FORTS, ' CAMPS,' ETC., OF PERTH, FORl'AK, AND KINCARDINE. 43

I.

THE FORTS, ' CAMPS,' AND OTHER FIELD-WORKS OF PERTH, FORFAR, AND KINCARDINE. BY D. CHRISTISON, M.D., SECRETARY.

Havin t variouga s times visite d madan d e rough plan f nearlo s l yal the forts in this large district, and being privileged to use the plans and descriptions of the few that I have not seen, but which have already been publishe r Proceedingsou n i dr Alex M y . b Hutcheson, thum a s I 1 able givo t etolerabla y exhaustive accoun e wholeth f to . The course of my investigations naturally led me to see also a good many of the ' Eomaii Camps' of the district, and of the obscure field- works, the precise nature of which cannot be ascertained without ex- cavating them—i lattere fth then f I ,hav O . e attempte givo t d e some account I hav d ean , referred briefl e samothero yt th f e o skin d thaI t have not seen, classing them all in a group by themselves ; but I have not described the 'Roman Camps,' because the subject is so large as to require separate treatment. (Platp Th ma e) showeI. e generath s l e elevatio th e lan y th b df o n contou d 100an 0 0 r fee50 line f te o principasaboveth e d seath "an , l • streams o avoiT . d overcrowding with names, the only inhabited places entered are the larger towns and such of the villages, generally of great antiquity, as are near the forts and serve as guides to their position. Besides the forts and the obscure field-works noticed in the text, the ' Koman Camps,' although undescribed for the reason just given, have been introduced on the map to give some additional value to it as a record of the fortified works of the district. All the works either have their special designation attached, when they hav e name ar one r o d, afte hil e whicn th ro l h they e nearesstandth r o , t inhabited placee Th . different classes are distinguished by the marks explained on the map,

1 "Notes on the Stone Circle near Kenmore, and of some Hill Forts in the neigh- bourhood of Aberfeldy, Perthshire," by Alexander Hutcheson, F.S.A., Architect, Broughty Ferry, . Proc.Scot.,A . S xxiii. 356. 4 4 PROCEEDING E SOCIETYTH F O S , DECEMBER 11, 1899.

and as the objects in each class are taken in the text from the south and west northwar d eastwardan d ,e p shoul th thei ma e b dre th plac n o e more easily found. Certain obscure works in Glenlyon, which lie beyond the map, are given from the Ordnance Survey on a separate little chart (fig. 55). My plane Figurese orienteth ar sf d o an ,p d to wit e e nortth th h o ht e inchth o ,t unles. ft 0 s e 12 scalwhe th f o en n o thee otherwisar y e stated. The profiles are usually on twice that scale. All heights of ramparts, etc., are perpendicular heights. A few contractions of words that occur frequently arc used in the text. Most of these are easily enough understood, and all that seem to require explanatio O.Me nOrdnancr ar fo . ;p O.S.AMa eN.S.A d an . r fo . the Old and New Statistical Accoimts of Scotland.

GENEKAL DESCRIPTION OF THK DISTRICT.

e districTh t dealt with nearly corresponds wit e Soutth h h Pictland of Skene, including the vaguely defined Fortrenn, but without Fife, whic s reluctantli h y excluded I hav s a personao , n e l knowledgs it f eo forts. districa Is ti t well define e isolatio naturth y b dy b wels a es n a l s fortoit f s from othe rboundes ease i groups th Norte t i t n th O y hdb . e northth n Seaa wid,O . e elevated tract destitut f forto e s dividet i s from the Aberdeenshire group. On the west, the Grampians, also entirely without forts, limit the united groups of Angus and the Perthshirn i Mearns f i d e fortan ,Highlane ecreeth o th d s p pu d glent si is only to find themselves cut off by a vast space from groups further west. Finally, on the south, a tongue of high tableland projecting from the Highlands toward Ochile sth s opposite Gleneagles Ochile th d s an ,runnin g thence to the Firth of Tay, form a well marked boundary. Besides town f importanco s e presenth t a e t somn dayi r o e, instances long districbefore th , eit t contains Abernethy, Forteviot Sconed an ,e th , now decayed capitals of the Picts, besides Dundurn, the probable chief stronghold of Fortrenn, also many villages whose great antiquity is testified FORTS, ' GAMPS,' ETC., OF PERTH, FOKFAR, AND KINCARDINE. 45

by the carved stones still existing at them, and although the number of fort comparativels i s y small, therotheo n s ei r distric n Scotlani t d that contain greao s proportioa t f larg no importand ean t examples. The country people, particularly of Angus and Mearns, although allocated in our military system to Highland regiments, appeared to me to be of a marked Lowland type in character, manners, and appearance; and the very small proportion of Highland names on the tombstones in the churchyards, even those on the Highland border, indicates a small admixtur bloof eo d wit neighbourine hth g Celts I hav. e been furnished frieny m y d b Dr Beddoe wit followine hth g not thin eo s subject:—

"Surnames in Laureneelcirlc Kirkyard.—Every tombstone was counted separately, but four or five identical surnames on one tombstone were reckoned as only one. r centPe . Surnames: Highland,... 20 8'5 ,, Border or Lothian, . 10 4'2 Doubtful, ... 14 6- „ . Others.19 . , 2 81'3 Some of those stated to be of doubtful origin may have been really Highland ones' others e th ; ' wer variouf o e s Lowland Scotch typesr o , common to England and the Scotch Lowlands. The inference I draw is tha t beet no ther ns muceha h immigration fro Highlande mth s since this part of the country was Saxonised. e followinth d n illustrationI ad gy factma sI , extracted froe mth Parish Eegisters of Muthill in Perthshire :— Out of 200 entries of names in the years 1697 to 1700, 83, or 41 "5 per cent., were of Highland type. But in 200 names taken from the Register perioe th f so d between r cent.184 pe 1886d 6 5an ,1 r , o onl , y32 were distinctly Highland thin I s. cas e changeth f populatioeo n seems to have been due to a current setting in from the Lowlands, or perhaps partly from the Highland families moving into the large towns." The distinction of race is also strongly brought out by the difference 6 4 PROCEEDING SOCIETYE TH F O S , DECEMBER 11, 1899.

coloue th f haieyesn d o i r an ascertaines r a , r BeddoD y db n observa i e - tions made in 1898, but not yet published.

CLASSIFICATION.

strictlA y scientifi r accurato c e classificatio f objecto n s that have suffered so much from gradual decay, and so much more from the ruthless han whicd f man an do e ofte ,h ar overgrowo s n n with d turan f weeds that without excavation we cannot eve e surb n e whether they are of earth or stone, is obviously impossible. The best I can make of it dividio st e objecteth s under the following chief heads :—

I. and probable Earthworks. II. Stone Forts and probable Stone Forts. III. Sites of Forts with little or 110 remains. . DubiouIV s work sitesr so , marked For O.M e - Camr o th t. n po V. Dubious works or sites, possibly military, not marked Fort or O.Me Camth . n po

I. EARTHWORK PROBABLD SAN E EARTHWORKS.

(a) EARTHWORK RESEMBLING A TYPICAL MOTE. 1. If division of our subject be difficult, subdivision is still more so, but it may be said that only one work can be structurally classed without doubt as a Mote. This is the Cairn Beth of the O.M., the Cairn Beddie, Cuer Bed, Beth r Macbeth'so e N.S.ACastle th e tru f Th o e . local name appears havo t e been Cair Betr n o probabl e CaeBeddiee d har th rBe d y,an interpreta- tion Macbeth'o t lea o st p u d s titla , e whic hI canno t earlietracn a o et r source than the N.S.A. The site is 5 m.-N.E. of Perth, 700 yds. N.W. of St Martin's Church at the bottom of a gentle hollow, close to a small rill, ft0 . wore 25 abov d Th mucs e kseai an th e. h ploughed down t stilbu ,l shows the plan (fig. 1) of a typical mote with a squarish base court surrounded by a trench, the mote or mound descending on one side into the trench. The mote is now only 8 or 10 ft. high and the trench almost filled up, but the writer in the N.S.A. says that twenty-four years before he wrote, a great quantity of earth was removed eved an , n afte s tim e hi O.Mr eth . represent wels a t sli preservede th , niote having a flat top 50 ft. in diameter, the base court measuring 230 by 200 ft. inside, and the trench 40 to 45 ft. in width. FORTS ' CAMPS,, ' ETC. PERTHF O , , FOKFAB KINCARDINED AN , 7 4 .

) EARTHWORK(5 S WITH SOME STRUCTURAL RESEMBLANC MILITARO ET Y MOTES, OR WITH MOTE on MARKED AT THEM ON THIS O.M. I have placed togethe l thal re works wit e terhth m Moat Moter o attacheo dt them on the O.M. Unfortunately the term is there used in various senses, but by favou Colonef o r l D. Johnston.A , K.E., Directo Surveye th enablef m o ra ,I d to state in the following paragraphs in which cases it designates a mound, and in which merely a trench connected with a fort. The information, however, cam lato enablo to eo recaset t e m te this unsatisfactory subdivisione th n I . present revisio mapse th distinctio e f nth ,o n between Mote, militarya , generally circular eminence, and Moat, a trench, is to be observed, and neither term is to b eordinarn a use r dfo y mound questionabls i t I . e e whethe casth y n an ei n i r distric e terth tm s traditionai d localan lwhether o f ,thes o y e an workr s have been military motes. s structurMosa r f themo fa t s a e, goes, have nearln ya equal claim to the title of mote or fort. 2. Inclibraltie.— Ona leve l field e f 2|groundCrieffo -th . inn E i f ., o s Abercairney . abovft e sea0 th es thii ,15 , s puzzling structure. Moat marked on it on the O.M., as shown in fig. 2, signifies the surrounding trench, which with its appurtenances is so drawn there as to be not quite intelligible ; I have therefore, while adoptin for e dimensiond gth m an Ordnance th f o s e Plan, represented the enclosing trench, etc., so as to correspond with my section A B, which is given on a larger scale. n for I e wor longa mth s ki , pretty regular oval, rathe. r E broade e th t a r e Th tha W. dimensione e ft nth .th end0 d 35 an , y e nearlb s ar ove0 l y90 al r structure rises gradually int oa much-flattene d dome t morno , e. ft tha 5 2 n above the level of the surrounding park, but it is pretty steeply scarped for a perpendicula trenche . dowth ft 8 o r nt ,o heigh t whic 6 f onlo . s thwidi ft y3 e and is bounded outside by a mound 18 ft. across and 3 to 4 ft. high above t trencr ou fa g s ha botfielde du trenc e th a thao h th e s trenc ,d t th th an no s hi as appearances go. An entrance, much modernised, 30 ft. wide, comes in from the N. near the W. end, and at that part there seems to have been some levellin interiore th f go . Her d therean esligha t indicatio parapea f no y ma t be seen at the top of the scarp. ' Castl e O th f fInchbrakiee o seene b traca o ,t t buneglectee s ei no ' th t d fine old trees and dense undergrowth make a proper examination difficult. On the whol t seemi e t unlikelno s y that havy thima es bee nMotea f unusuao l form sized whicn an o , hmediaevaa l castl afterwards ewa s built. 3. Moat, the Law, are the names given on the O.M. to a mound, IJr m. S.E. of Dunnichen Church and 500 yds. S. of Idvies House. It is situated 464 ft. above the sea, on the summit of a broad flat elevation or ridge, com- mandin n extensiva g e . above ft viewmoune Th 5 th e1 . do t rise 2 1 s vers i field d y conspicuouan , s wit stees hit p green sides crowned with treest I . has a slightly domed top only 18 ft. in diameter, green like the slopes, but with a small heap of stones on the top. The slope falls on a retaining stone wall 4 ft. high which girth e footh s t wit ha circumferenc . ft Thero n 0 s 28 i e f o e sign of a trench, and the term Moat (mote) was adopted on the Ordnance Plan for the very unsatisfactory reason that criminals were said to have been executed on it. The small rounded top seems ill adapted to the use of the work as a ydsMote.0 30 . "e N.E t i siti s th Idvief f eo o . s Church," wit remainso hn . Scale for all riiePlans 100—TO—o . Earthwork11 Figs o t 1 . Perthn si , Angus Mearnsd an , . FOliTS, 'CAMl'S,' ETO., 01-' PERTH, KOKFAli, AND KINCARDINE. 49

4. Mimt, Gallows e Laionexth ts i (tigexample, 3) . 3 111, . E.N.E. from Uie last, J in. N. of Gardync Castle, £ in. S.S.E. of Outline Church, and about 200 ft. above the sea. It is conspicuously placed at the very AY. end of a singular, narrow, artificial-looking but natural ridge which runs E. a fiel d Wn do an . sloping gentl yo t ITfro .. S m This mound rises gradually from the E., with a gently-rounded crest and steep sides, 15 to 25 . n HearinendW O , e . afteth N g e r th n o 5 4 o t 0 3 d an . S fte . th hig n ho

» iO Fig. 12. Castlehill, Inshewan.

a course of 70 yds., a trench cuts- off the mote, which rises 8 ft. above e trenchd fall 8 ft.th a fla1 an , s ty b steepltop,ha 7 2 , y aboue th fee0 2 o t t e N.. fac th s ,W quarriei eo upo t e e field0 Th nth 4 d . d awaan t . y(a S pla e sectiond th n an n show6d o )an s nothin t san rolled gbu dan d pebbles from bottomo t e namp Th to .e Gallows w seem t La inappropriate no s , froe mth following abbreviated referenc e O.Sth . n i eaccount artificiao : "Tw l conical mounds called laws exist at Idvie and Gardyne. An old man told Thomas VOL. XXXIV. D 0 5 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R .

Lyel that he saw two Highlanders taken with stolen cattle, judged, condemned, and hanged on the Law of Gardyne" (early in the eighteenth century), but this is no reason for calling it also a Moat (Mote). 5. 'Passing now to the high tableland upon which the high road from Forfar to Brechin runs, we come on the Mote of Melgund of the O.M., the only instance, withi r reviewnou , ofe spellin.th g mote placn i f moat.site o s ei Th If m. N.E. of Aberlemno Church on the broad summit of Angus Hill, 451 ft. above the sea. It is represented on the O.M. as a very small, shapeless s probablmoundwa d an ,y then undergoin ggraduaa l destruction, which seems to be now complete, as I could find no trace of it on the site in a field then under turnips. 6. Scarcel f thiso . E ,. yS.8.E J m . 2 m Jf Aldbar o . . m Castle,2 d an S.W f Brechino . , Moat s markei e O.Mgentla th n n o . o dy sloping fiel6 31 d ft. above the sea, at a place called Chapel. The term here means a trench, but it must have disappeared, and I could see nothing but a cottage and garden besid ea squar e enclosure f whica o s p hto wa mound w e , lo fenceth a n o ,y b d ruined wall, of which the N. side, 104 ft. long, and part of the E. and W. sides, 84 and 48 ft., remained. The site seems to have been of some import- ance formerly, as on the O.M., besides the names Chapel, and Priest Shed (apparentl name fielya th f edo clos o Moat),et yds0 ther 20 o t Court.a w s ei La the S. and Site of Church Bams 400 yds. S.W. I have retained this and the preceding No. 5 in this class, because they are marked ' Moat' on the O.M. But strictly they should have been relegated to the class with no existing remains. e las7Th t. occurrenc e worth f deo Moat (here signifying trenche th n o ) O.M. is at Castlelrill, Inshewan (fig. 12), which is situated on the N. bank of the South Esk, about 300 ft. above the sea, 2 m. "W.S.W. of Tannadice 0 ydsChurch50 . d W.N.Wan , f Inshewao . n House e riveTh r. here flows between perpendicula fortrese r th cliffs d formes an si , cuttiny db gdeea p trench landward, where there is a bend of the river, the land side of the long oval beinn i . gft protecte0 4 e e trenchrive r th th o y e clifrd 0 d b th 3 sid an , f y eb e foaminheightth d an g, river rushing s rockalon it belowd e n gi be yTh . trenc remarkabls hi y deesteepd d pan fro configuratioe an ,m th groune th f no d e counterscarth s highepi rft.5 1 tha ,e proportio e scaro th t nth n 5 pi 2 f no where highest, abou middlee th t . Eastward, however, fro e naturamth l faln i l the landward ground, the height of the diminishes rapidly as it curves toward e entrance e riverth sth d heres i an ,, C ,, stream e closth e o et Th . nearly leve prettd an l y regularly oval interior . measureft 0 10 y s b abou 0 15 t A sligh t landwarde bankinth f o p gu edge maybe remain ramparta f so . There are no signs of stone work and no stones lie about.

(c) EARTHWORKS WITH SOME RESEMBLANCE TO MILITARY MOTES, BUT NOT MARKED MOTE OR MOAT ON THE O.M.

8. The only earthwork resembling a mote in the Highlands of Perth is on the Torr Hill,% m. S.S.W. of Aber/eldij, on the left bank of Moness Burn, 400 ft. abov e seaeth , withi e publi0 ydsth n5 f o .c roa Crieffo dt r Hutcheson M . , from whose pape I rtak e plaeth n (fig. 13), callvera t i sy distinc welld an t - marked earthen, fort, measurin gft.4 ,internall 12 rectangula y b 4 y15 formn i r , I'ORTS ' CAMPS, / ETC. PERTHF O , , KORFAR KINCARDINED AN , 1 5 .

with two trenches on the slope, varying from 24 to 36 ft. in width, and rising 40 to 45 ft. above the surrounding ground. He also states that, on the ascend-

Fig. Earthwork,13 . ^ Ton- Hill, Aberfeldy Hutcheson.r (M . ) forte th ,f thero severa. e S ear in e gth hile l slopn otheth l o f eo r well-marked trenches, which may have been thrown up as to protect the fort on that the most vulnerable side. 2 5 PKOCEEMNC4 E SOCIETYTH F O S , DECEMBE , 189911 K .

This work seems to be altogether of rather a Unique kind. It is so in the first place because earthen fort exceedingle sar yScottise rarth n ei h Highlands; s alsi t oi remarkablt bu s rectangulait r fo e r form t I wil e observe. b l d from Mr Hutcheson's plan that, although the interior is rectangular with rounded angles, the trenches assume a more circular form. In these respects it resembles ' Roma e th n Post f Kaimso ' Castle between Ardoc d Strageathan ho n d an , other work with whicacquaintedm a I h papen O Torhile .th r l wor somes ki - what suggestive of a terraced mote, and what with that and its position so near lyinw lo t goa inhabited site t seemi , t unlikel e worsno a th e f b k o y thama t ti mote period. 9. Dundee Law.—This conspicuous green eminence rises on the N.W. side of towe th f Dunde. abovn o . ft bein ft 2 seae heigh a e lase0 th 57 go eth ,t 10 f t o t

,,.,...-i"..-;/,.>^ ' W////^> .>•£'- n -':W/u,. W///S/.

Fig. 14. Dundee Law.

very stee summi e l roundpal Th t particularlnearls i . tbu , N e y th leve n yo l e forth t d (figan . , 14)t witE. bu d gentl,ha measurin an . eS incline th g o et 260 by 170 ft. over all, occupies the whole of it. The work is peculiar for a hill site, being strictly rectilinear, excep roundea r fo t. endN d e anne, th t xa whic outsids hi maie eth n and"follow contoue hile sth therep th lto f o r . The inne northere rth ared aan n ; measure part. ft a , 0 r A8 fo ,sy b abou 0 14 t breadth of 36 ft., is very slightly raised above the southern part, B, and has a FORTS, 'CAMPS,' ETC., OF PERTH, FORFAR, AND KINCARDINE. 53

comparatively rough surface. The whole area is enclosed by a rectilinear . wideft 0 1 , . acroswhico ft t d wit6 0 3 an sh moun p ha flao riset to t 4 s2 d only a foot or two above the area, and 3 to 6 above the outside. On the N. side this rampart gives on the little level space, C, 33 ft. wide, with a rounded head, spoke f aboveo n , which shows slight remain a ston f o se wall t necesno , - sarily ancient ,e roundeath t d e edgeotheTh r. three sides hav seconea d dan lesser mound a little in front of the inner one at the edge of the descent. In connection with these chief mound s prolongei e rampartsr o son . N d e eastth , - ward, partly naturally perhaps hile dow d e edg th hillan e th f Dn o th o et l o , t , e approacth r ba ho t whic s sa o h e easterskirtth p e u s nhil th e ilano th lt f ko rounded head. This limb, D, also flanks a possible zig-zag entrance, E, F. The S. rampart is similarly prolonged, flanking a direct entrance, F, from the east, and it also gives oil' short branches at either end towards the outer ram- parts. The work seems to be of earth ; at least I saw few little stones and no oneg bi s lying about excepe roundeth t a t d head, which lies beyon e maith d n worke rectangulaTh . re wor th for f kmo give t i ssom e claia Roma o mt n origin t provethero n bu , s ei d Eoman fort perche similaa n do r sit Scotlandn ei . The rectangular form, and the greater elevation, slight though it be, of one cud, alse ar o suggestivt characteristicno e motear a f d eo hila an , f lso fort posie th ; - tion, on the other hand, is much more that of the hill forts than of the motes. mediaevae b y Thama s alst i i lt o possible, although histor s silenyi t abou. it t Altogether this is an anomalous and puzzling work. 10. Oastleton.—8J in. along the high road from Forfar to Perth, and 3 m. E.N.E . f abovMeigleft o . e sea0 a th nearle 18 s ,i y square fortress, raised considerably abov e surroundinth e g countr n threo y e sidest bu , less so on the S.E., where the entrance is directly from the high road by an ascending broad ramp, either moder modernisenor approacthe as dhousa hto e and garden in the interior (fig. 4, which omits the western part, broken down house garden)d th site fof an eth ro wide Th e . N.E. e trencth 1 a sidh01 s eha scarp, steep and straight, 12 to 14 ft. high, crowned by an earthen parapet 3 ft. high to the interior. The counterscarp is only 4 to 6 ft. high, so that the interio greaa s ha tr command e fortificatioTh .e S.E th n . o nsid mucs ei e th h same, but the trench is not quite so deep. The S.W. side now shows only a rather easy slope, but the construction of the house and garden no doubt caused great alterations and the destruction of the defences there. On the N".W. the steep scarp again appears, but without the trench. A small burn runs at the e S.Wth foo f o t. sidd ther an e sign e ar f anothe o s r little watercourse th n eo N.W .dimension e side0 Th .e 20 interio e O.Mth y th aboub e f o sn ar 0 . o r 30 t widtramparthe the hof trenca ft.and and t in h ft. varie 80 s to fro m35 straight intelligenlinen A . traditioo tol n there n ds m ma t mediaevaa ewa f no l castle, but he spoke of the trench as 'the moat,' and this may perhaps have

been the traditional name of the whole fortress. It must also be the work said JameDr sby Playfair havto 1 e been formeEnglisthe by dthhin e reignof Edwar s positiondit I.s a , , dimensions d structuran , e correspond exactly with his description e givet goei H authorito s .n a s s r statementhi fa r s a yfo t bu , it confirms the mote theory. If not a mote it must be regarded as a Roman work rather than a native fort, although it has marvellously escaped being

dubbed ' Roman Camp.' Picture1 of Scotland, James Playfair, D.D. . 433i , . 4 5 PROCEEDING THF O SE SOCIETY, DECEMBE , 189911 R .

. If11 m. S.E f Glami.o s Church . , N.Wm -j f Arniefoulo . hamletn i , Hayston wood (now muc t down)cu h . above ft , e th seafull0 th n e70 o y, top of a lofty ridge commanding a fine view, is this circular little work (fig, consistin5) . a flat-toppe f o g d moun t . morhighft no d 6 e, r thao 5 n diameten i e top. 50ft th ,n o rslopin g gentl a sligh o yt ty rampartma t i f i , be so-called, 12 ft. broad and only a foo't or two in height, which environs it l roundal e over-alth , l diameter bein g. ft Manabou 0 y12 t rounded pebbles of considerable size show in chance breaks of the mound, but no stones suitable for building e groun a greaTh r . dfo t distance round abou bees ha tn surface- quarried for slate. 12. At Canterland, J m. N.N.W. of the farm of that name, lj m. E. of Marykirk, 4£ m. N. of Montrose, 406 ft. above the sea, on the level summit of a ridge at the very S.W. end of Garvock Hill, with a commanding view to the S. and W., stands a work (ftg. 6) closely resembling the last, being nearly circular d havinan , ga centra l n i moun . . higt ft ft onl bu h0 6 d 2 y diamete e flath t n topo r , with sides sloping gentln encirclina o yt g mound . highft 2 ,t whicno h differs from tha t Arniefoua t n beini l g double with some compartments, perhaps changeo diggingt e e centradu Th s . l moun alss doha been muc intog hdu , revealin same gth e compositio t Arniefoula s na .

(d) APPARENT EARTHWORKS THAT ARE PROBABL MOTEST Y FORTNO D . SAN 13. Beginning from the south and west, as under the previous head, the first in this class is the Fort of the O.M. and Post of Roy on the Grinnan Hill of Keir.—The site is less than f m. S.W. of Ardoch camp, 420 ft. above the sea, on the edge of a steep descent, 40 to 50 ft. high, to Keir Burn, but only slightly elevated above the field towards Braco village. It has apparently been n earthwora k wit ha semiova l triple lin f defenco e e (fig. 15), partly ram- parte trenchedd dan , partly terraced broae th , d oval being rudely completey b d unfortifiee th d e steeedgth f peo bankentrancee Th .alons i , e narroa ,g th w crest of a ridge, i, from the E., and it is likewise approached by a rude road- way, c, from the burnside below. Roy's plan make e worsth k nearly complete middle th t linee ebu ,th halsf o f no longer exists. He says that it may have been a work of the natives before the arrival of the Romans, but calls it a (Roman ?) ' post.' There can be 110 doubt that it belongs to a common type of native fortresses. Its extreme length interioe isth abou havy d ft.0 an rma e, 32 t bee n170 y aboub .0 t20 14.^11 m. E.N.E. of Blackford Church, 180 yds. N.W. of the cross-roads at Loaninghead, opposite the mouth of Gleneagles, the easiest pass through the Ocliils from Perth to Fife, is another Fort (fig. 16), 5j in. nearly due E. of the last, which it closely resembles, and marked ' Roman Camp ' on the O.M. The site, |- m. W.S.W. of CriefF Junction, occupies the whole of an oval eminence, 520 ft. above the sea, rising about 25 ft. above a field on the N.E. side, and only 8 or 10 above one on the S.W. side. This eminence forms the central part of a little ridge, but is approached from it at each end by narrow necks, B and D. The entrance, I K, is not from either neck, but from the field to the S.E. The , consis D a steeC f o d tp defence an scare endsB th pt A , a sfro interiore mth , falling o trenchna , beyond whic a rampar s hi second an t d trench scare Th p. and rampar somn i e ear tplace . abovft botto e trenche9 e sth eth f mo thein si r El CO

IT!

t> aS3 w

fz! M

Fig. 15. Earthwork, Braco. Fig. 16. Earthwork, Loaninghead, Crieff Junction. Ob PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

front. On the N.E. face, E, F, the trenches pass into terraces which have no parapets. On the S.W. face, where they are most required, the defences are entirely dimensiongonee Th . thif . oveso ft s 0 rova20 l y forb e abou ar t0 38 t allinterioe th , r bein g120.y aboub 0 1 23 t steeedge a th f eo pn O woode . 15 d bank. highft 0 0 , which a haugfall n so h enclosed to the S.E. and W. by a, loop of the River Almond, 2 m. from its mouth, 10 . abov0ft seae eth , at Pitcairwjnen, this i s semicircular work (fig, 7) . the base being formed by the straight and unfortified edge of the bank. To the e semicirculaN"th . r lines face nearly level ploughed land havan d e beea n good deal destroyed. Where most perfect, in a plantation at the W. end, they consis a scar. high ft f o . tp6 hig ft , hwit3 moundo beyondd htw an 5 sn o , a pretty steep slope ; but in the centre towards the N. there is a terrace 10 ft. wide in rear of the outer mound. The inner scarp is parapetted only at the entrance, whic ft.e lengts fro 0 e basi hN.We , th Th 27 mth e f s h.i o ove l al r and the width of the interior, which contains a rectangular foundation, from N. . 135toS . Apparentl pura s i e t y i earthwork . 16. The next on the list, within the grounds of Cultoquey, close to the high . road N.E f m f GilmertoOrietfo o 2 .. , E ,. 1m - n Church. abovft e 0 th e37 , seas i marke, ' Eomad n e OutpostO.M. th t ther bu n n hardl,o ca e1 y ba doube t tha t belongi t a commo o t s n clas f nativo s e works n whici , h the extremity of a ridge or tongue of laud, whether on the coast or inland, is converted int fortresoa s simpl drawiny yb intrenchmenn ga f o t p acrosto e sth the ridge, the elsewhere showing in general no , and having been probably defended by a wherever the natural strength was insufficient. In this instanc e ridgeth e , . highabouft 0 ,3 t rises steeple th n yo Nascene .th sidegenerallt s i t bu , y gentlee intreuchmenTh . r S fro e mth t runs nearls i e ridge th d yf an o ,straigh p . acrosto ft r abou e 0 fo t th s12 t carried partially down its slopes. It consists of two ramparts or mounds and two trenches covering a width of about 60 ft., and is peculiar in this respect, that it is apparently designed to defend, not the small extremity of the ridge to the E., bnmaie th t n ridge. TheydsW 0 lengthn i e 20 ,". th easterl o t , y mound in particular is massive, and at one point rises 9 ft. above the trench to the E. The straight course of the intrenchment is probably the only reason why it has been ascribe Romanse th o dt . 17. Passing to the north and east side of the Tay we encounter two earthwork e Sidlath n wo s t EvelicJc,Hillsa e firs s i Th tyds0 . 50 . W.S.Wf o . the ruined castle of that name, and 1J m. W. of Kilspindie Church. It is . abovft reare e ol0 th de20 d castl . abovft e e e on 0 seaanth e89 n d o , of several little rather flat-topped eminences which fore tablelanth m d summi f Polo t e Hill, the highes f whico e greet Th hn . attainheighft 0 94 st on which the fort stands is conspicuous from below and looks down upon the castle, but the fort is invisible till the hill is climbed, as the works face N.E. and tablelande N.Wth o t . , whence the easile yar y accessible stronn I . g contraso t t the accessibility on that side, the ground falls directly from the fort with extreme steepness 200 ft. to the S.E., 500 ft. to the S., to the chasm of

1 By an unfortunate error in the iigures of the scale, fig. 30, p. 95, of my work oil Early n Scotland,i thie precedinth s d foran t g one wels a ,e als a lth l othersaie th d n figuresi givee ar , n only half their proper size. FORTS ' GAMPS,, ' ETC. PERTHF O , , FORFAB KINCARDINED AN , 7 5 .

Pitrodi e S.Wth e o .Burnt form d an , a sdee steed pan hillpe th clef , n liki t ea natural trench. As the position is a characteristic one I have given a general plan of it (fig. 17), while the fortifications are shown on the usual scale in fig. 18. A slight mound at the edge of the steep descents is sufficient protection for the sides tha se o ar twel l defende natury b d t etoward , wher;bu E. groune e eth sth d weakests i a scar,. hig ft 7 hp terracea fall n o sd thaan , t again upo na secon d

I \ ^ ^^^^ Fig. 17. Position ot'Evelick Fort on Pole Hill.

terrace, with a wide trench and two slight ramparts or mounds beyond. To the W. there is.a much longer . highscarft 0 singla 1 ,p e terraca . wid d ft 9 ean e trench and single mound beyond. The change from one system to the other takes place at what seems to have been the chief entrance, which passes obliquely through the lines, but there was probably another where the lines end eastward steee th t p broaa e faceTh d. pear-shaped interior measure ay nb averag0 s35 e of 200 ft. on the O.M. and the dimensions over all 460 by 280. secone Sidlae 18th Th .e littl f do wth f eHilo d l eartheen . E e nth t forta s i s 58 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R .

hamle f Bait,o t f m. N.E f Kilspindio . e Church . abov,ft e abou0 th e 10 t e verth e yBraet th a f foof o Cars o sd f tseaGowrie) an (o e, . This narrowA e . fortressamth 16 ,f . e o steep-side typsNo s i (figs ea ) 8 . d spif o t land, a sudden contraction of a gently rounded little ridge that runs down to the Carse threy fro b e stee mf th e of pslightl t braescu s i y, curve d massivan d e

«

'•'..very steep descent, c

Fig. 18. Earthwork, Evelick.

intervenino . tw abovft . e wid ft 8 d th risine 0 o ean t 4 mounds g g7 o t 0 3 , trenches ramp-entrancA . e crosse linee sth s obliquely lengte cute th Th - f .h o ofabrup n t fa i end spiy d b s abou0 ft.i t t 3 an ft.0 ,s heigh it o ,t 50 t 5 2 t descene highwa th e ver o th t t y t e Carse ya edgth s naturall It f o e. y scanty widt s beehha n much narrowe e quarryinth y db f sangraved o g an d l from 1'OETS, 'CAMPS,' ETC. PEKTHF ,O , FORFAK D KINCARDINE,AN 9 5 .

its western face, so that in some places the top is now a mere crest a few feet wide. 19. Passing into the county of Forfar, we do not meet with any unequivocal earthen fort in the interior, but on the coast there are three which are little vock fastnesses, so strong by nature as to require but little aid from art. They are all of the ' cut-off promontory' type. The first is Maiden Castle (fig. 9 and sketch, fig. 19), If in. E.N.E. of Arbroath at the S. end of Carling- heugh Bay flat-toppedA . , narrow point . abouhigft s protecte0 hi 8 t y a,b d mural precipic l roundal e , except wher t i approachee. N e th e lan n th so d

Fig. 19. Maiden Castle, near Arbroath. (wher e precipiceth replaces ei vera y yb d steep grass lane th y d n slope)o d an , side itself, from which it has been cut off by a deep trench, above which towers loftya , massive rampar f eartho t , rising abou. abovft 4 e trenc1 1 e2 tth d han abov interiore th e . This mound e shows sketcha ,th n i n , fig, take19 . n from the S., is very conspicuous on a coast where the land seems as if it had been close-shaven by the wind. 20. e N.E.1£th mo t ., further alon e coast f th Auchgd o f an . ,m S .- mithie, is Lud Castle (fig. 10 and sketch, fig. 20), forming the S. point of Castlesea . highft Bay 0 , ,10 girt witd precipicesre h , except towards the land, where it has a broad, straight grassy front, which, falling steeply 40 or 50 ft., contract narroa o st w neck mainland e joininth o t t gi . This nec pres ki - cipitous on the N. side, and slopes steeply to the rocky sea, so that the fore acces s difficulth i t o evet d s an nt dangerous t contenNo . t with thise th , defenders have reared a massive mound, 5 ft. in height above the interior, from precipice to precipice at the top of the slope. The present path skirts obliquely e rampartth f o d . en Probabl . S e e slope originath th yth o et p s thi u wa sl access thers a , some ear e signpassaga f so t througecu e ramparhth t e hereTh . interio t firsa f fuls ro i t l breadth irampartne th rea f o rt quickl bu , y contracts to a narrow passage leading to a small level square at the far end, the highest 60 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

point. The sketch (fig. 20) shows the precipitous N. side of the neck, and the perfectly inaccessible 1ST. side of the fortress.

Fig. 20. Lud Castle, Auclmdthie.

. Facin 21 e picturesquth g t decayinbu e g fishing villag f Auchmithio e e (the Musselcraig of The Antiquary), from a distance of a few hundred yards, is Castle Jloclt, (fig. 11), a level-topped, square projection, measuring about 100 ft. each side, and about 100 ft. high. Three sides are of perpendicular rock, and the fourth is cut off from the mainland by a curved double mound with two trenches. The profile of these is very slight, but they stretch from edge to e precipicedgth f eo d theian e r combined widt ft.0 thao 6 s , s hti e b ther n eca little doubt tha place bees th t e ha fortressn a . interioe th Mearne 22th n I f . o r s thercertaio n s ei n exampl coasa f eo t rock- fortress, although Bowduns, near Stonehaven, may lay some claim to the title, and in the interior there are but two forts of the earthen type, one of which barelw isno y recognisable. Thi si Dikes, O.M.e th f o , Arbuthf ^o m. E . - t Church . abovno ft fla0 seae a eth n t15 , , o promontor angle th junctiof en o yi n e steeoth f p little Chape witn Bervie De lhth e Water, raised abou ft0 .5 t above the latter, with pretty stee distincA p . slopee E.S th , td N.o t senough an , , wide and straight undulation of the ground, 350 ft. long, crossing the flat and cutting promontoryoile th ' t correspondseeni e b t l than ,bu al ca s t,i position si n wit ramparte hth . wide ft d treiic 0 Ws an 3 ,it . n sidho e marke O.Me th n ; .o d and reasoning from analogy, there can be little doubt that this was a 'cut-oft'' fort with an earthen rampart now nearly ploughed down, the interior space . ft bein0 25 gy aboub 0 35 t 23. The other Mearns fort of the earthen class is 2 m. N.N.E. of Fettercairn Church . N.AV, m f f Phesdo . o House d abouan , a milt e froe sitf th mo e ancient Kincardin e remarkabl th remaine e castlee s th towi th d t I f nan so ; y strong Green Castle (fig. 21) of the O.M. (although I could not find that it was known by that name locally). The site is a strong one, 400 ft. above the sea, FORTS, ' CAMPS,' ETC., OF PERTH, FOUFAR, AND KINCARDINE. 61

on a high point of land overlooking the Fevduii Water, and it is fortified by a single rampart of earth 8 or 9 ft. broad on the top, rising as much above the interior fallind . an i, nft gperpendicula0 3 steepl o t 0 y2 r height upo na wid e trench with a counterscarp 3 to 8 ft. high. The height of the rampart above

0 j(SP) 100!'' Fig. Gree21 . n Castle, near Kincardine Castle, Fettercairn.

the interior is a peculiar character, and shows that it must have been defended from the broad top, probably from behind a palisade. The dimensions over e th t a 0 20 e broa th d t an da 0 30 e O.M th y b n .. o pla ft aboue l 0 nar al 40 t narrow end of the pear-shaped structure. The interior within the rampart is 22 12y 0b 0 and 5. 0ft 62 PEOCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

II. STONE FORTS AND PROBABLE STONE FORTS. The Stone Forts of the district number twenty-two ; not many for so large an area t severabu , f theo lamon e mar e largesgth d mosan t t interestin Scotn gi - land. They are so distributed as to be conveniently divisible into groups, situate certain di n mountainou r hillso y districts. ) STON(a EHIGHLAND E FOETTH N SI r So PERTH . The stone forts in the Highlands of Perth are seven in number, of which six e county wese th n f th i e solitar e o Taytbasie r n i th o Th ,th f e .n n o i y on e ar e Forthe basi th probabls th f no wa , y outsid e Pictiseth h kingdoms i t i t bu , noticed here t completei s ,a lise Perthshirf sth o t e forts. 24. Dunnwre milesLedio n tw Be a ,n ' o wal k from , 1100 yards W. by S. of Bochastle Farm, and 350 N. of Tarandoun, perched upon a prominent knol l. abov abouft hige 0 eth h40 t roaLocd dan h Venachar abov0 e 60 eth d ,an sea, is a fine example of the type in which a semioval front of fortification defends the accessible side of a fort, the other side or base resting on the edge of a precipitous bank (fig. 22). In this case the bank, which is on the east side, is about 150 ft. high, and is broken by rocky faces, and, although it can be climbed, is impracticable to an attacking force. To the west and north the ground falls pretty steeply from the interior for about a height of 25 ft., below which there is on the north a small enclosed but not fortified plateau, and on the west a quite gentle descent. On the steep part four parallel abrupt green

mounds circle round, but large stones are everywhere embedded in them, and doubto n e thers b Mis a , n seca Maclagan 1 points out, that the realle yar y stone walls overgrown with turf. These walls being builsteea n o tp slope were much higher on the outer than the inner side, and must have had much the character of revetements with ston e inneth e f w topro walr parapetsno p so to s i l e Th . only faintly indicated e heighe moundth Th .f o . t ft s 0 varie1 o t s7 fror o m6 on an average, but in one place is as much as 13 ft. The second and the third mounds unit s theea y e northcirclth d thero et frone ,an f fortificatioeth o t n becomes straight wale flane th lf weakes e Th o kherth . s ei tforte parth ,f o t as the ends of the lower walls are raised only a few feet above the moderately easy ascent fro east e theid mth an , r front rises fro mth l plateau. At the south end there is a smaller plateau before the entrance, which passes only throug firse hth t mound continues i d an footpat,a y db h which passes over rather than throug othee hth r mounds probabls i d an , y modern. . sideW approaches ,i s it entrance f o norte Th d th en o hdet . plateauS e th t a , somethiny b g lik a eroadwa y i ne fortifie th fron f o t d linese inneTh .r area measures about 180 ft. from N. to S. by 150 from E. to W., but a part of the latter measurement fee0 6 , t broadese width t ea t part rathea s ,i r awkward slope to the precipitous edge. A deep hollow in the interior, moist at the bottom, probably marks a partially filled in well or cistern. The ground plan is founded on that of the O.M., with which my tape- measurements of some of the chief dimensions substantially agreed. 1 Hill Forts and Stone Circles of Scotland, Miss Christina Maclagan, p. 53 ; and Proc. S. A. Scot., ix. 36. FORTS, 'CAMPS,' ETC., OF PERTH, FORFAR, AND KINCARDINE. 63

f a t Ji e r

70'...... 9o' .... 60'

Fig. Dunmore22 . , Bochastle Ledin Be , . ' 4 I'KOCEEDING6 E SOCIETYTH F O S , DECEMBE , 189911 U .

f^%r#3rw*%m FORTS, ' CAMPS,' ETC., OF PERTH, FORFAR, AND KINCARDINE. 65

EXPLANATION OF THE PLAN, FIR. 23. North Side of the Fort. A Covered way, ascending fro e plainmth . EartheX B mound) n(? . E First plateau, with levelled space, I). FirsG F t wall entirely thrown down reste , th lik l , ewital h entranc. F t ea AdvanceH d mound t foo a ,rockf o t y ridg whicn eo stood G hF . I Second plateau, a rough hollow. K s nort LIt h wall continued, rising abruptly fro thire mth Ldo t , plateau. N" Its south wall, near the foot of the precipice, cutting off a possible but difficult topascene th . o t t 0 Third plateau . abovft seconde 0 5 e,th . M Its west wall, the second line of defence. P End of its north wall at edge of southern precipice. Q Third wal defencef o l , running from south precipice steea edg p epu slope abovd an erocka y bluf cliffooa o t f o ft nea e topth r . . diameterft 0 7 o t ,0 LeveT defende6 p to l littly db e cliffs joine wallsy db . West Side of the fort. S First wall, about half-way up, drawn along edge of a steep rocky slope from f norto p f southo to p precipiceo ht to narroA . w passage s runit n i s rear, from which rise e rockth s y e seconascenth o t dt wall. Near the north end is an entrance, approached by a path up the hill from the plain. SeconE d wall edgn o ,stee f eo p rocky slope, drawn from precipic o precit e - pice, and very nea tope th r. 25. Dundurn.—Elsewhere1 I Lave given a full description of this interesting hill-fortress I shal d lan , quote here merel e chieyth f points, elucidating them viewo tw b y plaa s d takenan n from photograph name , 25)Th s e24 (figs. , 23 . e O.Mt marke th e hilno s DuniJiUan,i n .oth s o fli s havin dt a i d gforta an . But I found that the recognised name of the hill on the spot is Dunurn, and that it was occupied by an undoubted fort. It is therefore in all probability the Dumduirn, identified by Skene as the principal stronghold of Fortrenn, although lie doe t sees no havo mt e been aware tharemainay an tfortres a f so s still existe thn deo hill. That distinguished authority point t thasou t Dundurn is twice mentioned by the early annalists as the seat of historic events :—(1) A.D. 683, Obseeaio Duinatt et Dumduirn ; (2) A.D. 878-9, when mortuus est in Dundeorn recordes i f Kindo g GiriDd ran gSken; e holds tha e identitth t f yo place th proves linee i th Sf sto y db Bercha n (llth century)— shalm attackee hi b l y " powerfuB e dth l house. banke e Ear th Ahth hear y f n so m ,o ! t Ked shall be the colour of the house before him, Fortrenn.f o n shale me H e l"th faly b l The site is on an abrupt, craggy, isolated hill, rising almost to a point, and the walls of the fort have been built from crag to crag, or along the edges of precipices 1 Early Fortifications in.Scotland, 1898. 2 Celtic Scotland, i. ]>p. 264, 330. VOL. xxxiv. E 66 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R .

* Kfc- ' >.f •• rsiffirtr ii.1 i is FOETS, 'CAMPS,' ETC. F PERTH,O , FOBFAR KINCARDINED AN , 7 6 .

steed an p descents .nature ,groune justh th s f a teo d dictated hile lTh .rise a o st height of 500 ft. above the sea, and 200 above the haugh of the Barn, a mile below its exit from Lochearn, in the midst of one of the most charming scenes Scotlandn i fors It msomewhas .i t triangular (fig. 23) southere ,th n side rising by a nearly inaccessible precipitous face from the Alt logain, the western by a very abrupt ascent, broken by steeply sloping ice-smoothed rock-faces and little cliffs, while the northern, rising abruptly at first, eases off suddenly into three rough irregular plateaux , raise0 above , 1 on e dothe, E th e, r fro, m E. Wo t . and bounded on the S. by a mural precipice, X N, which rises to the top of hillthe , completel lowetwo y r cuttinplateauthe off g x frotopmthe ,but leaving a difficult access from the third along the edge of the southern precipice. e vieTh w froe N.Wth m . (fig ) show. 24 e defencee northerth s th n o s n side. Abov outegroue e eth lefth e trees f ri th pto moun n so plan)e th dn o . (BX Abov nexe eth t three mounde tree th advanc n i s si e firs, th H , - tf eo de wal f o l fence. The next two trees stand on the debris from the N. wall of the second plateau. Then come e continuatiosth f thi no e thir e slopsth th wald o t ep u l plateau. This plateau lieo hig to sshoo t h . walwN s see i s lmuch it n t bu , running abov steee eth p tree-covered e sky-lineslope greath th o et t d debrisan , . wall W e secon mass th , it f so d wal defencef o l s conspicuoui , s runnine th o gt rockfooa f o t , above whice thirth s hdi wal f defenceo l , wit a htred bus ean h on the sky-line. Highest of all is the debris from the wall round the summit. Strictly speaking, the building of the walls is only inferred, as none is visible, but there can hardly be a doubt that the masses of debris conceal the bases of regular walls. All these points can be made out in fig. 25, taken from the W. (some trees are omitted), the N. ends of the two walls of the western side being also visible. The fort is about 600 ft. in length by an average width of 450, without reckoning the first plateau, E, which perhaps was not intended to be seriously defended. The contained area, therefore, is very large for a Highland fort, but the habitable space is limited to the hollow of the second plateau, to the third plateau (much curtaile summite th rocks)y o dt b d , witan , h some tolerably level . S d an littls groun . it fooe E f eth o e t t clilfth da n so ' Ports o marke26e e O.MTw .ar 'th Glenalmonn n i .do d befor t openi e p su on the lowlands of Perth, both of which I have seen. One of these, however, at South Tulclian does not seem to me to bear any resemblance to a fort, and I have placed it in the class of dubious remains. The other is More, 6 m. . sidN.N.E N 'thf e o heathera e e f th f Crieffo Sinao . n p O to y' . Glen,e th n o ' hill, 1520 ft. above the sea, with a steep descent of 850 ft. to the Almond Eiver e S.W.o th nwitd an ,h short, steep, rocky descents elsewhere, except eastwards, where the approach is nearly level, but by a narrow neck. The enclosed, irregu- larl ys almosi ovap to l ts girdle i n overthrowlevel a d y an b d, n stone wall, resting with its inner side on the very edge of the slope and the outer one several feet lower, with rudely built portions still standing, but no well- built face remaining. Many of the stones are large. The cross diameters e thicknesth e wals probd th an wa f l o s-, ft 0 9 d an 0 e interioth o15 f e ar r d mosan . t. ft AnnexeE accessibl 0 e 1 th ablo t o t d ey8 crescentia sid s i e c work wit ha greates t interiod enclose. ft an a wal. 0 ft ry 30 b dlwidt 0 3 f ho long, similar in size and condition to that of the main work. I have given no "6BSI 'u tiaawaoaa 'Aiaioos SHX AO 89 FORTS, 'CAMPS,' ETC., of PERTH, FOKFAK, AND KINCARDINE. 69

plan of this work, as Mr Hutcheson's plan of Gaisteal Dubh, No. 30, might almost althougstanit, dfor lattehthe mucris h larger. an y s uppedit Ta 27 vallee e .r th th Passin tributariesf o t yo w gno e mosth , t remote work is Dun Geal, on Creag a' Chaisteal, 900 f r. above the sea, -!>- in. N.E. of Fortingall, Glenlyon. It has been well noticed by Miss Maclagan, as well as by Mr Hutcheson,. who describes it as standing on the summit of a hill with a considerable extent of easy sloping ground around it; circular in form, 58

Fig. 26. Dun Mac Tual. (Mr Hutcheson.)

ft in diameter within the wall, which averages 10 ft. in thickness, showing outside th insidd ean e faces intac severan i t l place e masonrth s; y very massive, being erected almost wholly of stones separated either naturally or by art from the hill, and splitting easily into cubes. 28. The fort highest up the Tay proper is the very interesting ' Fort' of the M.'0 , Gaisteal c Tual, c Tuatlialr . HutchesonMa E M Ma . n f 1\o m Du r o , 70 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

of Fortingall, 1000 yds. N.W Taymotitf o . h Castle it0 . 90 abov, e the e on sea n o , of the eastern spurs of Drummond Hill. The chief facts in Mr Hutcheson's very accuratd fulan l e account, whic I hgive , wil clearle b l y understood from his reproduced plan (fig. for26)e Th t .occupie summie sth bola f o td projecting rocky knoll with precipitou , huE. t d descendinan . sS slopee th g o lesst s steeply on the other sides. The walls follow the contour of the hill to suit the exigen- cies . of the site, and thus form a very irregular somewhat square figure, the inner area measurin e pla. a shor ft th As ni 0 n B to 21 wal gy b aboul 0 30 t branchin o includt f of g a enatura l platform e curveTh . de N.E th wal t .a l angle enclose sa spac e. long ft abou 5 , 16 whert e ther a spu s i er rising towards W.e summite e mosth th th , n tO vulnerabl. e sideadditionao tw , l walls have been thrown up on the side of a hollow, some 25 ft. deep, that separates the fort from the ascending slope of Drummond Hill. At the bottom of this hollo sprina ws i waterf gmaio e Th n .fore wal mucs th i t f lo h broken downt bu , in places the outer and inner faces still stand, and I give a sketch (fig. 27), which

Fig. 27. Remains of Wall, Dun Mac Tual.

I took many years agoportioa outef e o ,th f nro face, wher bess i t eti preserved. Mr Hutcheson judges fro quantite mth stonef yo s lying below this face thae th t original height must have been'12 to 15ft. The width here at the base is e outeth ; som r. facft es almos9 i e t perpendicular e inneth t r bu ,slope s con- siderably. Mr Huteheson states that the name is attributed to Tuathal, son of Argutso, Abbot of Dunkeld, mentioned in the Annals of Ulster as having die A.Dn di . 865. e Dun,29Th . Tyndun, 1 m. S.S.E f Aberfeldyo . , occupie e summith s t oa commandinf g height, 110 . abov0ft e sea,-anth e s describei dr M y b d Hutcheson as being an almost complete circle, 110 ft. in diameter inside the wall, which is from 14 to 15 ft. thick. It is almost entirely destroyed, but outside th insidd ean e clearleb facen yca s see o havt n e been forme f uno d- FORTS, 'CAMPS,' ETC.F PERTHO , , i'ORFAR KINCARDINED AN , . 71

dressed blocks 4 ft, . froo t long m2 , closely set. Ther e appearancear e a f o s narrow entrance on the W. side. Three trenches or earthworks, 150 ft. long wide5 1 o t , giv2 1 ed additionaan l protectio mose th to n t vulnerable side^which is on the S.W. 30. Castle Dow, most unaccountably called " site of a supposed Pictish fort" O.M.e o nwitth d remaino an ,h n s r Hutchesos markeshowi M , it y nb t d a o nt be an undoubted fort (fig. 28). It is situated 2|- m. W. by S. of Logierait

Fig. 28. Caisteal Dubh, near Aberfeldy. (Mr Hutcheson.)

Church, but on the opposite side of the Tay, 1050 ft. above the sea, on the top of an isolated height overlooking the valley of the Tay from Aberfeldy to Dal- 3 ft.,21 somewha y e b s aTh i guise0 n 30 t . ovaI tE f . o sinuoule th n o s wal mucs i l h e S.Wdestroyedth . n o sid t ebu , thera lengt s i e f abouo h 0 5 t ft. tolerably well preserved, allowing that the masonry resembled that of Dun Mac Tuale wal s thickerTh i l. , however,s i t I . ranginft 6 1 go t fro 4 1 m faced outsidwitn i d h elargan e blocks crescentiA . c annexn , exactlDu t a s ya More (No. 26), strengthens the most vulnerable side: it projects about 80 ft. 72 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1890.

and. is 240 ft. long. This side is further strengthened by a trench. The entrance to the main fort was apparently about the middle of the W. side, and the contains remains of a circular structure 15 ft. in diameter. In this fort and in others, as at Dun Mac Tual, Mr Hutcheson noticed evidence of bankin witp gu hwalle bace eartth th f . ko t h a

(J) STON ENORTE FOETTH Hs OCHILSE o SSLOP TH F O E. 31. For a stretch of 12 in. from the W. end of the Ochils eastward the northern slope of the chain is destitute of forts, and the first to occur, errone- ously marked 'Roman ' on the O.M., is 3 m. N.E. of Gleneagles, on the top of Ogle Hill, If m. E.S.E. of Auchterarder Church, 800 ft. above the sea, and 600 above the Ruthven Water to the N. A ridge descending gently northwards from Coul Hill (100 e maith 6 ft.nd e )an chaith f o n yardOchilw forfe o t ss a m riset Oglbu s e Hill before falling abruptle th o yt valle f Ruthveo y n Water n apparentlA . y natural trench-like hollow. ft 5 7 , wide (B, fig. 29), separates the ridge from the site of the fort, A, which occupies the whole of the little summit of Ogle Hill, and is defended towards the ridge, reckoning from the interior ,scar a firs . highy ft b pt 9 (sectio, , withouB) nA t parapeta t coverebu , d with loose stones, probabl e remainyth a wal f o sl fallen ; secondlyrampart e scarpw p th lo e footo fro f a th e o , t my t th fallinb a , g ou a terrace, which with a small trench forms an outer line of defence. The terrace has a rampart, D E, at the E. end only. These lines are much injured, and the only signs of defence on the flanks of the site is a doubtful double terracing. The interior may be about 150 ft. in length. nexte 32Th . , also erroneously calle d' Roma n Outpost e O.M.th . u -s m jo i ,' E. of the last, on the top of Ben Effery (1200 ft.). This hill lies between dee e e steesmalled th th pan pd Pairnee glerth an f ravin. no W y Burne eth n ,o of Green' se E. Burth , n whicno t theiha . belorft junctios 0 topwit 60 .e nar From this directio Effera remarkabln s nBe yha y bold, sharp, conical appear- ance, and the fort on the top is very strong by nature on three sides, particu- larly on the W., where it stands on the edge of a mural precipice (E, fig. 30), accese th t s eassi bu y fro e ridgmth e descendings i upo t i t frond i S.e an m,th overlooked from the summit, D, of a little elevation on the ridge, 250 ft. from the interior of the fort. The S. end, therefore, was defended by three parallel drystone walls, B C, now cast down, 12 ft. apart, occupying a width of 60 ft., which run straight from the precipice edge for 70 ft, and then curve for about 50 ft. more to the angle of the E. face. The entrance is through this shod curveen w . dN e steepart. d sidE Th an pe.e faceth f o s traces of three .or four terraces in continuation of the defences, till no longer neede oncn do e more touchin precipice gth e edge. f Auchterardero . . furthe 1. m 33 |E . rm N.E.3 , d a ifor,n an s o t Rossie Law, 1064 ft. above the sea and 900 above the valley. A single oval e O.M.th n o , surround . p ft (figto 'Ring,0 e . 50 31)th s y 'b .abou 0 60 t a level s e i actua , Th p dryto l , grassy, pleasan0 15 t spacey f aboub o 0 , A ,27 t ft. The ring approaches it on the S. within about 30 ft., and at a lower . distanft 0 d an 30 t t separates ft.0 i . tillevebu 1 ,N t i f l o e l th s n froo t mi about 70 ft. lower. The approaches are long and difficult, usually steepest 011 reaching the ring, when they ease off to the top. From the W. and S. the FORTS ' CAMPS,, ' ETC. ? PERTH01 , , FOKFAR KINCAEDINED AN , 3 7 .

Figs. 29, 30, 31. Forts on the Ochils, near Auchterarder. 74 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

descent rocke almossar d yan t inaccessible a smalonls i r t yfo I le . spacth n o e defence N d . thath .approac e E an th t. eE remainine s gentlehth i n a O .s i g grass d stonyan y mound e . entrancebroaft th t 5 a d1 , , whic. E s fro hi e mth parte stretca r masonryo on n 1 coulfo ,f abou n e i ho ft.0 t se d,10 ttherbu , e n lengthf largi o . ft w e ro ,3 stones placea e loweo t th s i 2 t ,d a re edgth f eo ston (elevatiop . ybelo to ft mound s 7 w it d section d nan an , , D), perhape th s base of a former wall built on the slope. A very large stone lies uncovered at the entrance. On the inaccessible S. and W. sides there is no mound or wall, but merely a terrace at the edge of the descent ; and there is no sign of an inner defence for the level top. 34. Continuing stilnorth-easterla n li y directio n. e slope N alonth f e o sth g range for a distance of 5 m. we reach Castle Law, 2 m. S.S.E. of Forgandenny, on the top of a steep slope 700 ft. high, and about 950 above the sea, isolated on the other sides by short steep declivities. This fort was excavated in 1891 by the late Mr J. Weston Bell, Rossie, with the co-operation of Mr Lindsay-

Wood, the proprietor, and was planned by Mr J. H. Cimninghain, C.E., then the treasure r Society.ou f o r 1 Some years afterward I investigates e groundth d below the fort to the N. where several trenches and mounds were 'visible, and foundes i r Cunningham' fig2 M 3 . n do e th f o s e for n th pla tf ow no wit y hm ground below. Two irregularly concentric oval walls . thicke inneft e oute th 8 th ,5 1 r, r"1 were found, unde grass-growe rth n rubbish, still standin l roun gal a heigh o dt t oft.6 f ,o frot wit m2 h well-built e facesdimensionTh .0 36 e s ar ove l al r e onl . endTh E y e . entranc, th ft approache t 0 a 19 s i e a causeway y b y b d , and 10 ft. wide. But it only pierces the outer wall and there is no entrance through the inner wall anywhere. The gateway had bar holes, and the entrance within it was flanked on the S. side by a mysterious broad wall con- nectin e outed inneth g an r r enceinte d prolongean s d inte interioroth - Be . ginnin entrance th t ga ea considerabl anr fo d e distance . sideS alone th e , gth outer surface inneth f ro e s piercewalwa l squarf o d w witero ha hole d an s slits running deep into the substance of the wall ("fig. 33), containing a great quantity of charred wood chiefly in powder, evidently the remains of timber thabeed ha tn incorporated wit walle hth . To the N. the ground sloped gently to the edge of the steep descent, along which ran a low mound, concealing the remains of a substantial wall, enclosing between itself and the N. wall of the fort a space (a, fig. 32) equal in area to the interior of the main work. shora . S te steeth o pT descent fro wale mth l falls upo na trencn pla o n6 h( and enlarged section), with a mound on its outer side. The trench begins at the S. side of the entrance,'whence a ruined wall projects so as to flank it; and at its W. end it is blocked by a little mound, d, which descends steeply beyond it upon the lower level space, e. Beyond the trench is the gently sloping space , ,whic/ h bears trace f terracingo s , bounde secona y b dd trench, e firstt diverginth bu f ,g o , beginnind en g . eastwarW ge d th closan do t e ending abruptly after a course of about 150 ft. In the first half of its course it has a rampart on its inner side. Beyond the second trench comes another nearly level space, h, bounded on the S. by a steep little slope or scarp, slightly mounded on the edge, and 1 , Proc.Soot.,A . S 1892, xxvii. 14 . FOBTS, ' CAMPS,' ETC., OF PERTH, FOBFAB. AND KINCABMNE. 75 '• x-\s^>^''

Pla nUppe£ r Section,

>/777* K 7777m7777wro ^ W7r77rr7W7mrw- „ ^V t/r/rniri

Sections AB,CD Fig. 32. Castle Law, Forgandenny. 76 PROCEEDING E SOCIETYTH F O S , 'DECEMBER 11, 1899.

falling on a narrow marshy space, i, with a slight decline westward, on the S. side of which rises the mass of the hill-range. This slight mound is prolonged westwar encloso t s a e rectangulao eth ds r marsh e W th .e foof yth o tt flata , e , end of the fort-height. From the W. boundary of this apparent reservoir rises the prolongation of the range, and from the N. boundary the ground falls to the plain 600 ft. below. The spaces / and h unite eastward in a wide, nearly flat esplanade, k, whic s unfeiicedhi thao e s easter, th t n flank of/and li are quite open althoug . fronS fortifieds e i th th .

Fig. 33. Holes for Beams in the Wall of Castle Law, Forgandeuuy.

35. 5 in. W. of Castle Law the Ochils strike upon the junction of the e uppeth Ear d Ta d rFirte nan yan th par Tayf hf o hereo d t an , , perchen do a little knoll, looking down upon Abernethy, the ancient capital of the Picts, stands a fort excavated in 1896-98 by Messrs Alexander Mackie and James Mar thaf ro t place, latterly with som d fror 0 Societyai e m ou 75 e sit s ei Th . ft. above the sea. and about 700 above the strip of carse land that separates the Ochils from the Tay to the N., and is strong by nature, being only approachable FORTS, '.CAMPS/ ETC.F PERTHO , , FORFAR KINCARDINED AN , 7 7 .

- o 78 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

o FORTS,' CAMPS,' ETC. PERTHF ,O , FOEFAE D KINCARDINE,AN 9 7 .

fronarroa We y mth b . w neck s wil A e seeb l. n fro e planmth s (figs , 35)34 . , a wall has been drawn across the neck, connected by a short branch with the inner and proper wall of the fort, and parting gradually from it as it crosses e necd descendth an k s toward a damrned-us p little d locn marsa an hn i h elevated flae r hill-masreceso th t f o s s (fig. 34). This wall, although com- pletely concealed before excavation stils wa l, standing unde e debristh d r an eart . widft h t accumulatioa 8 e1 s wa d a heigh0 ft. 1 an o n,t f froo t o t mG basee th , faced with excellent masonry. The wall proper of the fort, also showing no sign of its existence on the level top of the site before excavation, completely enclosed the oval interior, which

Fig. 36. Outer Face of inner wall, west end, Abernethy Fort. (Mr F. K. Coles.)

measured 136 by 51 ft., and, as in many other Scottish forts, was built with its inner face at the very edge of the slope, and the outer one some feet, in one plac , perpendicularft s muc a e9 s ha , dow e slopth n e below e widtTh f .ho this . higinnes stilft wa 7 hleslo rn t i walss d thaft.5 wa l2 an ,n o t fro 8 m1 t Castl a place e (Nos w on A eLa .. n 34)i , timbe beed ha r n construce useth n di - a muc o t tion ht largebu , r extent whereves a , r bot outee d innehth an r r walls were examined, squared channels, in a double row (wherever the wall still stood high enough to show the upper one), were found running in from the outer face for about 8 or 10 ft. The regular loop-hole appearance of the openings of these channels arid the general character of the masonry are well shown in the measured drawing taken by Mr F. E. Colea (fig. 36). On dissecting a part of the inner wall, longitudinal channel r timbefo s r were also found. Full details of this very interesting fort wil e founb l. n Proc.Scot.,i A d . S 1898-99, vol. xxxiii. pp. 13-33. . ...

(c) STONE FORTS ON MONCREIFFE HILL, PEKTH. 36. The isolated Hill of Moncrei/e, 725 ft. above the sea, 3 in. S.E. of Perth, rises from the N. with a moderate slope to the edge of a precipice, from 80 PROCEEDINGS OF THE SOCIETY. DECEMBER 11, 1899.

which a very steep wooded descent, 600 ft. high, falls towards Moncreiffe. e edg th vallee e f thiEarn Houseo n th th sO f d yo mura. ean l precipico etw forts have been perched firs whichthe , of t , coming alon ridggthe e fromthe

. CarnacFig37 . , Hil Moncreiffef o l . scarcelw no . abovs i ft yseae 0 d recognisableeW.th an 60 , s i , O.Me Th .. givet si an oval form measuring about 280 by 150 ft. 1 had some difficulty in find- evidency an moung a f in et o las a r wall do t t discoverebu , 'distinca d t masf o s rude masonr chanca n yi e grounde breath n ki . I'OKTS, ' CAMPS,' ETC., OF PERTH, FOEFAR, AND KINCAKDINE. 81

37. j m. eastward, about 120 ft. higher, and on the very summit of the lull, is a comparatively well-preserved fort, marked Garnac on the O.M., which gives the title of Moredun Top to the summit of the hill. Perhaps this is an error suggested by the fort, i.e., ' big Fort,' because J m. N.E. there is a house marked not Moredun but Moreden. The fort (fig. 37), perched on the precipice edge, is quite inaccessible on the S., and on the N. has the protection of a long descent Wd ,approache s biii .an . t E fro e m th shory db t slopes. The remains are so dilapidated and overgrown that it is difficult to plan themmaie Th n. wal apparentls i l y double, except toward precipicee sth , where f o p e edgeto th t retire singles e i n bu , th o t d i biiil t t da an , no . tabou ft 0 2 t a dangerous e totaslopTh le e wall-mas widt. dowth it f o o ht n s aboui s t t standi d p nowher0 ft.u an 4 s , o buildinN es . morwa gft 4 e r thao 3 n . fouW r e a poincarefull th t visiblwalla n e o t th bu ,n e yi laid stone roa wn si look like the base of the outer face. At the W. end, another strong wall springs from the fort close to the precipice edge, and circles at a somewhat lower level three-fourths of the way round the inner wall, nearly enclosing a considerable crescentic space on the W. and N., but no connection with the inner wall at the E. end can now be made out. On this wall also there are traces of a facing. Still further out to the W., at the edge of a very steep short slope, I found remains of another less substantial wall, but it could not be traced far. Separated from the fort, by a slight hollow to the N.E., is another little stony, flat-topped eminence t mucno , h lower tha forte nth , wit a steeh p descent to the N. and N.E. The site was much obscured by weeds, but seemed to be surroundee ova)b o t remaine . ,long abouth ft d a stron0 y an ,db f 22 to s g wall. Thi beyons si plany dm . The fort proper measures over all about 210 by 180 ft. On the N. and E. sides direcn i , t connection wit wale l . alonghth ft al l extendind 0 5 an , o t 0 g4 into the interior, is a mass of overgrown debris 3 to 5 ft. high. The only entrance to the fort is at the E. end, piercing the double wall and mass of ruins obliquely from the N.E. In the raised mass near the entrance a circulae for s th i . sid t f N o en diameter i re . spaceth ft an n 1 o d2 , , sur- rounded by a mound-wall 3 ft. wide. Two others of about the same size are near it, one of them touching it. Further W. are several small, round, saucer- shapeentrancthe sid S. dof eraise the ehollowsthe d On mas. s contains eigh ninr o t e simila saucers,r' irregulao tw n interioi 'e r th res rowse f o tTh r . vacants i , excep another fo t r circular enclosure, measurinft.0 2 , y whicb 5 3 gh . touchesideS e .wale th sth n o l

(W) STONE FORTS ON THE SIDLAW HILLS. We have seen that on the extensive Sidlavv range there are but two earthen .forts e easter, th bot n ho n sid e; one , Evelick a hig n ho , sitee otherth , , Rait, low down at the very margin of the Carse of Qowrie. The number of stone forts on the E. is the same, but there are also probably two on the W. side, t clearllattee althougno th f s ri o y stonee f proveo hon e b . o dt 38. On the E. side of the Sidlaws, on the gently rounded top of Dron Hill, 684 ft. above the sea, i m. W.S.W. of Dron Farm, and If m. N.W. of Longforgan Church a for s i t , e O.Mwhicth n . o hlook s liko ovaletw s inter- secting each other longitudinall ye groun th (figs pretti n t .o i d38) yt Bu . VOL. xxxiv. F 82 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

evident reallthas i t i tovan ya l work, witsmale hon l crescentio e anneth n xo N. and another large one on the S., springing from the same points of the oval. The fort proper measures 330 hy 250 ft. over all. The greatly dilapidated wal generalls li y grass-covered onld an ,y abovo rise tw e surfacea foos th r e o t . A wal I lbelieve f largo e W th , .ew howevern sidro o s bee a es ha a n t i , stones, extending . southwar abouft 0 10 t d closeldan y sethardln ca , e anyyb - thing else than the base of an outer casing, and a less complete row shows the position of the inner face, giving a width for the wall of from 9 to 11 ft. There are basal stones, also, here and there along the S. face, where the

Fig. 38. Fort on Dron Hill, Longforgan.

mound is 27 ft. wide, perhaps from digging operations to remove the stones ; and some very large stones have been torn from their place and taken a short . sideE d s r cartinan havdistancfo . f N ei g bees e awaya e Th n almos. t completely destroyed, and are merely traceable. The enclosing mounds of the annexes are not defensive; they are very low, only 6 or 7 ft. wide, and are probably ruined walls. 39. Tlie Laws, on the Laws Hill, Drumsturdy, an outlying eminence rather tha e Sidlaw na parth f o t s proper s situate i ,f Monifiet o . dW y Ihfb . mN . Church, 431 ft. above the sea. The site is on a small isolated height, partly FORTS, 'CAMPS,' ETC. PERTHF ,O , FORFAR D KINCARDINE,AN 3 8 .

precipitous, the easiest access being at its E. end, and the fort occupied the wholnearle th f eo y level top. Thi lon so perhapo s fort gt ag no ,bese sth t preserved in Scotland bees ha ,o ns knocke confuseo s de recorde d contradictoryar aboutth t dan i d f so an , , thao t t for intelligibly man e , difficulideit f ao t enoug beste th impossibls i t , ha e with- out some reference to its modern history. The earliest notice, published in the Trans. . Soc.R of Lit.papea n i r dated 1827 t fro,bu m observations made about 1790, is by Dr John Jamieson, who described an outer wall, 500 paces in circum- ference, neaE. e bottoe ,th hile rth slantin n th o lf mo g graduall southere th p yu n slope till half-way betwee precipice e brine summith nth th f d ko tha n an eto t sidecontinuind an , asceno gt t circlei s da d round til t reacheli N.Ee dth . corner where one entrance was, another being apparently at the W. end. It is strange that no one else seems to have noticed this low-level outer wall, but perhaps it may have been covered by the operations to be described immediately. The inner wall, according to Dr Jamieson, surrounded the summit at a distance of several paces fro outemthe r one varyin,but g accordin naturthe the gto of e ground alse H o. mention threr o o e sothetw r short walls running fro outee mth r one to the brink of the precipice. n 185I r Neish9M t thaa , t tim e proprietoreth , excavatee th e sited dth an , result recordede sar , with additional note Johy b s n Stuart voln i , r . ou iii f .o Proc., . 440-54pp Neisr M .h states that before 1834 great quantitie f stoneo s s were carried off for farm purposes, and Mr Stuart estimates, on the authority of a man who took part in the spoliation, that 9600 cartloads were removed in the four winters ending 1818. The author of the Statistical A.ccount of the Parish, 1842, confirms this by the statement that persons then alive remembered when the ruins wer. higheft e5 r than whe wrotee nh . Mr Neish also states tha 183n i t whole 4th e surfac levelles e wa portion d dan s of the wall thrown over near the E. end. Finally came his own excavations in 1859, the discharge of the rubbish from which seems to have been down the hill S.Ee ath t . corner. I have found it impossible to form a connected idea of the disposition naturd fortificatione an th f eo s fro r Neish'mM s accoun s excavationhi f o t d an s the very sketchy plan.1 But for the difficulties and anomalies suggested by observationn ow truste d y spoe ha m th I to d n t thesealonef si o d ,an ,shoul I d have hesitatioo n d ha pronouncinn i wore gth havo kt e bee ordinarn na y stone fort of excellen masonry dr t y wite th hduplicatiot a a t no wale f i th t eac a lf d no hen side withoud san t reckonin longeo n e grth visible oute rr Jamieson walD f o l . onle yTh part tim d tha studha o e t measurI td yan e with some eth cars ewa E. end, where Mr Neish's excavations were either most thorough, or are best preserved showplas y sectiod A m n an . n ni n (fig. 39) e firs,th t thin t witgme h from the outside is a terrace, D, 15 ft. wide, at the edge of the descent. Mr Neish explains that this was formed by the rubbish thrown out in the levelling operations of 1834 and in his own work. I may add that a pleasant terrace walk runs all round outside the wall, and probably is everywhere due to the same cause. mai1A n difficult understandinn yi Neish'r gM s description arises fro treatins mhi g n aree placa th as a e divided into irregular space y wallb s faces (insteaa s a f o d fortification surrounded by walls. 84 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

Within this terrace d standingan , e origina e th edgo doubt n th f , o en o l, descent, comes a wall, C, 8 ft. wide, .with good facings, exposed for 3 ft. of height inside6 rubblee r Th o outsid. 5 , d whicean h doubtless exists betweee th n facings s perhap s hiddewa ,i t lai w greeno sdy r Neish nb no M bar nd y eb an , turf. Then follows an interval, B, between this wall and a second wider one, A, s outewhicit s r hha facee first th , f , liko e excavateinne e th e fulon rth l o t d inne e depthplace th on d rft.e7 n fac,i an , e only sufficiently cleare shoo dt w that ti is there. This I take to be the wall of the fort, going all round. The outer wall, on the other hand, diverges fro e inne mth e eastward on r , so thae intervath t l between them increases from 8 to 12 ft., and is probably 14 or 15, at the far end. Thus t Abernethya s a , e outeth , r wal s apparentli l y detached fro e maimth n

A B C D Lawe th f so . EasFortd 39 Figen t. , Monifieth. wall of the fort at one end at least, although it may have - run down the slope to join Dr Jamieson's problematical outer wall. It could not be that wall, as he say shile that th thioutefooexcavae e e la f th th o sttTh t endra wals .- wa l tion of these walls has been a good piece of work, and they are still well seen, C for 160 and A for 180 ft. of length. The wall of enceinte, A. as I took it to be, is exposed at intervals all round, and the W. end, more fully opened up, seemed to have an additional wall, althoug t distinctlhno y shown. Neish'r M n I s excavations great quantitie rubbisf so h testifie formee th - o dt ex r istence of extensive building in the interior, but the only structure of consequence remainin foundatioe th s gcirculawa a f no r tower wit fee w hwala tfe higa l d han pavea . 1 thic 8diameterft n d i . dkan ft are 6 3 a, accordin e planth t I o .gt was of brooh dimensions, therefore, but there was no opening in the wall, save FORTS ' CAMPS,, ' ETC. PERTHF ,O , FOKFAR, ANJ) KINCARDINE5 8 .

narroe th w entrance evidenco n d staira an ,f thaeo o ,s t cannoi t provee b te b o dt a . Vitrifaction.— usuae There commoth o s to i l n contradictory evidenc thin eo s point. Jamieson makes the startling statements that " This fort consists of two wall f vitrifieso d matter, which surroun buildinge th l Mild e "al dth e an "s; ar as regularly vitrified as the walls" ; also " the vitrifaction is as perfect as that of the Fort of Finhaven." Mr Neish, on the other hand, declares that vitrified masses were indeed found all through the ruins, and often as if used for backing up the walls (and here I may explain that what he calls the walls are really wale th l faces thao s ,appeare h t meao st n thae vitrifieth t d masses were used in the rubble between the faces), but never as if the walls (i.e., wall faces) had been vitrified. There can he little doubt that we should prefer Mr Neish's state- ments, as they resulted from a careful excavation, while Dr Jamieson relied on surface observation neee w d d hav an lese s; e th s hesitation t seemi s a , s extremely unlikely that anyone would remove 9600 cartload f suco s h useless building material as vitrified masses must be. Moreover, I saw no vitrifaction on the extensiv y observationm e s wala r lfa , visiblefacevitrifactios w a go s no sd ,an n is never found unitin e block a th handsomelg f o s y built walls i face t I . possible, however, tha vitrifiea t d fort existed here previously, from whice th h vitrified masses used in the rubble work of the present ruins may have been obtained. 40. Dunsinfian, the first of the two stone forts on the N.W. side of the vere primitivSidlawsw th yf fe o e ,on e seem e Scottisb o st h fortresses mentioned e earlth y b annalist t leasa s; t o othethern s i er existing claimane th r fo t honour of being the Dunsinoen of the tragedy which led to the murder of Kenneth, King of Alban, in 995,1 at Fettercairn. modern I n time firse noticr o Johth st t Si e placs neth Sinclairwa e n i o wh ,

1772, when he can have been only eighteen years of age, gathered its local tra- ditions. 2 These wer f littlo e e f thao consequence e verd tth en yt a t bu , century, James Playfair, D.D., then ministe f Meigleo r , afterwards Princi- t AndrewS f o l spa University, made some excavations, which were recorded

wora firs t generalln i no kt y accessible, d agai3an n (very briefl n boti y- in h

5 stances) in a work of his own, 4 twenty years later. Chalmers makes some sensible remarks about the place, and James Knox sums up the information known about it down to 1831, besides giving his own impressions from a per- sonal visit.0

1 A.n. 995. " Cinaeth MacMalcoIai i AlbamR a suin s occisus est. Tigh. (per dolum, Ann. Hit.)." " Interfectus est a suis hominibus in Fotherken per perfidium Finvelae filiae Cunohar comitis de Engus, cujus Finvelae unicum fiHum predictus ICyneth interfecit apud Dunsinoen." Ghron. Pictsd Scots,an 175 (Skene'7 ,28 s Celtic Scotland). 2 Enmities of Scotland, Forsyth . 319-21iv , . 3 General View Agriculture e oth f f theo County f Perth,o James Robertson, D.D., 1799, i. 380. 4 Description of Scotland, 1819, i. 488, James Playfair, D.D.

5 Caledonia, \. 503. 6 The Topography of the Basin of the Tay, 192-202, 1831. 86 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R .

But it was not till 1854 that a second attempt was made to reveal the nature of the place by excavation, undertaken by Mr Nairne, the proprietor at that time, the results of which were recorded by Dr T. A. Wise.1 Unfortunately, he seems to have paid only a fleeting visit to the works, so that he was led into serious errors, which I am now able to correct from my own observation and fro mcarefua generalld an l y reliabl . recorexcavationse th eMS f do e , kepth y tb late Mr Andrew Stewart, Kinrossie, Oollace, who took part in the excavations, for the loan of which I am indebted to his heirs, through the kindness of the r EevBaxterM . , F.S.A. Scot., Free Church Ministe f Cargillo r e lasTh t .

;^W:y>^ m ^V^s^^; f ^' "^WlC^^W^^ t^Si&AX %^'^-\fe ^-•-o I

""lUuuk, 1° «^ ^J?-|&=S 'ii»mih,,.-"V""IIIUlL ^...A^-,"'* <^ "-"'-^ • "Y/ -?s•: - /s^S--

Fig. 40. Dunsinnan. (Mr Hutcheson.)

printed notice is by the Eev. Thomas Brown,a Gollace, which, however, adds very knowledger littlou o et . followine th n I g accoun I tshall thesal lf emako sourcee us e informatiof so n and point out the very serious misconceptions to which some of them have given rise; thus illustratin e evith g l result f unskilledo s , incompletd an e hasty excavations, undertaken too often with the object of proving preconceived theories. e placvisitn th M yeow o t s wer n 1898i e n compani , y wit e Kevr th h M . 1 1859 , . Proc.Scot.,A 93 . . .S ii tit.,. . 378ix Op 2 , 1870-72, communicate Alexy db . Laing. FORTS, ' CAMPS,' ETC. PERTHF O , , FORFAR KINCARDINED AN , 7 8 .

Baxter assisteo takinn i wh , e d m ga roug h plan, corrected afterwardf o d ai y b s Mr Ritchie, factor on the estate ; but this has been superseded by the plan and section place) disposay 40 kindnese m th t d s d Alexandea r y (figsan lb M 9 f 3 s.o r Hutcheson, architect, F.S.A. Scot., who generously waived his intention of writin gspeciaa l articl Dnnsinnann o e , whe heare nh engageds thawa I n td o the present work.

18'

Fig. Profile41 . Dunsinnanf o s r Hutcheson.(M . ) The Site.—The hill on which Dunsinnan stands, 7| m. N.E. of Perth, is e Sidlawsridga parth f f o teo , running paralle e N.Wth ln .o with sidd ean , maie ofth , nt height inferio n wels i chaini no t i d t l isolateo an ,bu t ,r d froe mth other summits on the continuation of the ridge to the oSf.E. and S.W. The ascent from the N.W. is gradual, but is steep and difficult in other directions, and the flat-topped fort is perched, 1012 ft. above the sea level, on a little green knoll which crown narroe sth w summit. 88 PROCEEDINGS OP THE SOCIETY, DECEMBER 11, 1899.

The name, therefore, corresponds well with the site, if we accept the deriva- tion from Dun and sine, 'nipple,' as the fort stands on 'the nipple' of a swell- ing hill top ; but those who desire a more romantic origin will no doubt prefer the derivation, which has also been suggested, from shaiu/an, 'ants,' because " Macbeth's men swarmed up and down like ants when building the castle." Present e Fortifications.Aspectth f o — The nearly leve p disheto l s clown e middle th onlr e Baxter th M o t y s ascertaine aboud y . a ,b ft an , e 3 t m r fo d suggestio fortificatiof no n roun aree sligha dth s ai t ridgin2 1 edgee o th t t g6 ,a ft. wide, not always perceptible to the eye, but made out by measurement. On the steep short slopes of ' the nipple ' are a series of narrow terraces, some- times level, sometimes higher toward e outeth s r edge, sometimes becoming veri- table trenches shows Hutcheson'r ,a M n i s plasectiond nan s (figs e 41), Th 40 . . trenched character is most marked on the N., particularly on the line A A, but

Fig. 42. Diinsinnan. (The late Mr A. Stewart, Collace.)

even there the depth of the deepest and lowest trench is only 3 ft. The number of these terrace trenched an s s, varieS. e th s n froo e m N.E e on thre th o t .n eo where, however, another may have disappeared owing to the steepness of the extreme slopeth t . A endeE . e slop,th e end precipica sooo ss n no e that there is little room for terraces, and I could not make out more than one. Mr- Ste wart's plan (fig give) .42 versa y cramped representatio f the no named man s them ' Ramparts, ' bat this seem vera s y inappropriate title -for them. Neither is there the slightest sign of a lower wall hinted at by him. Plateaux below Fort.—e th Immediately e lowesn fronth i f o tt e trencth n ho N. there is a little level space, but a much larger plateau, 500 by 200 ft. on e O.M.th , interrupt e steeth s p southern descent . belo, ft abou e fort0 wth 10 t, f a towi ther s a s d welCastl na ewa s an a lf Dunsinnano e e sitth ,e would naturall thesn o e eyb plateaux somd an : e neighbouring easy slopes e tracTh k. FORTS CAMPS,,' ' ETC. PERTHF O , , FORFAK D KINCARDINE,AN 9 8 .

of an ancient wall round the edge of the large plateau is marked 011 the O.M., bu, lookin sigit I o t coul n f no e dse g down fro forte mth . General Dimensions.—The upper area, from cres e slighcreso t th t f o tt ridging e edgeth t ,a measuren greatesi 5 n lengti 10 . ts y ft b breadthhabou 5 18 t , which is towards the W. end of the oval space. At the corresponding part of widte th hd narrowen ft 5 th. .6 E e Measureo t s d ove e nippler th all ' , ' with its terraced or trenched sides comes out 325 by 240 ft. on the plan. The Fortifications as revealed by Excavation.—As we have shown, the green top and sides give no clear indication of a rampart, but shortly before 1799,

originally constructed, founde e rock d th neatl an n , o d y buil largf o t e stones." Passing to the excavation by Mr Nairne about sixty years later, it is amaz- ing, after Dr Playfair's experiences, to find Dr Wise describing and figuring the fortification of the top as an earthen , 20 ft. broad at the base and tapering to a height of 10 or 12 ft. (fig. 43).

. SectionaFig43 . l View of Dunsinnan r Wise.(D . )

But however this erro confidentle h ry arosema t i y, asserted tha suco n t h rampart ever existed, and Mr Stewart's MS. account of the 1854 excavations amply confirms Dr Playfair's description in all respects but one. Mr Stewart relates that the workmen, going in from the outside, cut through a mass 21 ft. thick, findin walga l "strongly buil f largo t e stones, both insid outd ean , while the filling in is a mass of loose stones, entirely without manipulation or mortar." We may dismiss as fanciful, therefore, the ' red mortar' of Playfair, who was probably misled, as will afterwards appear, by finding red clay among the ruins. Thus the wall of Dunsinnan falls into line with the well ascertained walls of the other Scottish stone forts that have been properly investigated. Mr Stewart gives no section along with his plan (given in outline, fig. 42), but I have con- structed a section (fig. 44) from his very clear description which will aid the reader, as we proceed, in understanding the position of the wall and interior buildings. I need only point -out just now the position of the wall at B. Entrance.—At present ther vera s ei y distinct entranc e N.Eth t .e a corner , penetrating obliquely through the defences to the central area, and this entrance is distinctly mentioned in the Old Statistical Account of 1798. It must have existed, therefore time th f Nairne'e o t a , s excavation a statemen r d M an , y b t Stewart tha tracewale s th t wa l roun dal d without findin entrancen ga , muse b t intended to apply only to the stone wall. This is quite likely, as there was certainl entrancyno e throug innehthe r stone walls leasat , t near their baseat , Forgandenny and Abernethy forts. Interior Buildiitys.—Dr Playfair states that, " having diligently exploree dth 90 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

. a deebelo ft t e surfacep3 wcu th area trenc w d ,no an , h across it, foune "h d no vestige of building, but he cannot have gone deep enough, as the excava- tions of 1854 revealed much building at the E. end of the area. Here agai r WisenD , misleimaginars hi y db y vallum, regarde chambere dth s foun undergrounds da , i.e., beneat originae hth Stewarr M fortle t flooth bu f t; ro clearese showth n i s t manner that the ye floor stooth n . do Kevertiny m o gt section (fig. 44), constructed fro descriptions mhi roce ,th ksees i floor, nD slopD , - ing up gradually to the centre of the fort: on this stands the wall of fortification, withi d passagechambersa Ee ,an d th t nan betweei , , H ,F ,F walle n theth ,d man besides the " Quern Chamber," G, the whole enveloped in a mass of ruins and covered by accumulated earth, etc., under the grassy surface, ABC. Unfortu- nately constructioe th , f thesno e chambers remains quite doubtful e planTh s. Stewarr Wisr oM ) fD d 43 utterle an d tar (figsan y2 4 .irreconcilable equalld ,an y so are their descriptions, Dr Wise giving them a rounded figure with converg- mortaro n wall g d in an s, whil r StewarM e t declares tha l tsquare al the e yar , with perpendicular walls thatd an , , althoug fore mortarho th t n wald t ,lha "ye e squarth l al e buildings contain considerable quantitie mortar.d re s f o sHi "

A

D

Fig. 44. Excavated Wall and Buildiiigs, Duiisuman (from Mr Stewart's description).

expressions on this point are rather obscure, as if he had not seen this mortar actually joining the stones, but only lying about, and he probably really means use e claytere h th s m a , mortar very loosely anothen i , r sentence speakinf go admixturn "a f decomposeeo d bone charcoad an s f woodo l s ""a compouna d mortar." It will be observed that Mr Stewart gives on his plan several remains of build- ing t notice r Wise'sno D n di s accountconsideree h . S mas A e .f ruith so d n no on very slender grounds to have been a tower, and two passages leading to it he regarded as the only early British works, because they alone had converging walls alse H o . says that " within one-hal larga f o fe circular opening towards the west of the area, which seems to have been an open court," foundation walls still stoo fooda t high t tha othee bu , tth r half contained none. Vitrifadion.—Dr Playfair make mentioo n s f vitrifactionno d Williamsan , , originae th l discovere f vitrifiero d forts Knod an , x declare thert nons bu e ewa ; Wise found some deep in the excavations of the chambers ; Laing also dug out several pieces Steward an , t asserts t onlno , y tha e wholp th t to ee ruinth n o s were full of vitrified stones, but that many stones of the wall were fused with trap rock, sandstone quartd ,an z intlumpe oon . This read f vitrifiei s sa d blocks FORTS, 'CAMPS,' ETC. PERTHF ,O , FORFAR D KINCARDINE,AN 1 9 .

wer meanin s emasonrya e uset th no n di gd thaan , wale vitrifieds th t wa l n O . whole th e ther hardln ea doubca e yh t vitrifiedt no tha wale s th t wa l , although vitrified blocks, either brought fro ma distanc r takeo e n fro n oldema r for1 01 t the same site, may have been used as biiilding material. Water Supply.—The excavators foun withiy traco an dn e fort f t nth o e bu , springs exist not far off down the hill, and Stewart speaks of a ' Flatt' im- mediately below the rampart wall to the S., covered by a " body of red mortar " (cla ") y ? whic havy hema bee nwatea r tank." The Finds consisted of (1) a quern in a fixed position in one of the chambers. (2) A spiral bronze finger ring, described by the Rev. Thomas Brown as of most perfect workmanship fore a serpent th f m o n i , e eyebace scaled th ,th an sk n so carved regularl verd yan y yeaa Nairneminutelyr kepn s M i r y t wa b t t bu , I . wa authorite s th ) Andlost r Robern (3 o M ., f yo t Chambers rouno ,tw d tablets of metal resembling thef brasso me on ,engrave d wit legene hth e sconc d: " Th e of kingdom come until sylphs in air carry me again to Bethel."l Conclusions.— mucIts regrettee i b o ht d thae excavationth t s wer impero es - fect and so unscientifically conducted and reported ; but on the whole we may conclude that Dunsinna stony defendes dr newa a e waly th b f d o l p rounto e dth ordinary fort type and on the slopes by additional works, now assuming the aspect of slightly trenched or level terraces, the true character of which has still to be ascertained. Also, that it contained much building, the ruins of which, with the accumulated soil of centuries, favoured by burnt timber, of which many traces remained, brought the top to a nearly uniform level, completely burying both wall and buildings: that the evidence as to the form and structure of the building contradictors si quitd yan e unreliabl saie e s b : tha d same a y th t ema to the presence of mortar, by which probably clay was meant: that the few relics cascleao n t e rbuildinge origi th lighth f n no o td finall an : y that thers ei nothin separato gt e Dunsinnan fro e clas mth f ordinar o s y Scottishy fortdr f so masonry. 41. Denoon Law.—The work here is a somewhat questionable example of a stone fort, as I could only see faint evidence of rubble work rather than building in the very few breaks in the massive green rampart. But Dr Playfair and Mr Knox both state distinctly tha wale buils th tf ston lwa o t e without cementd an , absence th trenchesf eo presence th , stona f eo e threshol entrancee th t da wels a , l as the whole aspect of the enceinte, is that of a stone rather than an earthen fort. e We positioth . Th sidn o f Denoo o es ni n Glen, her broaa e d open valley, t soma t ebu distance froe streamth m , 2| m. S.W f Glamio . s Churchd an , . abovft seae 9 eth . 68 The Law stands up conspicuously from every near point of view, being quite isolated and 100 ft. in height. The sides are steep, and the south-eastern ascent end lina precipicef n esi o edge th f whic o en o , e forhth t stands (fige . 45)Th . area of the fort occupies the whole of the level top, and hence has a squarish oval form, and is girt by a single mound, massive'and lofty at the N.E. end e N.Wth ann .o d face , wher considerabla r efo e stretc leso t hii sn s o t tha 0 n1 14 ft. high inside, a very unusual elevation, the base here, however, being

probably natural. But even at the precipice edge on the S.E. the rampart is Picture1 of Scotland, Robert Chambers, 1828. 2 9 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R . high. ft 7 . o t Nea 6 S.We rth . angle, however moune th , d seem havo t s e been altogether removed, and at the S. W. end it is only 4 ft. high, although the

Fig. 45. Denoon Law Fort, Glamis.

necessity for defenc greates ei s almosi r t therei e d tN.W gonth an , t a .e angle. Thus it would seem to have been much injured at the S.W. end of the fort. There seems to have been an entrance at the much injured N.W. angle, but FORTS, ' CAMPS,' ETC. ,PERTHF O , FORFAR D KINCARDINE,AN 3 9 .

the only distinct entrance now is near the S. angle of the N.E. end. It is 7 ft. wide, has a stone threshold, and passes between the massive ends of the ram- part, of which that on the N. side lias a broad platform in its rear, while the shorothea s trha terrace arrangemenn a , t which flank defendd san entrance sth e inside. The top of the rampart is a mere crest on the N.W. face, is somewhat r threo o e tw littl S.Ee s broaderth e' ha saucer. n d o an p , 1to hollowe th n o s N.Ee . broaft th 0 t .a 1 d end o muct e s shor a ,6 th s s i hface ta t par thbu ,n o te S. side of the entrance being pitted with several' saucer' cavities. In the area of the interior there are some inequalities of the surface, particu- larl ywida e shallow circula larg a . side r d N hollo ee ,an irregulath t wa r mound, perhap rubbissa h heap, near wher rampare eth s wantini t S.We th t .ga angle; these are probably modern interferences. The dimensions of the interior from crest to crest are about 380 by an average of 230 ft. It is difficult to give the measurement over all, as the position where the rampart merges in the slope is unknown, but it must be about 430 by 280 ft. On the outer slope at the N. angle and thence to near the entrance in one direction, and half-way along the N.W. face in the other, are two terraces, with a third below theshora r mfo t distance neae entrancth r d middle th an e; e eon subdividee ar shore e th on t d by compartments mannea n i , r difficul describeo t t , but shown in the plan. Possibly these spaces were levelled and subdivided as sites for houses. Three short parallel terraces are also faintly visible on the e entrance . th sid S f e terraceeo Th . t e noticeDenooa sar Alexy w b d nLa - ander Gordon, who compares them to those of Romanno. They are also noticed t describeno t bu Knox.y b d n planninI g these difficult e forobjectth t d san generall1 y I had the benefit of help from the Eev. Dr John Stevenson of Glamis, annephey dm r ArthuwM r Cassels Brown.

(e) ISOLATED FORT NEAR ALYTH. 42. Barra Hill Barryhillr (o ) Fort situates i d 1J m. N.E Alytf .o h Churchn o , spua Alytf o r h Hill. This long ridge, after risin gsummitso fro tw We mo th .t , 966 and 871 ft. above the sea, falls on the E. to a neck about 500 ft. above the sea, over which passes the high road up Glenisla, and to the E. of the road the ridge again rises abruptly 150 ft. to a little eminence, 680 ft. above the sea, conspicuous from being thus thrust inte e Islavalleth oth .f o y Here stands the fort with steep rocky descents on all sides. The earliest notice of Barra Hill is by Dr Playfair, who saw it probably at

eighteente th f o d en h alss e century,th ha o t i bee d n2an describe r GeorgSi y db e Mackenzie, whose visit must have been early in the nineteenth century.3 The fortress consists of a regular oval work (fig. 46), measuring on the O.M. ft.0 , 12 20occupyin y 0b levee gth l summit d variouan , s appurtenances, which can only be understood by reference to the plan, founded on that of the O.M., t witbu h considerable additions d witan ,h section y myselfb s . Approaching^

as a visitor would naturally do, from the W., we first meet a trifling mound- 1 Topography Basine oth f Tay,e o . th f172 p , James Knox, 1831. - Description of Scotland, i. 485, James Playfair, D.D., 1819. 3 Archceol. Scot., iv. 184. •6681 'IT aaawaoaa 'Aiaioos HHI ao somaaaooHd FORTS, 'CAMPS,' ETC. F PERTHO , , FORFAR KINCARDINED AN , 5 9 . stee e edge th th f p et o a ascent fence, A A , , boundin a gplateae foof th o t t ua th ' well,e e forwhicn o th 't Bs a circulai h, r grassy hollow, e stonth t ya bottom, 3 ft. deep, about 50 diameter, and quite dry at my visit. Looking up from this plateau to the N. we see a steep ascent to another . higherplateauft 5 2 , , C ,partiall d domiy mound w an enclose , lo -D a D ,y db maie nateth ny db wor k; lookin fro firsw e m, anotheth gno t E. platea e th r o ut ascent, E E, is seen, at first gentle, then steeper, interrupted by the wide terrace leadind an , F ,g directl rampare th for e o yt th t f propero t . s Ganothei r rough plateau, unfenced, highe d verr an tha y, C nslightl y dominate maie hile th th ln y descenddb H work t A . s direct fro e ramparmth t of the fort, but beyond that a terrace, I, runs round eastward and southward as far as J, where there is a fall of 2 or 3 ft. to the trench, N. The mound, K, about 50 ft. wide on the top, is probably natural, but carved into shape t bearI . a littls e mound . endsouthward N an ,e ,th Jc t ,a d gradually narrows to join the very regular, evidently artificial, mound, L/ the s rea . it wide a ft wid r n d whici cres6 an e,r s whicf o o t terraceha h5 s hi M ,

Fig. 47. Profile at Barra Hill. (Sir G. Mackenzie.)

e regardeb (whic y s ma continuouha d s wit e terraceth h , IJ), rising onla y couple of feet above the 6 ft. wide trench, N, from which it is fenced off by a low moun . end. E BeyonO de , th onl t L therda littlea y s i e , perhaps natural, sharp-cut trench, T,' shown only in my section, having on the further side a narrow, rough ridge, U, beyond which is the steep descent of the hill. This ridge would for ma goo d natural firs t. sidS lin f f defenco ee o e th n o e forte th . Finally, wit hbola d comman l rounddal , except toward e plateaud th s an , G , mound-head t slighte rampart bu , e wherforth K s th ,i s ti , f t i eo QP , , R S proper r PlayfaiD . r describe s . higha thimounft s n s8 a o , f earto o t d 6 h whic hwala freestonf o l buils wa et without cement whicf o , e foundatiohth f no rough granite remained, 10 to 12 ft. wide, the same width as the top of the mound. I cannot help thinking, however, that he was mistaken, as there is no evidenc sucf aspece o moune th hwallth a w f d eno o texactls di an , ye thath f o t ordinary ruine partialld dan y grass-grown wall f Scottiso s h forts. Neitheo rd I know any example among them of a wall constructed on the top of an arti- ficial mound r GeorgSi . e Mackenzie tell thas sstonee u e rampar tth th f so e ar t a red conglomerate from the hill, but does not mention any masonry, and I coul nonee dse . Vitrifaction. — Dr Playfair and Sir G. Mackenzie saw only a few vitrified masses in the ruins of the rampart, and I noticed none amidst the great masses of stone outes it inned sn an ro tha re slopesli t r PlayfairD . , however, describes 96 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

a thoroughly vitrified 'bridge,'18 ft. long, only 2 broad in the middle, but widening to both ends, and covered with gravel, as crossing the trench (at the . MackenzieG r Si plan)y t m poin Bu , n whil.i tJ e confirmin e vitrifactionth g • , says and shows by a section (fig. 47) that this bears no resemblance to a bridge, and I can amply confirm him, in regard to its present state, as there is now nothing but a slight descent at J to the trench, N. Possibly a rough and pitted irregularit terrace th f yflat-bottomer o eo d trench visiblew , neaI ,no y , J rma , have suggested the idea of a northern side to this ' bridge,' which, unless it was purposely greatly reduced between Dr Play fair's and Sir George Mackenzie's thaw t cannolo i t o visitss s t wa hav, e bee anf no y bridgea us s ea . Dr Playfair calls this the only vitrified part of the fort, but Sir George, having found more vitrifactio pointe th t , nmakea I , extraordinare sth y sugges- tion ,e tur thath f f i twer e removed vitrifiea , d waly l woulwa e foune th db l dal roun same th d et froNtimA o t confesse.e m I eh s tha cannoe h t t accounr fo t a vitrified wall being found in so extraordinary a situation as the bottom of a trench. entranco N e throug rampare hth markes i tO.M.e th n d,o althoug r PlayD h - fair speaks of one, apparently at the E. end, secxired by a bulwark of stone. I could see neither the one nor the other.

(/) FOKTS ON HIGH Rin&ES BETWEEN FORFAR AND BRECHIN. A few miles N.E. of Forfar two narrow parallel ridges, their crests from a mile to a mile and a half apart, run north-eastward, between bhe broad fertile valley of the South Esk on the one side, and Rescobic Loch and the pass through which road and rail escape from Strathmor e Mearne th n routeth e er n fo o s other. These ridges rise boldly to a height of 750 ft. above the sea, and 500 above the low ground to N. and S., but the hollow between them is only s continuou i . deep ft d 0 an , 30 s wit e table-lanth h d whice hth f carrieo e on s roads from Forfar to Brechin. Where the hollow debouches on the table-land is Aberlemno, celebrated for its large assemblage of early Christian monu- ments. The furthest north of these ridges is 5 m. long, beginning at Carse d endine Wth Graan . t n Finavoa o gy n e HillotherTh . , beginning near Luuanhead and Pitscandlie, is 3 m. long, and ends at Turin Hill. A large fort occupies each of these ridges at or near their highest points, and there is als smaloa l wor doubtfua f ko le Turicharacteth f no d ren toward. W e th s Hill ridge. . Turin43 Hill Fort, e Campth Kempr o e StatisticaCastleth f o l Accounts and of Mr A. J. Warden's Angus, is a very remarkable work of almost unique character and extent, and has hitherto escaped description, beyond the brief notice sauthoritiee inth s mentioned abovels yi t inI .. S.S.W f Abero . - lemn . abovft o e sea 0 e summiChurclth e80 Th . d f Turio an ti n Hila s i l peculiarly narrow level ridge which runs E. and W. and is 1600 ft. in length by only breadth n aboui 0 fore 10 ttTh .wit appurtenances hit s occupie l thisal s space rund an , s besides some distance dow gentle nth e descen N., e thao th s to t t the total occupied area is about 1600 by 400 ft. The descents from the narrow K. and W. points of the summit are steep, and to the S. the fort looks down t murabu w lsteea frolo clif n edge a o mp fth f edesceno Rescobio t e Loc0 h50 ft. below. As it is on the dominating point of the two ridges, which, as I have FORTS, 'CAMPS,' ETC. PERTHF ,O , FORFAE D KINCARDINE,AN 7 9 .

explained interposee ar , d here betwee coas w Grampiane lo nth t e rangeth d ,san it commands a most extensive and beautiful view, and has a position of strat- egical importance s proximitIt . e verth yo yt ancien t Aberlemno alsy e ob ma noticed. worke Th s wil bese b l t described fro weste mth , whenca y b visito e d eth le s i r farm road gradually skirtin e hilth l p sidgu e from Clocksbriggs Station nearly summite toth , whe t turnni s northwards throug hlittla eAberlemne pasth o st o road shor A t stee. bu t p ascent from this pas se sharp-pointeleadth o t s d W.

A ~~-

Scal thr efo e Plan. (Sections double)

0 0 105 100 0 0 0 F30 ?50 20Z.O0O Fig. 48. Fort on Turin Hill.

end of the summit. Here (A, fig. 48) is a double row of cup or saucer shaped hollows, possibly foundation t circleshu r sfo . Fifty yard . ove sft ovan a further8 ls all i 10 work n y ,o r b coverin ,0 B13 , g the whole width of the top, and environed by a mound 24 ft. wide and 2 or 3 high, grass t witybu h small stones s entrancevisibleha t d i ,an s. E fro e mth whole th W.d e an interio, irregularls ri y cuppe moundedd dan . Nearly 100 yds. further, after crossing a wide mound, 0, which we disregard in the meantime, we come to another mound, D, encircling the long oval space E, a VOL. xxxiv. 98 PROCEEDINGS OF THE SOCIETY, DECEMBER. 11, 1899.

about 500 by 180 ft. over all, the sides being not far from straight, the ends rounded. The enclosing mound is grass-grown, but shows small stones, and is no less than 35 ft. broad and 6 high on the outer side at the W. entrance. On the N. nearls sidi t ei y distincts gonei t bu , , though much ruinede precipicth n o , e s agaii . sideS d n e an ,wel th edg ln . o eendpreserve E e ,th wher t e a d th e entranc e peculiath s eha r for mbehins showha plane rampar e d th dth n an ,ni t . sidS spac ft.ea 5 e 3 , straigh th f abouw enclosey eo n b lo o 0 a 7 t y t b dmound , divided an d into fou r fivo r e compartments. Within the oval, much nearer the W. than the E. end, and touching the N. side, is the nearly circular ' ,' F. A mass of rubble all round, with many fine facing stones at the base, proves it to have been about 90 ft. diameter, insid e fin . thickth y eft ewal a dr d 4 qualit 1 e an , l o th frot f 2 yo m1 masonry is testified by a part of the outer face, 8 ft. long and 3| high, still standing (fig. 49). Many of the facing stones are 2£ to 3 ft. long.

Fig. 4.9. Masonr Turif yo n Hill Fort.

Fifty yards furthe s anothei . E r r nearl. ft y 0 circula13 y b r0 work14 , G , ove grassw r lo d stonall ye an ,th y encircling mound being wide. abouft 5 ,1 t wit n entrancha e S.Wth o et . Withi secona s i t dni encircling moun. ft 0 d1 wide innee th , rl bein areal f adiametero n gi . abouft 0 5 t . Abou yds0 . 10 t furthe e com a massivw ro t e e. moundft 0 4 , abouo t H , 0 3 t e eastere edgth wide th e entrancf . o et th sidS a n ,f eo e descent t th i e n O . rapidly diminishes in bulk, and soon reaches the precipitous edge, but on the other sid t i continuee a considerabl r fo s e distance curving roun Ne .dth sid e tracede b n dowca , hillmore d n th til y an ,e t joinli oftewa terracea e ss na th l ,al e siteth precipice ,f o wherth d n o moundt e run en i et e . th ou sW , nea e C , th r edge; a very large semi-oval space of about 1100 by 370 ft., enclosing the inne re entranc . th ovasid d S citadelf o ean e l th a es thu i n , O s shu . in t second . moundoutsidft 0 7 firse d lowe , eth I ,an t r dow e hillnth , circles round with it. At first it is 30 ft. wide, but soon becomes a terrace, and I could not trace it more than half way round. 44. Finavon.—The ridge of Finavon, Finaven, or Findhaven, running parallel with the last, after attaining its full height of 751 ft. above the sea, FORTS, ' CAMPS,' ETC., OP PERTH, FOKFAR, AND KINCARDINE. 99

gradually falls north-eastwar contou0 50 e rth lineo t d , wher windina e g road e ridgpasseTh e s. theS ove o nt t froi r . risemN s again steepl. almosft 0 y10 t immediately to a little level summit, on which stands the fort, a mile due W. of Aberlemno Church in the plain below. This summit is approached by a long moderate ascent from the N., and by a short but narrow one from the E., a ver s yha stee t i e t S.pth ,bu falbeginnino t l a well-define t ga d precipitous

180' 150"

Fig. 50. Finavon, near Aberlemno.

forte edgeTh , .curiousl y t enoughthia t s withdraws i edgeno s t i , bu , n fro0 m5 to 100 ft. from it, and it is also slightly withdrawn from the steep ascent at the W. end. The plan (fig. 50) is very simple. Two long straight sides face N. and S., and end o wele tw ar s l e roundedth . Accordin e O.M th e dimension o th .gt s from . endW e nea,5 e th contractin 12 th t r a y b 0 . 11 ft o 0 gt cres 50 creso t te ar t s therefori . end t E I . a peculiarle y narrow e moun fortTh . d rampart, even 100 ' PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

on the side towards the steep descent, is in some places as much as 7 ft. high and 30 broad, and on the N., where its outer limit is not so easily made out, it ia at least 40 ft. wide where my section was taken. At the W. end, on the other hand, whict naturallno s hi y strong s singularli t i , y weak, being narrow and abov. onlft y3 littl ea e platead en u. E whic e th ht comeA . it s righo t p u t it is connected by an earthen ramp, F, with a perhaps natural mound running parallel with that end, bat extending in an irregular manner to within 40 ft. . N . dowe ft e d slopsomth e 0 stee th nan o th t 10 f. e p S descene th n o t Ascending from the N. the visitor is deceived into the belief that the space between this mound and the rampart must lead to the entrance, but finds him- cula sel n i def sac.t presenA cara t t road hile , windth A l, frop u N.e s mth , passes close beneath the rampart from E. to W., crosses the fort obliquely, and goes down the steep descent to the S. Thus the only two existing entrances formed e s somewhai ar t I . t remarkable thaflae th tte S. spacth , whicn o e h same th en o s i . end E d e an , th t a 0 and en 10 d . W e . wideabous ft th i t 0 a 6 t level as the interior, is neither fortified at the edge nor on its flanks. interiore Th sectionshows y a , m n ni , has . end, E nea e ,shalloa th r w pit, G, 45 ft. wide, and at the extreme. W. end a very deep one, H, shaped like an inverted cone, with grassy sides, descending directly froslighe mth t rampart there to a depth of 30 ft.

Accordin Williamso gt . JamiesoiiJ r D , Statisticae th , l Account O.Me th d . ,an this for vitrifieds i t . Jamiesonopportunite th d ha 1 f investigatinyo t gi whe n the tenant was clearing away part of the wall, but unfortunately his descrip- tion is almost incomprehensible. He says that after piercing through 8 or 10 ft. of rubbish the vitrified wall was found regularly built and standing from 10 to 14 ft. in height and 20 to 30 broad at the base. But the wall was not all vitrified, as he says that parts from top to bottom afforded no vestiges of fire, although others were completely burned. Apparently, also, it was not the built wall that was vitrified, for he says " the irregular concrete mass formed a buttress on each side for the regular intermediate wall." He also says that the stones were brought from various quartersmale on ln i hea s: foune ph d sever no eight different kind stonef so thad an ;great a t quantit ashef yo burnf so t wood was mixed with the stones. The rampart is now much overgrown, but a con- siderable part of its slopes are bare, and wherever that was the case I found the stones much bound together with a dull grey, ash-like substance, so that the fort seem havo st goos ea name titlda th o ee' vitrifiedt moss a ' t others onle yTh . true glassy piece I picked up was lying on the surface inside the fort, and had probabl lefd somy b tan eyp u recenbee g ndu t explorer.

((/) STONE FORTS ON THE HIGHLAND AND LOWLAND BORDER OF ANGUS. The Oaterthuns.—6 m. N.W. of Brechin the glens of the Grampians debouc a strea n o hm that a perfectlrun n i s y straight a . tm cours6 r fo e right anglestreae th themo . r t s4-|m s know m i e -Paphri Fo th . s na e Burn; t thei n join Wese sth t Water, which follow same th s e north-easterly coursr efo anothe halfra mil d , ean whe t ni take widsa e ben d flowdan s eastwar joio dt n Norte th h Esksouthere Th . nvallee sidth f formes eyo i correspondingla y db y 1 Tram. Jloy. Sue. of Literature. PORTS, ' CAMPS,' ETC., 0V PERTH, FORT'AR. AND KINCARDINE. 101

,,

,'.'.»' .•''•• " - - o-

eo \o

o o<-

''''/ l '' '•'•/''. /lf

'''•' ''.'>','', '/''^;;-'';''';i/;u"1"u"t'1:^ / ; '"' ''>,f 102 PROCEEDINGS OF TEE SOCIETY, DECEMBER 11, 1890. straight ridge, 7 in. long, which has no general name on the map, but might be^ called the Menmuir ridge, rising into a series of tops from 800 to 1000 ft. above the sea and 300 to 500 above the lowlands of Forfar. This ridge, therefore, form remarkabla e bulwark betwee Highlande nth Lowlandsd san upod , ,an n it

....'•;.'}•'•' .^^^•'•'''''''•'•''•''•'''•Vv^^'''"''.,.^'''2/,''/'V.;;;....

,- V/'^-'f'"' ' "'''''''''''^''^

"' " '""' '"• '

#'" ' """"''*v

3'f -'•>••'' ; ? ^

.% % * % ^ ^r ^

Fig. 52. The Brown Caterthnn. directly opposit Wese eth t Water pass throug Grampianse hth , stan famoue dth s C»f«r(AwM«, two great forte, placed on rival summits of the ridge, 978 and 943 ft. abov e seaeth , whence they look upon each other acros holloa s pasr wo s three-quarters of a mile wide, with the Highlands of Forfar at their back and the Lowland same th f eso count t theiya r feet. FOISTS, ' CAMTS,' ETC., OK PEIITII, FOKFAR, AND KINCARDINE. 103

generaA l knowledg theif eo r position, sized ground-plaan , derivee b y dnma from Roy and the Ordnance Survey t Roy'bu , s sectiont agreno eo d swit s hhi plan, and the best descriptions, those by Roy and Miss Maclagan, are brief and t varianca e with each other. The two forts dilt'ev in almost every respect except position and size. The White Catertlm ovas linens i e it drawn formi lar sd an ,n with some regaro t d regularity; wherea e Browth s n Caterthu s nearli n y circular d is an , very irregular in its tracing. An enormous mass of exposed debris testifies that the formea ston s ewa r fort, although strengthene intrenchmeiitsy db , whereae sth latter, as far as surface appearances go, was mainly an earthwork. Another contrast is in the number of entrances, the White Caterthun having apparently hat onedbu , whil othee hav y eseveth d ema r ha eight r no . All these differences are well seen in the plans (figs. 51, 52) taken from the Ordnance map. Possibly they indicat worketwo thas werthe t contemnot e - porary, and that the White succeeded the Brown Caterthun. 45. White Caterthun, nearly 5 m. N.W. of Brechin. It is impossible, with- out excavation e origina th mako t ,t ou e l structur e Whitth f o ee Caterthun. Indeed variet e dispositioe th , th n ye greai th f tno mas f stono s e debrie th n i s circuit of the inner enceinte is so perplexing that it is hard to lay it down in plan, and impossible to understand, even in an accurate plan, without numerous sections. Roy speaks " of the astonishing dimensions of the rampart composed entirely of large loose stones, being at least 25 feet wide at top and upwards of 100 at bottom, reckoning quite to the , which seems indeed to be greatly filled up by the tumbling down of the stones." And he accordingly shows this mighty rampart in the section (fig. 53, C), reproduced on a larger scale for comparison with my own. But Miss Maclagan correctly points out that the outer stony slope, which Roy took for a part of his rampart of loose stones, is merely the hile slopth lf covereeo d with fallen debris fro mwala l above. This comet sou clearly in ray general section (fig. 53, A, enlarged at D), taken with no other instrument than a tape, but with the expert aid of Mr J. H. Cunningham and Mr F. R. Coles. Section D, across the N.E. face of the fort, shows—a, The inner slope of debris, rising onl. abovft ye interio4 eth t thia r n s i muc s pointa 7 t a ha bu , some other places. 6, The top, 31 ft. wide and hollowed, as seen also in section E, taken on the S.W. side. This represents the fact, which seems to have escaped observation hitherto wale ,s whol th s it tha i lf n o i te p circuito e th t pitted with shallow, mor lesr eo s circular ,diameter n hollowi . ft 0 3 s ,o frot 0 m2 forty-eighd an forty-ninr to numbern ei exterioe e,, Th c , r slope ftlengt5 n .i , 5 d han 14 in height, covered with debris. Here it must be remarked that the debris is not continuous; 28 ft. down there is an unencumbered strip of grass, d, 6 ft. wide, beyond whic e trench debrie e edgth hth th f eo o st . . lieft s1 agai2 r nfo I found this stri grassf po , although occasionally e overruth n o stones y , nb be o t , whole a marke, d featur l alonal ee N.Eth g e existenc. Th face , a e f ,o e trench herbees eha n doubte Misy db s Maclagan thinI evident s i ,kbu t i t enough, as show botn I nfouni generan i h . ver e t E sectiondi b d yo t lfre an esD from stones. /, A mound 25 ft. wide, rising only 2 ft. above the trench and 5 abov e exteriorth e . Theroutwaro n s i e d indication, tha I tcoul d s seeit f o , being anything but an earthen mound. Roy calls it " a simple earthen breast- 104 PROCEEDING E SOCIETYTH F O S , DECEMBE , 189911 R .

tr-

o FORTS ' CAMPS,, ' ETC. PERTHF O , , FORFAH KINCARDINED AN , 5 10 .

work," and probably he is not responsible for the extraordinary development and form given to it in the engraving (section C). On comparing section D with E, taken respectively on the N.E. and S.W. face f tho se work, differences come out e ,compara th probabl o t e -du y tive gentlenes e N.Ee slopth th n .f eo faceo s . Hers neithei t i e lono r s r gno interruptes i t i ; sD o n dhigi s abouha e middl th tlevea y eb l . spaceft 7 , g , wide, and the whole surface is thickly covered with debris, instead of having e grassth y intervae S.Eth f .o l section e trencTh s .agai i h n quite distinct, outeo n s r ha mound t i t bu . impossibls i t I e from these surface characteristic understano t s e originadth l structure, bu I tbeliev e that excavation' would reveal remain f masonro s - yre presenting a wall possibly 30 to 40 ft. thick. Whether the pitting on the top is due to mere searching for stones, or had a structural origin, is a question which also could only be solved by excavation. Possibly there were really two con- centric walls, the interspace being filled up with cellular building of some kind or another ; and there may have been another and slighter wall situated at the e trench edge th e slopdivisioe th f debri e footh e Th o e n f th eth .o o t f n no so slope into an upper and lower strip, as shown in section D, favours this supposi- tion. It remains to speak of the outer line of defence. Eoy calls it " a double in- trenchment, s sectiohi , figd nC an " . 100, represent moundso tw s , h,, witi ha trench between. Wher takens sectioy wa m e) , howevern(D I ,foun d nothing trencha t bu , It, wit ha 15-ft.-wid e level terrace s rearit d thi , n an I,,si con- tinued to be the case along a great part of the N.E. face, although towards endo tw s a slighe ther s th ewa t mounding both insid outsidd an e trenche eth . totae Th l dimension White aree th th f a seo d Caterthuan . ft 0 e 13084 nar y 0b of the citadel is 470 by 210 ft. It seems as if there had been but one entrance, situated at the E. end, where no wmera e footpath passes ove hollora stone th ywn opposite i debris th t A e. similao tw ens e possibli rdar t pathsi d f theeo an , thame on tindicate e th s positio secona f no d entrance t neithe,bu thef ro m passes throug hwell-markea d hollow like the path at the E. end. e Brown46Th . Caterthun (fig. 52) ,f m. e lastN.E th s describe i ,f o . y b d Eoy as being " fortified by several slight earthen intrenchments." Their present slight character may be partly due to the boggy nature of a great portion of the ground. The inner ring (fig. 52, and m, section B, fig. 53), now scarcely trace- able in some parts, is so trifling where preserved that it is questionable if it was intended for defence. It encloses the level summit, a space of about 280 by 190 ft.; Imt a little way down is a second much stronger ring, n, a mound 8 or 9 ft. high outside, which probably, as Miss Maclagan thinks, conceals a stone wall, and appears to have been the main defence. It encloses a space of 610 by 480 ft. The next ring is a trench, represented as perfect by Roy and the O.M. I foun t remainsomo n bu ;i dn t ei f stretchesso , althoug N.Ee th t .i o ht wat onlsno y wel d slighl ha seen t t bu ,mound n fron i sd rear an t . Abou0 12 t ft. farther fourte dowth s nhi ringa moun , ,p d measuring . acrosabouft 0 3 st heighn i e outside 6 th r o o t 5 . d fift. Aboue lowean r outeft th o h 0 s i r3 rt ring, q, consisting of a slighter mound towards 20 ft. wide and rising only 4 or 5 above a narrow trench, which is bounded outside by a very trifling mound 6 width n i totae . ft Th olr8 . dimension 112e sar 101y 0b . 0ft 106 PROCEEDINGS 01' THE SOCIETY, DECEMBER 11, 1899.

entrancee Th s throug l theshal e ring e numerousar s . Perhaps som thef eo m modere ar n breaches s the a simple t yar bu , y opening d nothinan s g moree th , point cannot be determined without excavation. Roy seems to have thought most of them genuine, as he says " the second or strongest [ring] has no less than seven gates."

III. SITES OF FORTS WITH LITTLE OE NO KEMAINS.

1. Craig Olmey, 2i m. S. by W. of Dunkelcl, 1323 ft. above the sea, is marked by a dotted circle, 120 ft. diameter, as the site of a fort. I have no informa- tion about it. 2. Jackshairs, lj m. E.S.E. of Forteviot, 339 ft. above the sea, on the highest point of a little wooded ridge on the skirts of the Ochils. The O.M. represents two, perhaps three, concentric rings round an area of 180 by 150 ft. wite 0 wholft.. 6 th n inne,ft h a y e0 b rmeasurin25 0 8 rin y f b go 0 g35 I could only find three faintly marked trenche sidee on . n so . Auchterhouse3 Hill.—This well woode d finelan d y shaped summie th f o t

Fig4 Profile5 . sitf o Fortf s eo , Auchterhouse Hill.

Sidlaws . N.N.Wm 6 , f Dundeeo . , 140. abovft 0e sea th ee , th . N.E 1f m Jo . Parish Churc f Auchterhouseho s 'sit ha f ,fort o e 1 markee O.Mn th o .n do its very top. A century ago, Dr Playfair saw only faint traces of a fort. All tha I fount dstone w werfe esa suitabl r buildin fo e a edgstee th f o et p a g slope of 30 ft., on which were two concentric, low, artificial looking mounds which could not be traced far. This was the natural boundary to the S.E. hile th l . slopW . eN froOe blune nth m th t conical sit s continuousei d offeran , s no natura S.We l th e descenlimit o .th T s .ver i t e N.Eyt th i abrupt. o t d an , considerablfalla n so e plateau generae Th . l nature sitth e f wileo l -be under- stood fro rouge mth h sections (fig e interio s probabl. fore Th 54)th wa t . f ro y circular, and may have measured about 200 ft. diameter. 4. Dwmbarrow, t Kirktoa f Dunnicheo . uS y Farmb n. E ,Church . 2m J . A single oval ring of 120 by 100 ft. is marked on the O.M. on the top of a . abovknolft seae 4 eth , 54 l risin t littlgbu e abov hige prettd eth han y level land to the S., but with a descent of 300 ft. to the low country on the N. The site is much overgrown with brushwood, and I could see nothing but a few large stones lying about. FORTS, ' CAMPS/ ETC., OF PERTH, FORFAR, AND KINCARDINE. 107

IV. DUBIOUS WORK SITER SO S MARKED FORT OR CAMP ON THE O.M. . tSouth1 Tuklian.—A for s markei t de O.M. th her n 2 111o e, . N.N.Ef o . Fowlis Wester, 1 m. S. of the Almond, 600 ft. above the sea, in a field sloping northward to a small burn. The green elevation, 5 to 9 ft. high and 80 across, shows natural rock cropping out on the top. Breaks reveal quantities of large and small stones, and there is something like a circle of stones under the turf about a foot high, round the foot of the mound. This puzzling place beao resemblanct y e ran t seedi m dno mo t forta o et . 2. Duncan's Camp, Birnan o . S.S.Em m . abov ft Hill2 f seae 0 o d .e th ,60 ,an Dunkeld Church, is drawn on the O.M. as a low, flat-topped elevation, triangu- lar, with rounded angles I havt , . no emeasurinft 0 10 y gb abou0 18 t havd seeinformatioan o e n t n i n abou. it t 3. Caledonian Camp.—This pretentious name is given on the O.M. to an extensive enclosure, 3 m. W.N.W. of Blairgowrie, on Lornty Burn, Gormack Muir. An irregular rectangular space of about 4000 by 2000 ft. is enclosed by the ' Buzzart Dykes,' described to Dr Joseph Anderson by some natives as ' lea irregulan lo dikes.abov. greas a i ft sea e t d 0 I erth t, an ' 85 ground o t 0 60 , many cairn e markee O.Msar th n . o d e bot muith h r n o insidbeyond d ean . Dr Anderson saw some of them, and thought they were not sepulchral, but might be remains of shielings. 4. Rob's Seed.— e peculiaThisth s i r a wortitl f o ek also marke ' fortd n o ' the O.M., but which seems to me to be a dubious fort. It is 1J m. W. by N. of Rescobie Churc n irregulaa n ho r platea e summit quite th no uth n o ef o t ridge, here . abovabouft 0 e seaeth 55 t , tha tTuri o t run n no s Hille fortTh . naturao worn s kha n inne la strength s r ha circula d an , . rdiameter ft are 5 6 a , which . belofalle ft exterior levee th 5 s w th f . froo o t l W m3 , e excepth o t t It is enclosed by a low mound, 18 ft. wide and only 1J high, of small stones, as far as can be seen, with a few large flat ones not of the kind commonly used for fort wall-faces. 5. Druid's Camp, site of, is marked on the O.M. 3| m. S.S.W. of Stone- haven on high table-land 480 ft. above the sea. There seem to be no remains, an informatioo I dcouln t dge n abou e e nameorigiplace th th th t t r f eo o nbu , spote th .o t o g t no d di I 6. Camp, Montgoldnm.—Near farm-house th e th f o tha f eo p to t namee th n ,o very gentl t prominentye e ' Camp . abovO.M.e Hillft . th sea e 0 m f eo th '2 45 , , N.E. of Arbuthnott Church, is an irregularly circular mass of stones about 60 ft. diameter outee th , r edg f whiceo s markei h d irregularl somy yb e large rough stones not suitable for building. One huge block, apparently pulled from its place and blown up, must have been 6 ft. square when entire. But the mass consists of smallish water-worn stones, such as are abundantly ploughed up in the adjoining field. Mr F. R. Coles, after making a careful examination and takin gplana , believes that thiruinea s si d cairn numbeA . largf o r e stonee sli scattered to the N.E., which he thinks are the remains of a second cairn. 7. Blackill Camp.—The site indicated on the O.M., f m. N.N.W. of Dun- nottar Castle f Stonehaveo . S , . £m n harbour spaca I ,foune b e o encloset d d S.Ee a oprecipic nth y b . e rising froe hea mth f Strathletha o d n e Bayth n ,o 108 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899.

N.Wa straigh y b . t moun e S.Wa d6 wideth . y higonld b ft n . o ,han 2 y branc firste t encloseTh -t N.Erighe no ha th s . i n tt d o i angle last e d th an ;o s t named mound is on the summit, 250 ft. above the sea, of the enclosed area, which slope steepl o sprecipics e th o yt e thae worth t k cannot have bee na fortres f o s any kind, more particularly as the mound is so trilling.

V. DUBIOUS WORK SITESR SO , POSSIBLY MILITARY, NOT MARKED ' FORT' OR ' CAMP' ON THE O.M.

1. Dunnichen.—On the top of this detached height 764 ft. above the sea, -j m. N. of Dunnichen Church, overlooking the supposed scene of Ecgfrid, King of Northumbria's defeat in 685, the remains of a fort were visible according to Knox, op. cit., p. 118. I have not been there, but nothing is marked on the O.M. 2. Green Cairn, 1J m. W.S.W. of Fettercaim, 200 yds. S. of Cairnton cottaghige th hn enamo e road a littlth f s ei o , e eminence, measuring abou0 20 t t notey i hav M f O.Me o seth . been o y recollectio. nft m lost 0 t 14 bu , y b s ni that it was raised about 20 ft. above a level field, that it had an irregularly flat top, rough with many closely aggregated indefinite little mounds and frag- mentar t circleyhu s t (?)wit bu clea,o hn r sign fortificationf so . . Intrenehnient3 O.Me a straighname th f th n o .s o ei t mound. ft abou 0 45 t lon Drumsleen gi d Wood f Fordoun. ,S.Eo jm . abov. ft seae 0 e,th 30 , running paralle a littll witd ean h withdraw n fro e edga steem th f eo p descentn a n o , elevated flat. The parish minister, the Rev. J. Menzies, pointed out to me that several other mounds run off at right angles from the main one, but they are little more than broad undulation e groundth f so , wherea 'Intrenchmente sth ' is substantial in height and width. The remains are puzzling but not sugges- tive to me of fortifications. Knox says that the country people call the place Scotce Th ' h Camp.' 4. The Law, Tumulus, are the titles given on the O.M. to' a mound lj? m. W.N.W. of Kinneff Church, and 300 yds. N. of Largie Castle site. An ex- tensiv . side W ecuttin e showth n go s roundesandw fe ya soid stoneslan e Th . flat top measures 75 by 33 ft. Tt may have served as a simple mote or for- tress, but it is, I should think, of natural formation, as is a similar mound, with its long axis in the opposite direction, however, about 100 yds. to the N. . Castle5 Hill,f Kinneto . 100W f0a y b Churchse . ydse S edga th .f o en o , cliff representes i , strongla s da y marked artificial I founO.Mmoune th . dn o d it, however, to be quite inconspicuous and entirely without remains of building r fortificationo . 6. Malcolm's Mount, . W.N.Wm 2 f Stouehaveuo . f . ,Fetter N.Eo m £ . - esso Church ft0 . abov15 , seae e.th Growing crops prevente froe dm m gettinp gu to it, but it looked on a near view as if a natural mound rose to an artificial one. Taking the enclosed and' planted top as the artificial part, it is a regular circle 100 ft. diameter as marked on the O.M. 7. Remains of Rampart is the title on the O.M. of a mound, only a foot or two wide6 hig d h,an tha e promontort th cutf of s y calle Boivduns,e dTh . S.E Jm . of Stonehaven harbour, on the N. side of the Castle Bay of Dunnottar. This precipice-girt point consist e landwarth n so d a squarissid f . o eft 0 h50 flaf o t TOUTS, 'CAMPS,' ETC., Oi' PEKTH, F011FAK, AND KINCARDINE. 109

10 . abov anothef 0ft o e sea d th e ,an r fla extreme t th beyon t a et i d point about . •lowerr)ft 0 f abouo , t nece hale extentth th fk narroa t A . w geo cut fron si m the S., and at its head, nearly on a level with the flat, a ravine begins which gradually descends to the sea on the N.E. The artificial mound also starts d runan o s dowge fro e bottome e ravine heae rockth th th nm th th f f do o o st e of Bowdun e seath .t a s Thus, althoug t i hcut s acros e neckth s , fros it m positio d triflinnan g dimension n hardlca t e considerei sb y d defensiveA . similar mound in its rear courses along the top of the ravine, but this, like the other, seemed to me to be a fence rather than a defence. 8. Circular Stone Structures in Upper Glenlyon.—In this remote locality, 16 m. abov e sufficientlth e y remote Fortingall witt s squarhit Ha e eth forn i t Gean andDu l overlookin t frog i remarkable s hill i mth t i , findo et , withie nth lengt a mile f e pooho th , r remain f fouo s r circular structures o largto r , fo e ordinar t circlesyhu t apparentlbu , t sufficientlyno y stron welr go l buil havo t e been fortified towers, besides being on indefensible sites in the level bottom of

Chart oF Rums >n Upper Glmlyon

. FigChar55 . f ruineo t d "Towers," Upper Glenlyon.

vallee th ^ Towers, however e Gael, theth e O.M.e calle th sd yar n calan ,o d l them "Caisteal,...". as show, n i•n thee char• t (fig- . 55--)- -reduced from the 6-inch O.M. These names are also recognised by Mr Duncan Campbell in his Book of Garth Fortingall.d an Miss Maclagan e HillTh , Forts, &c., of Ancient Scotland, , call85 . s p them circular buildings d describean , d figurean s s thre f themeo , one of which seems to have a single row of stones, set on end, and therefore not very suitable for the foundations of the outer and inner facing of a broad rubble wall. Mr Cunpbell also finds no less than seven MotelviUs in Glenlyon, but they do not seem to be known as such by the people, and the O.M. does not notice them eithe namy b r r drawingo e , excep e SitJieanth t Camslai, . abov1m 7 e Fortin- gall, whic s e drawi O.Mh th a triangular n s no a . , flat topped mounw lo , d with a tower on it, but is not marked Moteliill. The tower appears to be a mis- r JoseptakeD s a , h Anderso a smal ns i inform lt i gree e m sn mounde On . might wish, however naknoo m t , Cuairteig,To w e morth f eo " directly above eth 0 11 PROCEEDING SOCIETYE TH F O S , DECEMBE , 189911 R .

Kermclach round fort, which Campbelr M , l says o ' put,t e Tinwaln th s Ma f do open shame.'" Circular Stone Structures n Strathardle.—i A large numbe f 'huo r t circles' and other circular structures of larger size in a very ruinous state have been de- scribed by Dr John Stuart at Balnaboth, and by Miss Maclagan, who also gives plan f themso f others o wela ,s t a lGlenderby,a s Strathardle I hav d ean ,see n similar remain s' Roma neae th rn Camp f Raedykeso ' , near Stonehaven. Doubtless they occur elsewhere in our district. They are always of an obscure, little understood character fald an ,l very doubtfull l undeal e heat a th r f di y of fortifications. GENERAL REMARKS. I. CLASS OP EARTHWORKS. (a) Possible Motes. Number.—About half of the twenty-five earthworks (including two e Postscriptth n i ) have some claim n structurao , l grounds, although generall a feebly e e onetitl f th Mote.o eo , t Wit e exceptionhon e th , most that can be said is that they are, on the whole, more like motes than forts. Height above e Sea.—th Like the other earthworks, they contrast strongly with the stone forts in this respect, as their elevation is quite moderate e onlthao Th tw yt . clim b0 contou 50 abov e rth e line ar e Arnieful, d Dundee0 ft.an ,70 Law, 2 ft.57 , and only four others contour0 reac60 e hth . Sites.—As position1 01 a rul t ef natura no o sthe e yar l strength—the exceptions being No. 8, the Tim; No. 9, Dundee Law—the last of which has perhaps the feeblest claim of all to be called a mote. Plans.—No. 1, Cairn Belli, the best claimant, appears to have been a typical mote with a base court. No. 11, Arnieful, and No. 12, Ganter- land, are simple moated mounds on paper, but the resemblance is not so meano n e y Torr,b s Th s unliki satisfactor , 8 . e a efield th No n . i y terraced mote. As to No. 2, InchbraJfie, if not a mote, it is certainly difficult to cal t anythini l e moundth e s surroundingth ha , else t I . g trench, and an outer rampart, but its unusual size and the feebleness of the rampart are contra-iiidications. No. 10, Castleton, and No. 11, Dundee Law, are possible examples of the square type of mote, ramparted and FORTS, 'GAMPS,' ETC., OK PERTH, FORFAU, AND KINCARDINE. Ill

trenched. The four remaining were possibly motes of the simplest type, mere artificial mounds without rampart or trench, always of necessity an uncertain class. The Name District.—e Moteth n i Its veri y doubtfu e namth s f eha i l been applied locally to any of these works, even to Nos. 3, 4, and 5, and I failed to get any evidence of it in the neighbourhood. In the Regis- trum Magnre Sigillx I have only noticed one reference to a mote in the three counties—A.D. 1546, "lie Moit dt knoo no Errole. wf o i d 1 " t non remaine markebu y e O.Mar , ean ther it th e f n .do o sar e Three ' motehill sn jVlearn i ' o ine als tw ar Pertos d mentionedan h e th t bu , signification is probably different from that of ' mote.'

(b) Earthen Forts. Number.—The total number of apparently earthen forts in the district is thirteen, including two described in the postscript. Elevation above e Sea.—th As wit e motesth h e elevatioth , s veri n y moderate, with the exception of ~No. 17, Evelialc, which attains 890 ft., a very unusual heigh r anfo t earthwork , howeveris deepla t I t . no ,y trenche t rathedbu terracea r d fort. The next highest barely passee sth 500 contour line, and only two others reach the 400 contour. Sites.—It is one distinguishing feature between this sub-division and the motes that, wherea e mosth l t al sprobabl e wean moteo e k ar ssites , e fortth s without exceptio e placear n d where they derive considerable strength fro e naturgrounde mth th f eo . Plans.—Where constructore th a fre d e ha shand e forth ,m approaches e oval t th a singlthaonl bu f n ,o i yt e instanc e ovath ls i eenceint f o e fortification complete positioe Th . n being generall straighe th t ya t edge a osteef p bank, whic s lefwa ht unfortified, compelle e forth do mt assum somewhaea t semi-ova lcase figure th f fortifiee o n I . d promontories, inlan coaste th d n o,o r wher e fortificatioeth confines i n tho t de necke th , e forforth tf mo depend s entirel e cut-of than th y o f o tf point. The only complete fortified enceinte is at Green Castle, IS~o. 23. It is single, consistin rampara f go trenchd an t f uniforo , m strengt l roundhal , 112 PROCEEDINGS Oi' THE SOCIETY, DECEMBER 13, 1899.

and this assimilate e stonth eo t fortst i s . Bu sigo I n tcoulf no e se d stone about it, and the fine broad rampart was amply accounted for by the deep and wide trench from which it was no doubt taken. In three instances f cut-ofo l al ,f points fortifiee th , d lins onlewa y n singlei t bu , othee th r example ordinare th f thal so al ty n typei fortd an ,s with incom- plete semi-oval fortifications, the lines were double or treble. Terraced fortification seems onl havo yt e been employe t Eveliclc,da No. 17 . Water Supply.— e eartheth Most f o n forts motese wels th a d ,s a lha , a pretty direct acces o streamt s s running close under their sitest bu , whether they had springs or cisterns inside is uncertain, as not a single native earthen fortress in the district has been excavated. Relics.—For the reason just given, we are equally ignorant of the relics they may contain. ImportDistinctionsthe of between EarthenStonethe the and Forts.— Although the earthen differ from the stone fortresses in the lowness of their situation, their structure, and their generally incomplete of fortification, it would be rash to conclude that these distinctions depen differenca n do e builders f f racth dato o e r n ei eo . l al The n yca be explained by the differences natural to sites on the hills or in the ease th evalleys y witb d h an ,whic h stoneworke th throwe b n o n p sca nu former, and earthworks in the latter.

IT. STONE FOBTS. Number.—The numbe f stono r e forts, reckonin sucs g a l thahal t have the citadel of stone, whatever the nature of the outer defences may be, is twenty-three—but few for so large a district. Elevation above e Sea.—th The great comparative elevatio f stono n e forts is shown by the facts that only one of the twenty-two comes down to contour0 50 e th , whereas.onl e twenty-fivy th fou f o r e earthworks come up to it; and that none of the latter come within 100 ft. of the 1000 contour e stonth , f whileo fortfivd x e abovesi ean ar s mor, eit e within f reachine o highes . Th ft f all. n 0 o More,gtit ,10 Du Gfenalmond,s i no less than 152 . abovft 0 e the sea. FORTS, 'CAMPS,' ETC., OF PERTH, FORFAR, AND KINCARDINE. 113

Sites.—"With the exception of No. 27, Dun Geal, which, although on heighta approachabls i , easy b e y slope l roundstone al sth l e al ,fort e ar s strongly r moro protecte e e on side n precipicesy o db s , ravines r steeo , p slopes. Plans.—The great majority are oval in form, although in some the broao o approacovat s s s i la d e circleth h . Onla few n i y, whice ar h chiefly tower-like structures e forth ms i , circula r vero r y nearl. yso In seventeen of the twenty-three stone forts the artificial zone of fortificatio a complet f s uninterruptei o n e on e d stone wall, often with diminutio0 littl11 r s thickneso eit f o n s wher e naturath e l defencs i e strong, greatly contrastin n thesgi e respects wit earthee hth n fortsn I . one, No. 25, Dundurn, the enceinte is also complete, but this is effected y drawinb a gserie f stono s e walls irregularly froe inaccessiblon m e precipic anothero et , Rossie Non 33 , OgleI . .31 . No Law,Hill, d e an th fortified zone is also complete, but a great part seems to have consisted not of stone walling, but of terracing. In No. 32, Ben Ejfery, the same double system appears to have been used, but one side, a straight precipice edge, was unfortified, and this with No. 24, Dunmore, which a semi-ova has l fron fortificatioof t n fromin- edg an edgto e of e accessibl eonle bankth ye instancear , incompletn a f so e artificial enceinte e twenty-threeth n i , wite possiblth h e, whic exceptio36 . s i h No f no so dilapidated that no opinion can be formed about it. Development of the Fortification.—In four, Nos. 27, 29, 33, and 41, the enceinte is single and simple. No. 44, Finavon, has in addition a mound ende on ,t a doubtfully artificial t connectebu , earthen a y db n ramp with the main work. In three, Nos. 26, 30, and 35, there is an annex at one end. In No. 37, Carnac, the annex goes so far round as almost to form a seconMacTual,n Du , d 28 enceintea secon n i d d an , enceint s alsi e o nearly completed by an annex at one end, and advanced stone walls at the other. In Dron, No. 38, two large annexes give a false appearance of a double enceinte, but they are merely slightly fenced. In the remaining eleven, excluding the too dilapidated No. 36, the enceinte is either doubl trebler eo , sometimes partially quadruple. VOL. XXXIV. H 114 PROCEEDINGS OF THE SOCIETY, DECEMBER 1), 1899.

Structure of the Walla.—This has been determined with or without excavation in ten of the twenty-two, and in all these it is of the same type, the faces bein f well-builgo masonryy e cordr tth f rudel o e d an y, built rubble work. Two—No , Castle , Abernethy—34 . 35 . Law,No d an have wooded ha ne Scottis beamth l al n additioni sh n i ston s A e.• forts that have been investigated, there o prepareappearn e b o st d foundationd an , when the wall, as generally happens, stands on the edge of a descent, the outer face is lower—sometimes much lower—than the inner one. There seems little reason to doubt that the Avails of the other stone forts of the district are simply buried under the ruins of their upper part or under the accumulated overgrowth of soil and vegetation of centuries. Finavon, a muc s a h , Novitrifie44 . d worka possibl s i , e exception, although Dr G. A. Jaudeson, who saw sections made through the enceinte, declares in his rather incomprehensible description that it was a built wall buttressed with vitrifaction. Water Supply.—From their general sites 011 pointed heights at con- siderable elevation stone sth e fort e reads th hav t y eno acces o streamt s s whic e eartheth h n forts often possess. Neithe e springar r s generally found near the t mpresena th1 e01 t hillsides. s Abernethy,ha , 35 . No an exceptionally good, ready supply in the dammed-up loch close below it, and also had a large rock-cut cistern inside. Hollows in the interior of several, as at Denoon, No. 41, Barm, No. 42, Finavon, No. 44, and the White Caterthun, No. 45, may indicate the existence of old wells or cisterns, but in most of the areas there is nothing of the kind. Relics.—The numbe f relico r s foun bees ha dn remarkably smallt bu , only two of the forts have been scientifically excavated. The result, as1 far as it goes, is that the rinds are compatible with the existence of the forts about the Roman period, but not necessarily earlier, and that the range is considerably into mediseval times. Broclis.—Several circular foundations of Broch size remain, but it is very doubtful if any^rcally were Broclis. One that I have seen, Robs Reed, has been included in. my class of dubious works, and it is not constructed of the regular masonry required for a Broch; the FORTS, 'CAMPS,' ETC., OF PERTH, FORFAE, AND KINCARDINE. 115

e saidb samI e y 'Towers,think th ma ef o , n Glenlyoni ' s figurea , d by Miss Maclagan, and described to me by Dr Joseph Anderson. The circular work at Tyndun, No. 29, is too large for a Broch, and if a lesthi s si s difficulty n witGeal, e Du hwidt th f , o Noh 27 . its wall is scarcely sufficient. In all respects the most likely Broch foundation is the excavated circular Avail 18 ft. broad, with an area of 36 ft. enclosed, at The L>uos, No. 39, but in the absence of evidence of a contained stair or chambers it cannot be proved to be the wall of a Broch. Vitrified Forts.— It cannot be said that the authorities for vitrified districe fortth n si t have done anythin cleao gt p thu re obscurities that haunt this class of objects. Rather have they done much to illustrate the extraordinary discrepancies as to mere facts that meet the inquirer into their history at every turn. Take The Laws, No. 39, and we find Dr John Janiiesoii describing the two walls and all the immense mass of building withi thoroughls na y vitrified, wherea Neisr sM h asserts there wa vitrificatioo n s e walth ln i facesn e absenceth , f whico hI verified , arid that vitrified masses were only backine founth n d i r rubbl e go th f eo wall. Or take Dunsinnan, No. 40, where Playfair makes no mention of vitrifaction, while Stewart asserts that the whole interior buildings were alss thaod wale wa uset an th i t l, n dit facefulo f o l . The fort best entitled to be called vitrified is Finavon, No. 44, but only secone th n i d degree t cannoI . comparee b t d with Carradale, Argylen i , f continuouo . ft 0 wit7 s shit vitrifaction standin walla s a , p stilgu l less wite Arisaith h g forEileann o t n Gobhar,na recently verifier D y b d Rober e plac9 ft. a heigha continuou on ,f o to t e n , i Munrt be o t os vitrified wall wherever it is visible, and comparatively little of it is overgrown. Finavon rather ranks with Tap o' JVoth, Aberdeen, where ther mucs ei h vitrifaction e latr Macdonaldth e M s a t bu , , Huntly Farm, found, bearing but a small proportion to the un vitrified stone, and where in one of his two complete sections no vitrifaction was found from top to bottom. Thialss case o swa section y th prove eb e b t Finavon.o da t s Thes forto etw s differ, however e positio th vitrifactione th n i ,f no , which 6 11 PROCEEDING SOCIETYE TH F O S , DECEMBE , 189911 R . p o' Ta Noth,wa t a s y p certainl whereas to ma d nea e e an w th rt ya f i , believe Dr Jamieson, it buttressed a built wall. Lastly, it may be suggested under this head that the quantities of loose block f vitrifactioso n apparently Laws e foun Dunsinnand Th an t da y ma have been got from vitrified walls which preceded, on the same site, the unverified walls now existing.

DISTINCTION BETWEEN THE FORTS OF THIS AND OTHER DISTRICTS IN SCOTLAND. t woulI d tak lono emakto o g t comparisoea nothee witth l rh al group s Scotlandn i t confininbu , e comparisogth o Argylent e differenceth , e ar s very marked—perhaps more so than in any other case. One manifest dis- tinction is that one-half of the forts in our district are of earth, while it s i doubtfu a singl f i l e earthen fort exists in Argyle. Thiy be-ma s' e abundancth partiall o t e f stono edu y e everywherew lo n higi , r o h ground, in Argyle, together with the rarity of deep soil suitable for mak- g entrenchein d works t morBu .e marked distinction e largth oe ar s number and small size of the forts in Argyle, and the small number and larg r district e fortou e th sizn i sf eo . Whe s considerei t ni d that vast tract Argyln si e destitutear f fortseo , 171 seem slarga e numbee th r fo r occupied tracts compares a , d5 scattere 4 wit e hth d pretty generally over Perth, Angus e differenc th d Mearns an t , n Bu siz i es . i evee n more striking. Adopting the standards used in my work on Early Fortifica- tions Scotland,n i t comemeasurabli 4 t tha16 f sou o t e fort Argylen si 1 8 , are very small, 72 small, 10 considerable, and only 1 large, whereas in our three counties, of 44 measurable, 5 are very small, 13 small, 19 con- siderable and 7 large; or, dividing them into two classes as • large and e thre centr smallth pe e n i , 6 countie7 , r cent onld pe n Argyl i san ,y8 e are large, and 92 per cent, in Argyle and only 24 per cent, in the three counties are small. These contrasts derive some interese ar tt fro se face e mth t on tha e th t lane inth d first occupied in e Scotlanothee th e Scotsth th n d r i y dan b , chiee th sea f cannofe o t w powee Picts th f I tf o r.positivel y affirm that FORTS, ' CAMPS,' ETC., OF PERTH, FORFAU, AND KINCARDINE. 117

they are the forts of the Scots and Picts respectively, it is at least a very reasonable supposition. That our knowledge of the forts in the whole district is still so im- perfece mainlwane du th f excavation s o i o tyt . Onlo havtw ye been excavated scientifically, and two unscientifically, and the clear and instruc- tive results obtaine e firs th s encouragina t n casi de ar e o furthet g r scientific effort as the confusion and misconceptions that followed in the second should war againss nu t investigations undertaken with insufficient mean completo st e the r mwithouo t expert guidance t I seem. s strange that, while mean e founar s carro t d n scientifiyo c investigatione th t a s utmost ends of the earth, the hidden mysteries of our own country, so read o comparativel s t a hany d an d y easily unearthed, should remain neglected. But apparently neither the hour nor the man has come.

POSTSCRIPT. e accoun e Society'th th n f I o t s excavatio f Ardoco n s recordei t i h d that seven fortified works,' noticed by writers of the 18th century as existing in the neighbourhood of the ' Roman Road' that runs from Ardoch to Strageath and beyond Gask, had disappeared, and could not even be accurately localised. Two of them, however, have been quite recently rediscovered by Mr Alexander Mackie, and have been exca- vated under his supervision by permission of Mr Samuel Smith, M.P., anr KingtodM n Oliphant proprietorse ,th . Unfortunately discovere th , y was so recent that neither could the forts be marked on my map, nor coul e descriptionsth d , wit e plath h n r MackitakeM y b ne (fig. 56)e b , inserte thein di r prope rnarrative e placth n ei . OreMll, No47 . Fort, situates i d about \ e house mth th . f JSTf eo o . same name withid . abovan ,e ft policy th n0 ee woode60 seath , n ,o d Muir of Orchill, in the sharp angle of junction of two trifling rills. From grouninteriothe the S. dto rrisethe ft. to s 50 steepl one to the y40 n thiforte oO th sf . side, further strengthene marsha y db artificiao n , l defence was deemed necessary, at least no evidence of such was found at the edge of the bank ; but the comparatively weak northern face has 118 PUOCEEDINGS OF THE SOCIETY, DECEMBEK 11, 1899.

been fortifie semiovaa y b d l lin f earthworkso e , comprising, where fully developed . higha ratheft a scar ) ) 0 (2 1 p;(1 r , angled trencha ) (3 ; widemoun5 roundea 2 . hig ) d ft (4 ;dh5 an . ddeep ft trenchf 7 o , o t 5 , C __

I . S" HJ a. b-

. FigEarthe56 . n Fort, Orchill.

whic counterscare hth rathes pi r higher tha scarpe mouna nth ) (5 ; d rising f theso carrie e l e ar . al abov ft t dexteriore 4 no eth r roun t o e 3 Bu dth . whole fron f fortificationo t . The s westere founit yar n o d n half, where FORTS, 'CAMPS,' ETC. F PERTHO , , FORFAR KINCARDINED AN , . 1.1.9

the interior has a command northward of only a few feet, but eastward exterioe th r ground falls away d than e, slope froe interioth m steeps i r , and here only the upper scarp (1) and its trench (2), much reduced in t with t thA me esize .e easterar , n sharply-pointe f tho ed fortden , where the narrow front could only hold a few defenders, the first mound (3), and its trench (4), again appear, in front of (1) and (2), to strengthen this weaker point. On continuing the transverse section through the inner area, two small trenches, a and /;, about 18 in. deep, were discovered, curving round parallel wit e earthwork'shth f tho e p to e , th onrean i f e o . rabou ft 0 1 t scarp, the other from about 12 to 18 ft. in rear of the first. They con- tained flat stones or flags, generally disarranged, but in some places still standing on end, so as to line the sides of the trenches, e in the enlarged section, leaving a space about 8 in. wide, which was filled with earth, containing decayed or charred wood in small fragments but in large quantity. It seems a fair conjecture that these trenches held , and that the stone linings were intende o assist d n supportini t d gan fixing them. In some places, the same woody earth was found beneath palisadee ende th e stonesth f o th sf d i bee s ha a ,n fixed int a ofounda -

tion-beam. Aor . 48, Kempy, Gask.—This fort, much levelled by the plough, so closely resemble e lasth st s unnecessari tha t i t givo t y. it a epla f o n positioe Th 1STf y Findogass \\no i b . . mW . k Church . abovft 0 e20 ,sea , on Kempy Knoll, from which there is a steep descent eastward to a rill, tributar f Cowgaso y k Burn. This naturally strong sid s unfortifiedi e , semi-avaa t bu l double entrenchment, wit hsingla e small 'palisade trench,' exactly like thos jSTon , protecteo 47 . N.We sth . front, whic accessibls hi e a gentle y b ascent s approachedi e S.\V d Th .en . a narro y b w ridge, nearly on a level with the interior, and here the abrupt end of the trenches, before reaching the edge of the descent to the rill, shows where the entrance had been. The dimensions over all are 385 by 215 ft., and of the interior 250 by 165 ft. The inner trench is about 15 ft. wide and 7 deep, and the outer one 8 ft. wide and 3 to 4 deep; but 120 PROCEEDINGS OF THE SOCIETY, DECEMBER 11, 1899. these dimensions would be greater wliun the plonghed-down rampart, now 17 ft. wide OH the top, retained its full height. The front of fortification is 48 ft. broad, including the ' palisade trench,' but narrows to 34 ft. at the entrance. The ' palisade trench' is there carried further than the defensive trendies t containi d an , s n edgestoneo t , sse diarre d woodd an , black mould, like the similar trenches of No. 47. As palisade trendies have not been previously noticed in Scotland, I am fortunate in being able to conclude my account of the forts of South Pictland wit a hrecoro interestin s f o d d nove an gn observation a l n i ; forts, moreover, which have been rediscovered, after escaping notice for more tha a ncentury o potterN . r otheo y ry kin relican d f o swer e found in either of them.