracks which crossed it lower down to have required causeways. This small burn is tributary Fangrise th o t t Burn deea , whicf o p y gully joinst hi wa y ,b ,jus f t o opposit d en e eth roae th d mentione lase th t paragraphn di deepla d an ;y hollowed track, most pro- bably once connected wit lattere hth , flank lefs it s t stretce banth n ki h above eth gully on an alignment virtually the same as that of the earthwork ditch further E. A resulting confusion between the road-hollow and the ditch has been made still worse by the presence of turf-dykes, no doubt once the boundaries of fields, along the southern edge of the former; and the published account of the earthwork should accordingly be treated with reserve. judgee b n dca fros a evidence Amr th s fa e visible today earthwore th , k begint sa about 705484, some 430 yds. W. of the stop-butt and fully three-eighths of a mile E. Fangrise ofth t Burn apparens It . (belowf o . t terminaW fenc. a ) ft 0 e2 runnins li . gN . whict marke S O.Se no d th s h i an .n dmapso jus s i t tI .possibl e tha bane y th t kma once have extende somr d fo yds 0 e6 . further W.t disturbanc,bu e resulting froe mth hollowing-ou roadwae th f o tobscures yha tracese dth exace th ; t poin t whica t e hth earthwork begins is not, however, of any particular importance, as it was evidently mean givo t e Snae placth wo et Bur soos na this growd na sha n large enougt ac o ht as an effective landmark. It runs from W. to E. across the moor for some 1900 yds., 2 and, before its eastern terminal was destroyed (infra), it ended on the Flourishwalls Burn at 721484. Its alignment is remarkably straight, apart from some minor sinuosities greatess ,it t divergence fro mose mth t direct course being less tha yds.0 n5 ; d sucan h accuracy argues both competenc procese d carth an en elining-outf i o s . wore Th k consist ditca f so h . sidewitN s hban,a it thoug n ko placesn hi especid an , - stretce allth n ystop-butti e h justh t larga te easp th f e,ga to ther alss ei osligha t spoil- shalloa ditce . sidd th S f h an eo we ban th superficia n ko e . sidth N f l e o e scooth n po main bank. The main bank varies in breadth, as more or less spread, from 9^ to 14^ ft. and the ditch from 7 to io| ft.; the ditch is from 2 to 4 ft. deep below the crest of maie mooe wideth . th e ft n spoil-ban e f 7 rth banko th . o p t d Froto p an ,u e mks i th ground slopes gently down to a swampy hollow draining into the Flourishwalls Burn, t approachei s a d an s this hollo earthwore wth k attain greatess sit t breadt less i sd han regularly aligned than elsewhere bottoe holloe th th t f mA o wt .disappeari s alto- gether for a distance of 70 yds.; it has probably been damaged here by traffic, as seee holloe trackb th n n f risinwo sca t northwardsgou , probably headine th r gfo peat-hags formerly exploited in Hule Moss. A further gap, 20 yds. wide, which ends 3 1 ibid. 2 For the analogy provided in this respect by the Catrail, see Roxburghshire, 482 f. 3 Gibson, R., An Old Berwickshire Town (1905), 247. The commoners' rights to 'fail and divot' were ex- tinguished afte lawsuia r 183n i t 6 (ibid). HERIOT'S DYKE 235 105 yds. short of the Greenlaw-Duns highway, is evidently associated with another bel tracksf o t , which moun lefe th t Flourishwalle banth f ko s Bur nshora t distance N. of the highway; these tracks no doubt represent an old road to Duns, traffic on whic presumabls hha y contribute damagee th o dt . Slight traceearthwore th f so n kca be seen E. of the second gap, reaching almost to the highway; and though it has now been obliterated E. of the latter, this final fragment appears on air-photographs1 and was also see y Cranb poine wn 1928.i th t t A 2wher earthwore th e ke jointh , it s Flourishwalls Burn enters a small but rocky and well-marked gully, and lower down isufficientlsa y importan landmara s t a featur t ac itself f ko o et ; consequentlyn i d an , view of the earthwork's relationship, on the W., with the Snaw and Fangrist Burns, purpose th constructios it f eo n would see havmt o naturat ou ee beeek l landmarkno t s where non presents ewa . 9. Chesters Brae. The remains marked 'Ancient Earthwork' (741480) on the O.S. (Provisionap ma l Edition) were regarde R.C.A.Me th y db 'os a .f doubtful origin', 3 Crad an w seem hav o st littld eha e confidenc theirn ei valu evidencs ea supporn ei f o t Spottiswoode.4 What appears today is a terrace, up to about 17 ft. wide over-all and hollowed along length s parit f o t , which runs obliquel sidwida e f th eo e p yravinu e slopee th f ,o givep li e s andplacth t a wha,o et t resemble three alternative hollow tracks superficiae spitn th ;i f eo l resemblanc hollowed-oue th f eo t stretc ditcha o ht - and-bank earthwork, the whole may be confidently regarded as an old road. The hollowing may well have been accentuated by some slipping of the soft clay soil on slope e stretc e facth e f Th th eo . 'trench f ho ' recorde Crawy db 5 abou quartea t a f ro mile further east (746479) could not be examined in 1961, as it was obscured by young conifers and heavy herbage; but in view of what has by this time emerged regarding Craw's line as a whole, this fragment cannot be supposed to possess signifi- cance, whateve trus it r . e naturbe y ema 10. Sisterpath Waulkmill. What Craw recorded here (757479), admittedly without much confidence,6 turns out to be a farm-road, still in use, and tied in with a ford which connects wit hpublia c highwa othee th rivern re y o sid th f .eo 11. Winfield. Craw's field-name 'Blackdykes', f Winfielo . E 7 squar? ( d e 8951), cannot be regarded as significant, as any linear earthwork is liable to be called a 'black dyke Berwickshirn i ' northerd ean n . III. CONCLUSIONS From the foregoing review of the remains as identified by Craw, it is clear that they fail to corroborate Spottiswoode's theory. The review has left us with only two certain finds—the first a turf-and-boulder bank between Haerfaulds and the Brunta Burn, probably never more than a mile and a half in length, and the second a ditch- and-bank earthwork on Greenlaw Moor, rather more than a mile in length and separated fro Brunt e abouf mth o p at ga seve Bura y nb miles addition I . n there ear two uncertainties - a fragment of a ditch with two banks beside the Brunta Burn and

106 G / Scot / UK 9 / 7450-1-2. H.B.N.C., xxvi, 363. 1 Berwickshire, . 159No . H.B.N.C., xxvi , 363.2 36 ibid. 6 ibid. 4 7 ibid., 364. 6 23 PROCEEDING SOCIETYE TH F O S , ig6l-62 doubtfulla y reported length of ditch-and-bank wor Craikn i g Plantatio theset bu n- , even if accepted and regarded as having once formed parts of a single unit, would account for less than half a mile of the gap, and that at its W. end only. All the rest of the supposed evidence has evaporated, and in particular we are faced with a gap fully thirteen miles long between the easternmost visible trace and the . Thus even when allowance is made for the destruction of some earthwork monuments since Spottiswoode's day statemens ,hi t remains unconvincing. However, if the theory of a 'long' Heriot dyke as set forth by Spottiswoode is not supported by remains existing on the ground, these remains themselves possess in- trinsic interest, and an attempt should be made to interpret them on their own merits. t wilI l have been noted already thaearthwore th t Greenlan ko w Moor cono t , - sider this first, resemble Roxburghshire sth e Catrai significano tw n li t respectse th n I . first place it is far too slight for defensive purposes, and would hardly, for that matter, provid eserioua s obstacl 'preya o et stolef o ' n secone cattleth n I d. place bots hit ends rest on burns, and burns running on courses which project its general line (fig. 2). Catraie case th th f e o n lI similar features have been take suggesno t t tha functioe tth n of the work was to mark parts of some boundary-line, recourse being had to the 1 earthwork in places where no burn was available as a landmark. It will be interesting to see what follows if the same hypothesis is applied to the Greenlaw dyke. bees restsd ha en nFlourishwalle s . a ,saidth E n e o , Th s Burn, which turnst a , the point of contact, from S. to E. Thereafter it is joined by other burns, the com- bined stream being marked on the map successively as the Swardon Burn, the Kirk Howe th Bur d ecourse nBurnan th d thif eo an ;s stream, though somewhat sinuous, is in general towards the NE., the Howe Burn finally joining the Blackadder Water just upstream from Kimmerghame Bridge (806512). As the crow flies, this point is about five miles and a half from the end of the earthwork, or nearly six along the waterway. The W. end, similarly, seems to hand over its function to the small Snaw Burn at a point where the bed of the latter has attained a certain breadth, though not very deep; in its lower part, however, it becomes a deep and steep-sided gully, which opens into another considerable gully, that of the Fangrist Burn, about 670 yds. belo lase wth t certain earthwore tracth f eo k (698484) course Snae Th th . wf eo Burn is approximately from E. to W. and that of the Fangrist Burn from N. to S., but immediately below their junctio Fangrise nth t Bur rund nan s turn. inte sSW oth Blackadder Water after a further 750 yds. At the confluence, the Blackadder Water turns from its previous easterly course into the deep and narrow valley occupied Fangrise th highe y b p t ru Burn finalld ;an y arrives, after further turns which carryt i thed an n . inflofirsHowe e NE.th th SE tf t wea o , Burn describes a , d above. This point is about six and a half miles, as the crow flies, from the mouth of the Fangrist Burn. Thus the dyke on Greenlaw Moor could be represented as the artificial portion o boundara f y otherwise define streamsy db whole th , e enclosin elongaten ga d strip of country between seve eighd nan t miles lon nearld gmilean o ytw s widheads it t ea . othee th rn O hand mucd an , h more probabl e Catrai vien th yi f wo l analogyt i , coul suggestee db d tha line th te formeearthwore th burn e y designedth s b d swa kan d 1 Roxburghshire, 483. HERIOT'S DYKE 237 to 'rectify naturaa ' l river-frontier Blackadde,e thath f o t r Water favoun i , somf o r e person or group established on the right bank of the latter but claiming, in addition, this left-bank strip, whic suitabls hi farminr efo r pasturego similaA . r explanation has been suggested for the so-called Deil's Dyke in Nithsdale.1 On the other hand, if a more far-reaching explanation were desired, it could be pointed out that the Blackadder and Whiteadder Waters, together with the mouth of the Tweed, would carry a river-frontier all the way to the sea, while the Blackadder Water, upstream fro infloe Fangrise mth th f wo t Burn aligneo s s i ,s a d r thatfa s a , 665484, hal Hallyburtonf o mila f . eW t wouli , d serve well enoug prolono ht a n gi generally westward directio line nth e denne Fangrise d th furthe y b e . t E rth Bur d nan earthwork. This part of the Blackadder valley is, moreover, a conspicuous natural feature, being dee possessind pan g steep sides thin O s. showin furthee gth r coursf eo the supposed frontier might be drawn, with some approach to credibility either (i) westwards, wit hele hth earthworkf po destroyedw sno othey b r ro , means through woodlanany d whic havhmay e existed, (iior ) northwards, alon uppermosgthe t 2 t soma t stretchee ou hille awar poin rivere e th fa sdi th o n ythai f s to defini se t ,o th t - tion of a frontier had ceased to have practical importance. Something of this latter kin thoughs di havo t t e happene south-eastere case th Catraile f th e o th n di f o d , nen on the high ground above Robert's Linn. Viewed in this light, the Catrail analog3 y makes possible a further suggestion. The Catrail is currently explained as a temporary frontier-mark, drawn by Anglian settler thes sa y colonise valleTeviot,e e Pictse dth th th f yd o ' Woran 4 k Ditca s ha similar boundary covering the junction of the Tweed with the Gala and Ettrick Waters.5 If these interpretations are correct, an earthwork-and-watercourse bound- ary crossing the Merse and the Lammermuir foothills might equally be thought of as separating Anglian settlers, engaged in colonising the lowermost reaches of the Tweed, from natives pushed north inte moorlandoth s beyon e e valleth dth f yo Blackadder Water and westwards to the confines of Lauderdale. Our other certainty, the turf-and-boulder bank, is difficult to interpret in a fully convincing manner s westerIt . certainld nen yimportann resta n so t valley, which contain largsa e burn Blythe ,th e Water thit ;bu s valley lies transversel bank'e th yo t s axis, and thus suggests the end of a line rather than its prolongation. Nothing being known about the eastern end of the bank, further inferences regarding it can hardly facn drawn e i Bruntt e resteb i f ti th t n do ;bu a Burnwoodlann o r ,o d alon righs git t bank samthe , e consideration would apply thaIn . tBrunt casethe as ,a Bura nis tributary of the Blythe Water, the only function attributable to the linear bank would then be to mark the third side of a triangle elsewhere defined by the streams. This well y true factth n ma e, i e explanatiob , remainse th landf e no th Blythef s so ,a , which occupy this area, formed a unit of landed property at least as early as I2g6,6 while earlier occupatio atteste s ni cliff-toa ford a y an td b p enclosure. possibles i t I 7 , more- 1 P.S.A.S., Lxxxvm (1954-6), 153 f- 2 See p. 232, n. 3. 3 Roxburghshire, 480. • ibid., 483. 6 Selkirkshire, 127. e Cat. of Docts., ii (1272-1307), p. 203. ' Berwickshire, Nos. 219 and 221 respectively. For more detailed descriptions see P.S.A.S., xxix (1894-5), 138. 238 PROCEEDINGS OF THE SOCIETY, 1961-62 over lespoino o t n , s o that t n four cross-ridge dykesimmediatele th n i , y adjoining region of north-eastern Lauderdale, which may well have been designed to fulfil an exactly similar function.1 On the other hand, the place-name Spottiswoode,2 which existed bees ha ns a ,said t leasa , earls a t 1296s ya , suggest presence sth woodlandf eo , t somthea r no e earlie Brunte th rf r somewhero dateao . Burnn E o ,e th . o et Th e exten sucf o t h woodland cannot coursef o , guessede likels b i , t i havo yt t ebu ; been larger in earlier than in later times, and the possibility must consequently be re- cognise e Bruntth n carries do a wa Burd tha e lineat den th t bu n t r no wor d di k forward through woods for an unknown distance eastwards, by blazing, felling or entanglement. Such methods of marking would leave no permanent trace for the information of later ages. The limits of legitimate inference having now been reached, what remains to be moro sain s di e tha exercisn na guessworkn ei have possie W th .e t -a arrive r fa o ds bility of two separate boundary-works, one coming from the W. and ending in a forest somewhere between Spottiswoode and Westruther, and the other beginning an eastward course on the Blackadder Water at, or perhaps somewhere N. of, the sharp bend W. of Hallyburton. It may now be asked whether these two ends may not have been linked together, acros intervenine sth r thre o somp o gga e etw mile s wideo t , produc continuouea s frontier reasonably comparable wit Catraile hth . That this might have been effected without recours earthworko et s will appear fro mstuda f yo the ground S. of Westruther. A great deal of this is flattish, lying just below the 700- ft. contour; it has evidently been reclaimed by the canalisation and straightening of the small streams that drain north-eastwards past Westruther Mains, while Jordon- law Moss, whic feehw tfe lie highera s , still remains uncultivable undrainen A . d swamp in this area, closely adjacent to the supposed Spottiswoode forest, would have obviated any need for artificial frontier-works, and in fact might well have made their construction impossible; and at the same time, from a point NW. of Westruther MainBlackaddee th o st r Wate t 661501a r naturae th , l hollow that carriee th f sof drainage of this ground is a sufficient landmark in itself. On this showing Spottis- woode's story might claim some credit as reflecting an authentic tradition of high antiquity. Finally, if anyone wished to out-Spottiswoode Spottiswoode himself, he would be fre poino t et tha cross-dykee ou tth t s mentioned above combination i , n with their associated mosse burnsd san , coul thoughe db extendins a f o t supposee gth d Catrail- like boundar Kelphop e heae th th f do o y t e Burn (fig. 2), some halsevea d f nmilean s NNW Haerfauldsf o . . Suc extensionn ha t existedi f i , , would naturall mattea e yb f ro great interest, as it would enlarge the supposed area of Anglian colonisation to include not only the region previously discussed, along the valley of the Blackadder Water, but also the whole of north-eastern Lauderdale to the main Lammermuir watershed. 1 For a description of these works see 'Note of Cross-dykes' at the end of this paper. I am much indebted . MacLarenA r M o t , M.A., F.S.A.SCOT. visiteo wh ,als o d behalf y othewh discoverem d mn o an , e thire th d th f do air-photographserien a n so othere Th mentionee . sar d briefl Berwickshire,n yi . 236 P.S.A.S.,d No an , LXXXHI (1948-9), 203. 2 With whic Spott. hcf smoder e Bur th . 231)f n(p o hal a n. mil a , d SpottiswoodefW ean name Th e . e'd Spot', on record in 1296 (Col. ofDocts., n (1272-1307), pp. 201, 213), is derived by G. F. Black from the lands of Spot Easn i t t Lothian (The Surnames of Scotland, s.v.). HERIOTS' DYKE 239 However, befor muco eto mere buils hth i en o texistenc thesf eo e cross-dykes shoult ,i d recallee b d that such uncommot workno e sar n elsewher Berwickshirn ei plene ar -d ean tiful in the Roxburghshire Cheviots, and that other, undiscovered, examples might accordingly exist hereabouts in positions of no significance. Cases in point can be seen, for example, in two works which link the headwaters of the Soonhope and Kelphope Burns, respectively, with those of the Hopes Water1; these works, though moro n e tha nmila e fro northernmose mth t membe seriee undeth w f sno o r r dis- cussion, would not find a place in the suggested Catrail-like boundary as the Hopes Water flows in an entirely different direction, down the northern slope of the Lam- mermuirs and into the Tyne.

Note Cross-dykeson foue th firse rf cross-dykeo Th t s mentioned above runs fro mmosa s communicat- ing with a headwater tributary of the Wheel Burn (c. 555523) to a point about 50 yds. below the lip of the steep eastern side of the Earnscleugh valley and 1200 yds. NE. of Earnscleugh itself (c. 553526). Its length is thus about 400 yds. Towards its south- eastern end it changes direction noticeably; the angle occurs at what is now a small quarry, but originally this was no doubt an outcropping rock such as might have con- stituted a convenient landmark. The material of the bank has been obtained from quarry-pits whicf o l hal t havno , e been joine intp du oregulaa r ditch. This dyke, with its connecting moss and streamlet, completes the enclosure of a large block of land otherwise defined by the Earnscleugh, Leader, Boondsleigh and Blythe Waters and the Wheel Burn. The second dyke crosses a neck from a moss (c. 541536) at the head of the Trow Burn, a tributary of the Earnscleugh Water, to the head of the Gladescleugh Burn (c. 537538), whic turn hi n runs dow nsteea p slop joi o Whalplae t nth w Burn some 1500 yds. approximately ENE Longcroftf .o . This wor abous ki yds0 50 t. longd an , similar in construction to the first; in its case the circuit is completed by the Cleek- himin Burn and the Leader and Earnscleugh Waters. e thirTh d dyke crosse e tonguth s f lano e d lying betwee Whalplae nth d wan Soonhope Burns, which coalesc t Longcrofthea e a smal a restth d f do n en ls o e on ; stream which enter Whalplae sth w Burn just below Glenburni 542560). e e(c th d an , other on that of a nameless tributary of the Soonhope Burn (c. 536560), its length being about 550 yds. and its course slightly sinuous. This work consists of a bank of eart smald han l stones taken fro mquarry-ditca northers it n ho n side bane th ; s ki sprea breadta stando d t . abov1 o t 8an in abov. . f 1p hgroun2e ft o ft s u e3 th ed dan the bottom of the ditch, which is 6 ft. wide. fourte Th h dyke . frorun mmosa sNW 531587. Bellhile e (c heasth th f dt o a )l Burn, a tributary of the Soonhope Burn, to the steep south-eastern side of Crib Cleugh whicf o t streaha ou , m flows westward joi o Kelphope st nth e Bur poina t na t 1100 yds. N. of Kelphope. Its length is about 750 yds., and it is similar in dimensions constructiond an lase ton mentionede th o st t enclosei ; northere cona sth f o - d nen

'P.S.A.S., LXXXIII (1948-9). f 2 20 , 0 24 PROCEEDING SOCIETYE TH F O S , ig6l-6a siderable area of land which is otherwise denned by the Leader Water and the Kelphope, Soonhop Cleekhimid ean n Burns.

Ac KNO WLEDGMENTS In addition to acknowledgments made in the course of the paper and notes, I . DobiewisR . thano hM t , r C.B.E.kM , B.A. helr fo field-wor, pn . i G . I alsd r kan oM Scott, D.A.(EDIN.), who drew the plans.