0 3 PROCEEDING E SOCIETYTH E O S , DECEMBE , 19128 R .

III. THE CIRCULAR FORTS OP NORTH PERTHSHIRE. . J WATSON . W Y B , • M.A., LL.D., F.R.S.E., F.S.A. SCOT.

0 remarkablThe e fortifications which for subjecmthe thiof t s paper occur, so far as my observations and inquiries have hitherto disclosed them Glen i , n Lyon, Loch Tayside, Upper Strath Tay, Strath Tummel, neighbourhoothe Pitlochriedof , Glenquaich Strathbraanand , The . distance instraigha t line fro e farthesmth t wes heae t th f for do t a t Glen Lyo Pitlochrio nt s aboui e t forty'miles, whil distance eth a n i e straight line north and south from the forts on Loch Tummel-side to the furthest, south of the Amulree group is about fourteen miles. Within that area I personally saw eighteen circular forts, while I have reliable authorit e existence th th r t f leasya fo o f morex e eo ^ si On . Glen Lyon grou s beepha n completely effaced within living memory, neae on r d Pitlochriean , mentione Pennanty b d , seem havo st e also disappeared totae Th l .number , therefore whicf ,o h ther remaine ear s at present, more or less distinct, is twenty-four. There is, however, reaso o believt n e that additional specimen e e founth b n y i d ma s neighbourhoo f Blai Statisticaldo d Ol r Athole th r Accountfo , mentions Strathgroyonn i Glen ei o ntw Tild an t . It is also possible that still more might be discovered by an examina- tion of likely situations in such places as Glen Almond and Glen Lednock, between Loch Tay and Strathearn. The basin of the Tummel, too, may contain more tha nI discovered particulan i , r about Kinloch-Rannoc d Fincastlean h , neithe f whico r n ha placed ha I s opportunity of investigating. The first direct mention of these fortifications, so far as I am aware, occur Pennant'n si s four. Pennant himself appear havo st e seen three e Pitlochrioth f e group frod e Revan m,th . James Stewar f Killio t n THE CIRCULAR FORTS OF NORTH PERTHSHIRE. 31

he got and gives information as to the dimensions and names of the forts in Upper Glen Lyon, at Fortingal, and on the south side of Loch Tay. Pennant, however n somi , e cases (Killiechangien Du d an , Mac Tuathai n Drummono l d Hill) confuse e circulath s r forts with hill forta differen f o s t typeappeare H . confuso st remarkable eth e terrace deminencn a sit n eo e south-eas e farmhousth f o t f Kerrowo e - mor Glen ei n Lyon wit roune hth righe d th for n t o tbankLyo e th f nso near the foot of that eminence ; and, lastly, the place he prints as Fiam- nam-bbinean, by which is apparently meant the fortified hillock on the left bank of Pobul Burn, and which is marked on the modern maps a"a s tower," doe t belonclase no s th f structure o so gt con e ar -e w s sidering. Statisticalw Ne writee e th Th f o rAccount e paristh f f Fortingao ho l states that fiftee sixteer no n such structures existed withi bounde nth s of the parish at the time of writing (1838). Mr Duncan Campbell, Inverness nativa , f Gleeo n Lyonboos hi kn i , entitled The Chronicles of Garth and Fortingal, locates the Glen Lyon towers as far east as Fortingal, and gives the Glen Lyon tradition regarding them. There are, however passageo tw , writern si muca f so h earlier date, t withouno e b whic ty somma h e relevance e firs f theso Th t .s i e from Gildas (ob. circ. 570)writeo wh , s: " De artissimis foraminum caverniculis fusci 'vermiculorum cunei, tetri Scottorum Pictorumque greges " (De Excidio et Conquestu Britannice, c. 19 ; p. 35 of Mommsen's edition); " Swarthy columns of vermin from their small caverns of very narrow outlets, loathsome hordes of Scots and Picts." The other passag quotee on y Skehs i eb d e (Celtic Scotland, . 261i ) from Eddi' st Wilfrid S Lif f o e , written abou A.D0 70 t. " Populi bestiales Pictorum feroci animo subjectionem Saxonum despiciebant, et jugum servitutis a se abjicere minabantur, congregantes undique de utribus et folliculis Aquilonis innumeras gentes, quasi formicarum greges in aestate de tumulis ferventes, aggerem contra domum caden- 32 PEOCEEDING-S OF THE SOCIETY, DECEMBER 9, 1912.

tern muniebant." beastle "Th y tribePicte th f sso bega warlikn ni e mood to despise the Saxon rule, and threatened to cast from off them the yoke of servitude, gathering together from all quarters countless tribes from, the bags and sacks of the North, like hordes of ants swarming in summer from their mounds begad an , builo nt a bulward k against fallina g house." Skene omit translato st curioue eth s expression de utribus et folliculis Aquilonis. The term cavernaculte of the passage from Gildas migh regardee b t a litera s da l translatio e Gaelith f no c uamhag, littla e cave, regularly use denoto dt e either earth-houser o s e chamberth e f seconbrochso sth n I d . passage, uter, a ski n bag, appears to be used figuratively and contemptuously for a circular buildin e brocth f gho type, while folliculus, sacka a bladde, r (much smaller than uter) may refer either to earth-houses or to the chambers of brochs or broch-like structures.1 The passages are curiously suggestive of the Fir Bolg, the men of the bags, who figure so pro- minently in early Irish history, but, so far as I know, the term bolg is never used among us in connection with brochs or earth-houses. At the present day, these forts are termed individually caisteal, castl Duncar e M (or s a , n Campbell. Inverness, says, castal), collecd an - tively caisteilean nam Fiann, castleFianne th f so . Ther widela s ei y known saying earliese th , t notic whicf eo h occur Pennantn st i go o wh , it doubtless from the Rev. J. Stewart : "Bh chaisteia ad l deu Fionng gai , An Cromghleann na n dub clach.n hna "

"Twelve castle Monnd sha , dare th k n Bent-gleI nstones.e oth f "

Here Cromghleann alwayss i Scotlandn i , ",a take kennine b o nt g " for Glen Lyon. This distich whico t , hI shal l refer later, condensee sth traditional accoun e raisonth f o t e fortsd'etreth f .o Tha e terth t m

1 Mr A. 0. Curie suggests that the reference in cavernaculcK may be to hut- circles, which ofte lond nha g narrow openings. THE CIRCULAR FORTS OF NORTH' PERTHSHIRE. 33

caisteal, thus a s applied olds i , s proveface i , th t y thadb t occuri t n si the place-names Pitcastle of Tullypowrie and Pitcastle near Pitlochrie, both named after forts still quite distinct l place-nameAl . s that involve Pit, a loan from Pictish, are old. The Pitcastles are now calle Gaelin di c Bail' a' Chaisteil, Castletown. Late , howeveron r I , shall give reason thinkinr sfo g that these fortoriginalld ha s ynativa e name much older than caisteal, which is borrowed into Gaelic from Latin castellum. t leas a instancee n on tI , Caisteal Dubh Baile Ceard,n na Blace th k Castle of Balnaguard, the term " castle " is applied to a large hill- fort wit hstona e Norte th wallsometime e n hw I . s apply caistealo t brochs, e.g. Caisteal na Crbice in Glencasley, Sutherland; Castle Spynie, in the Beauly district; Caisteal Grwgaig, Loch Duich. More often, however, we call them dim. The Norsemen called the brochg borgr.1 In the remarks that follow, I shall begin with the Glen Lyon forts, and proceed thereafter to deal with those that exist outside the glen, I hav s a er sascertaineofa d them. Referenc larga o et e scale map, suc s Bartholomew'ha e miles2 th o st inch (sheet 12), will sho e rangewth f hillo s s that divide Glen Lyon from Loch Tay and Glen Lochay on the south and from Rannoch northe onth , e wes e glewhile heath th sees th nti n f deo o sprino n t g froe watersheth m d between Pert d Argyllan h . These mountain ranges, north, south, and west, are pierced by a number of passes, which, in the days before roads, were regular and important means of communication e hea f GleTh o d. e enterenb Lyoy dma n froe mth west, either through Glen Orchy or through Glen Dochart. In the latter case one would strike northwards at Tyndrum, but the two •routes converge near the head of the glen. 1. Beinn a' Chaisteil.—Here rises Beinn a' Chaisteil, on a spur of which stood the most westerly of the Glen Lyon forts. This is the See, however, Loch Kanaird, Place-Names of Boss Cromarty,d an . 256p . 1 3 VOL. XLVII. 34 PROCEEDING E SOCIETYTH F O S , DECEMBER , 19129 , .

onlwhice y on practicall he havI sai b s seen i t o dt t eI no . y obliterated by landslides, and this, together with its very remote position, in- volving a two days' j ourney, discouraged me from visiting it. About eight miles down the glen we come to the remarkable group of towers between Pubi Cashlied an l . e spacHerth littlf n eo ei e more thana mil have remaine w eth fourf so , with traditio fiftha I f visiteno . d these (fothire th r d time 25tn )o h August 1912. S compan n . i , E r M f yo

Pig . Pla1 . n of Cambuslai Fort.

M'lntosh t e MasteRoyath Ar , f o lr High School whoo t m , a mI indebte e measurementth r dfo pland san s submitted M'lntosh'r M . s plans are here reproduced on a, scale of 30 feet to an inch. They are inserted so that the top of the page indicates north. 2. Gambuslai.—The furthest wes f thio t s groupe (figth lie) 1 o .t s south of the public road, in the S.E. corner of a field, and quite close to S.Ee Lyonth e circl.e n th sidO th f .e o there appea tracery fean wf i s of either the inside or the outside wall-face. Otherwise the outer face THE CIRCULAR FORTS OP NORTH PERTHSHIRE. 35

is distinct. Both faces are visible on the northern half. The founda- tions of two oblong structures, perhaps of secondary origin, and at right angles to each other, are seen within the circle. The remains of the tower consis greaf to t block adjoinine stonef th so d an , g dyke appears to have been built out of its materials. We made the greatest external diameter 73 feet. Its wall varied from a little over 10 feet to 13 feet in thickness whole Th exceedingls .ei y dilapidated.

Fig. 2. Plan of Castal'an Dui Port.

. Gastal3 Dui.—n a About quartea east e jusmild a th f o an ,t o ro e t e publinorte th th f ho ce ruin th road f whae o s ar , t Pennane th n (o t authorit f Revyo . James Stewar f Killino t ) calls i (figCastaDu . 2)n la . It was apparently the most complete of the group in Pennant's time (1769)relativels i d an , foundatio e stillo ys Th .greate th f no , circular wall is traceable all round with approximate completeness, and on the north and north-east some courses of the cyclopean drystone masonry are in position. The most remarkable feature is the great boulder, 6 •3 PROCEEDING E SOCIETYTH F O S . DECEMBE , 19129 R .

20 feet 8 inches long, 10 feet 6 inches high, and 11 feet thick, which f orme insid e outsidee norte dth walth e th e parn n th eh o th o O lt f to . builders erected a wall about 8 feet thick or, in-places, more,'and I suggest that their reason for doing so was that they were merely incorporating this huge block in.a wall which, when complete, was to rise considerably abov s higheseit t pointsituatioe Th e doo.th f rno uncertains i consideree w t bu , d the^ likelies southte jwsitiodu e .b o nt

Fig. 3. Plan of Caisteal a' Chonbhacain. distinca fore s flooe th ha t f Th o tr slope southwards. Ther smala s ei l Lyoe th moras reare s perhapni d th an ,n yard i 0 s s20 s distante Th . diameter and thickness of wall (except at the point where the great stone comes in) are very similar to those of No. 2. 4. Caisteal a' Chonbhacain.—The next of the group (fig. 3) stands a little more tha nquartea mila f eo r eastwardssoute th n ho sidd f eo an , the road, immediately within the roadside wall. It is known as Caisteal a' Chonbhacain, from a remarkable stone in its vicinity, which was till E CIRCULATH R FORT P NORTO S H PERTHSHIRE7 3 .

recent times practically an idol. This fort, as will be seen from the plan, is decidedly oval in shape. In Pennant's time the door was still extan ease narrowth d tt t "an enda w ,lo covered wit hflag.a " positios It stil y discernede lb n ma thicknese th e wale d th th t an ,a l f so door is a little over 11 feet. Inside the fort at the east end there are remaine th wala f sfeeo l4 t thick running from nort southo ht , forming completa e chord, excep breaa r fo tk immediatel doore fronn yth i f o .t Whether this wal secondars i l originae parr th yo f o t l structuro d eI not venture to suggest. The wall-faces are well denned almost all e S.Wroundth t s best.i bu , , showing some coursee originath f o s l masonry. The stones between the two wall-faces are rather small, the building presenting the appearance of two strong walls with stones throw promiscuouslyn ni I dra. w attentio curioue th o nt s " eken "i the masonry on the S.W. side, which is undoubtedly structural and not accidental structuree th l al .f o onl e sI havThie th yon s si e examined which show decidesa d divergence fro circlee mdimensions th It . : e sar longest axis fee5 9 ,inche 6 t s; wall fee8 , inche 6 tfee 4 inches1 C t o st . . Cashlie.5 —The fourte Cashlith f o h eyard w towerfe sa souts si h of the road, right in front of Cashlie farmhouse, now a shooting lodge. Thoug ha quantit f largyo e stones mark e sitee structurth s th , s eha bee badlo s n y knocked about tha foune w t t i impossibld o takt e e measurements sufficient for a plan. It was, however, apparently not circular, but rather oval. Its walls appeared to vary from about 9 fee2 1 inchefee6 t o t t thicknessn si larges wa t rI tha.. 3 t d Nosan 2 . The fifth of the Cashlie group is said to have been on the south side of the river, a little to the east of Dalchiaiiich farmhouse. There may be some remains of it, but I failed to see any, and Mr M'Kerchar, the tenant of Dalchiarlich, who was born there and has lived for over sixty years between Dalchiarlich and Cashlie, half a mile away, could helpo n giv e .em Between Cashli d Bridgan e f Balgieo e a distanc, f nearlo e y eight miles, no remains of forts appear, but I had some faint tradition of one 38 PROCEEDINGS OF THE SOCIETY, DECEMBER 9, 1912.

having been at Gealainn, where the house of the factor for Meggernie is now, about a mile to the east of the River Conait, a left-bank tributary Lyoe oth f n flowing from Loch Dam Locd han h Giorra thiy y B swa . ran a pass called Lairig Meachdainn, leading to Rannoch, coming out at Camghouran, on the south side of Loch Rannoch. There is no other pass either to the north or to the south between Cashlie and Bridge of Balgie. Bridg f Balgio e s twelvi e e miles from Fortingalo s r o , y halwa f between Fortingal and Pubil. It is an old and handsome bridge of great strength, spannin Lyoe gth n just above Linne Bhalgaigh greata , . deep pool, circular or bag-shaped, as its name indicates. Here the hill road from Edramucky on Loch Tay-side joins the main Glen Lyon road, coming from the south by a gap called Lairig an Lochain or Pass e Lochleto th fmile o Tw s . from Bridg f Balgieo e another important pass branche south-westerla n i f sof y direction. This pas knows si s na Lairig Bhreislich, and it comes out at Dunchroisg in Glen Lochay, five miles above Killin. 6. Just belo junctioe wth thesf no passeso egreetw a n o ,n terracn eo the left bank of Allt a\Mhuilinn, are the sorely battered remnants of. another tower. The terrace was in recent times the site of a homestead, the buildings of which and the remains of the tower are jumbled up in such confusion that I refrain from giving measurements. Glen Lyon tradition has it that a Menteith man leading a raid on Glen Lyon was slain by an arrow from this tower. The place where he fell is still known as Dail an Teadhaich, the Menteith man's 'dale. I had this from Mr J. MacDiarmid, Dunruchan, Muthil. Half a mile east of Bridge of Balgie is Innerwick, where a market called an Fheill Muice, the Swine Fair, was wont to be held. North- ward therefrom runs Lairig Muice, e Swinth e Pass, makin r Balgfo l e soutonth h sid f Loco e h Rannoch d branchinan ; e th g o frot t mi right, about a mile north of Innerwick, is Lairig Chalbha, coming out t Cariea mileo tw , s eas Ballf o t . E CIRCULATH R FORTS OF NORTH PERTHSHIRE9 3 .

. Kerrowclach.—7 Just opposite Innerwick e soutth n ho , sidf o e Lyo adjacen d towee n an th sit e forda f Kerrowclachf th eo r o o s t ti , , on the farm of Kerrowmore. Its stones were used, says Mr Duncan Campbell, to form an embankment for the Lyon at this point. The approximate measurement e toow s k shoo havt t wi e been nearly circular, wit a largha wal s f aboueo lwa fortfeet 2 t 1 I t ,. witn ha external diameter of about 86 feet. A quarter of a mile to the west of it is the ancient graveyard of Brenno, with the heptagonal font of the Kirk of Brenno. All trace of the church is gone. 8. Camas Bhracuinn.—About two miles east of Innerwick, the U.F. Mans t Camaea s Bhracuinn occupie towea site f sth eo r whic hI hav e not seen, but of which the foundations existed within living memory. Thentrance th e t subsidiara spoa o s et i t y pass which joins Lairig Chalbha above mentioned. 9. Roro.—A little to the east, but on the south side of the river, and less than half a mile west of the farmhouse of Eoro, on the flat, and touching the south side of the cart road, are the remains of another tower (fig. 4). The outer wall-face is fairly distinct all round, but the founinnes i e dron with certainty onl t intervalsya peculiaritA . y of this fort is that its floor is quite 2 feet below the level of the surrounding sur- face. Withi e broken-dowth e fore nth ar t n wall buildina f o s g which •was use schoola s da , constructed materiale th coursef ,o f e o th t f s,ou o fort. Mr Cameron, the tenant of Roro, told me that his father had attended that school. Though stone e mosth f o ts have been removed, there still remain enough of great blocks to make the site impressive. A wes e littlsparkline th tth o e t g Allt Caor comes down fro skirte mth s of Ben Lawers, through a pass which runs by the western flank of the mountain to Carie of Loch Tay. The Roro fort is nearly circular, wit externan ha l diamete feet0 wald 8 f abouf an ,o lo 4 r 1 fee1 1 to t t feet in thickness. 10. Cam Ban.—The next site is on the north side of Lyon, about five miles east of Camas Bhracuinn, on the farm of Balentyre. The fort 0 4 PROCEEDING E SOCIETYTH F O S , DECEMBE , 19129 R .

(fig. 5) stood on the very edge of the ancient lake terrace, which runs here at a little distance north of the public road. It is about a quarter oa milf ee farmhouswesth f o t f Balentyreo e e samth , whicn eo s i h terrace, and immediately behind the cottage which stands close to the public road. It is in the last stage of dilapidation, but neverthe- less the great foundation stones of its outer wall-face can be seen in

Fig. 4. Plan of Roro Fort.

situ for three-fourths of its circumference. The inner face has suffered so much tha coule t determinw t dno t witei h certaint pointy an t y.a s nearle forwa Th t y circular, wit n externaha l diamete f abouo r 7 8 t feet. The spot selected for its site is,a little to the east of the entrance to Lairig Bharra e pasth , s that goes from Inverva Cario t r Locn eo h Rannoch. About acrossy halwa f , another pass branche o tht f e of s right through Gleann Sasunn, the Glen of the Saxons, and comes out at Inverhadden near Kinloch-Rannoch. Here, on the north side of THE CIRCULAR i'ORTS OF NORTH PERTHSHIRE. 41 the Tummel from Inverhadden, is a place bearing the significant name ot Druim a' Chaisteil, Castle Ridge, but I have not visited this site. • At Chesthill, a mile or so below Balentyre, one enters the magnificent Pas Lyonf defile so Th quits ei .mile o etw s long afted an , r traversing the gorge the Lyon enters on the final part of its course along the level

Fortn Ba .Figm . Pla5 .Ca f no and peaceful plain of Fortingal, skirting as it goes the north side of Drummond Hill. 11. Dun Geal.—Here, about a mile to the east of the pass, a little N.B. of Fortingal Hotel, and immediately above the farmhouse of Balnacraig, stood a tower known locally as An Dun Geal, the White Fort (fig. 6). It is built near the edge of one of those precipitous rocky bluffs that characterise the north side of the valley from Fortingal to Killiechassie. The eminence on which it stands is about 400 feet 2 4 PROCEEDING E SOCIETYTH F O S , DECEMBE , 19129 R .

abov plaie levee th e th f nlo belowcommandes i d an , considerabla y db y higher eminenc vieN.We e th wTh o e.frot mt extendi S.Ee th o t . n so Aberfeld littla beyond y shore d S:We eth wa yth an o tn . t paso ; s from Fearnan to Fortingal. This fort is the only one of the circular towers, so far as I have observed, that makes any pretension to strength of site. mose th r t fa impressivs als i y ot b I e Gleth nf eo Lyo n groups it r fo ,

GealFign . Pla6 Du . .f no (No. 11.)

circl complets ei roundl mosn ei al d t an part, innee outesd th ran r wall- face seee ar sn clearly, wit he masonr th herfee therd f w o tefe an eya (figs. 7-10) s comparativelIt . y good stat preservatiof eo partlns i yo t be ascribed to its position, which is difficult of access for carts. Never- theless an aged mason of Fortingal assured me that he himself had suc- ceeded in taking an enormous quantity of stones out of it. The Dun Geal has within it the foundations of three or four structures, the largest of

. THE CIRCULAR FORT NORTP O S H PERTHSHIRE5 4 .

whic rathes hi r oval abuttind ,for en e wit e th t n hon wallgo othere th , s small and circular. Immediately to the east of it are the foundations of three very small circular structures. The narrow door faced west- e soutwards th e lin f he entranceo Th sidth . f eo e seems quite clear; northere th entranccleare o s nTh t .sid no es ei seem havo st e beena little over 3 feet wide. From about the middle of the southern part circumferencee oth f foundatioe th , strona f no g wall runso t righ t ou t the precipice, and from a point on the north side directly opposite this, a similar wall runs N.E. toward smalsa l marsh eminence nortth f ho e on which the fort stands. Just eastwards of the fort is another slight eminence , s easterwithit n o , walnd ol sidel n runnina , ge righth o t t precipice on the south, and northwards into and through the marsh, d beyon t havan no . ey dit Thes beema r ne e o parth wall y f o t ma s defences. The external diameter of the fort is a little over 87 feet : its wall varies in thickness from about 8 feet to 14 feet. 12. Litigan.—The next and last tower in the valley of the Lyon stood smala n o l eminenc farmhouse e N.Wlittlea th th f o o et . f Loweo e r Litigan. on the right cheek of Keltney Burn. The ruin stands in a cultivated field, and, indeed, part of the fort itself (fig. 11) is cultivated. Notwithstanding this e remainth , s mak efaia r sho two-thirdr wfo f so circumference th S.W.e th n o , however e; , every vestig gons ei e except at one point. The site is fairly elevated, and commands a good view up the pass through which runs the public road from Coshieville to Kinloch-Kannoc Stratd han h Tummel, e S.Ewhilth o .et lie s Appif no Dull. Directl e otheth n e Lyoryo th sid, soutf it no e f valleyo h , and on a precipitous spur of Drummond Hill overlooking Strath Tay to the east as far .as the Braes of Tullymet, is the large rude hill fort of Dun Mac Tuathail. The fort of Litigan was rather oval than circular, with a greatest external diameter of 73 feet, and wall 9 to 12 feet thick. Pennant mention forto Locn stw o s h Tay norte th lie.e n hso " On side of Loch Tay, about five miles east of Killin, above the public 46 PROCEEDINGS OF THE SOCIETY, DECEMBER 9, 1912. road. The other, called Caisteal Baraora, about a quarter of a mile a fromeasure e laked mth an , d mile eas f Achmoreo t e seaf th ,o t Mr Campbel of Achalader." The site indicated for the former of these two forts is just at the starting point of the present public road from Loch Tay-side to Glen Lyon, by Lairig an Lochain. I got no tradition of a fort at this place, nor could I find any remains of one.

Fig. 11. Plan of Litigan Port. (No. 12.)

It is just possible that the information got by Pennant referred to fort No. 6, at the Glen Lyon end of the pass. . Remain13 Caistealf so Baraora stil ldistance existth t ,bu e from Loch Tay is understated. The name is Caisteal Braigh an Radhaire, " the castle at the upper part of the outfield," or as Mr J. MacDiarmid has it, " Castle of Brianraray." 1 The site is approached through thick birch wood, and is not over easy to find. Last summer I got within a

short distance of it, but had to turn back, as grouse were being Transactions1 of Gaelic Society of Inverness, XXTI . 141p . . Blaeu' p showma s s Bareyra, placed with approximate correctness. E CIBCXJLATH K FORT F NORTO S H PERTHSHIRE7 4 .

driven on the moor adjacent. Pennant's description of the great fort on the adjoining height may be supplemented by that given by Mr MacDiarmid.1 There may be remains of others on the south side of Loch Tay, but I know of none between this and Grandtully, half way between Aberfeldy and Ballinluig. There on e nortth h sid f Tayo e , right opposite Grandtully station, there ar e two close together, the Black Castle of Pitcastle and the Black Castle of Tullypowrie. 14. Pitcastle.—The former stands on an eminence amid the culti- vated ground on the farm of Pitcastle. It is somewhat overgrown with blackthorns, and has been so mishandled besides that I could t welno l measur . eit Ther doubto n s i e , howevere th f o , s thai t i t same type, general size, and thickness of wall as those already de- scribed. 15. Tullypowrie.—The other stands about a quarter of a mile to e eastth holloa , w with small stream intervening disuseA . d quarry, which lies about midway fortsbetweeo tw , e formth n conveniena s t and known landmark. The remains of the fort are on the farm of Middleton of Tullypowrie, at the foot of a cultivated field. They are much more considerable than those of No. 14, but are so excess- ively overgrow y blackthorb n n tha l measuremenal f to t ou s wa t the question I judge .two largee e th th df .e o rthib o st The next group of towers which I investigated lies on the north side of the Tummel basin, where I examined three, two on the farm of Borenich and one near the Queen's View. . Ther16 , howevereis t leas a besides,e on t norte , th als n hoo sidf eo Loch Tummel, and a mile due west of that just to be described. 17. Borenich.—This fort stands on the moor about a quarter of a mile nort f Borenico h h farmhouse d abou yard0 an , 30 t e s th eas f o t burn which runs by the cultivated fields to'the north of the public road. This is quite a large ruin (figs. 12 and 13), in a comparatively fair state

1 Transactions of Gaelic Society of Inverness, xxvi. p. 141.

THE CIRCULAR FORT P NORTO S H PERTHSHIRE9 4 .

of preservation stretcA .nort e walf th h o n hlo sid stils ei l abou feet4 t e inneth hig n s robscurei ho face t a largbu , y db e quantit f falleo y n stones. The doorway faces the west, but its exact breadth is not cleare outeTh r. wall-fac s veri e y distinc l round s foral e t wa Th t . nearly circular, wit externan ha l diamete n feet8 wald a 6 f an f ,o l o r almost uniform thickness of 10 feet, in this respect differing from the

Fig. 14. Borenioh Fort. (No. 17.) Loch Tummel and Schiehallion.

walls of the forts described above. The site gives a fine view of Schie- hallion and Loch Tummel (fig. 14), but eastwards the view is cut off by Borenich Rock. The foundation of a wall starts immediately to the e doorwasoutth f o h y (left hand looking west) d run r abouan ,fo s t yards5 1 , terminatin stona n g i t havy no moun ey ma d r o whic y hma been a stone hut. Another wall starts from a point a little more to e leftth , run r abou fo se sam th t e length d terminatean , s similarly. VOL . . XLVII. 4

52 PROCEEDINGS OF THE SOCIETY, DECEMBER 9, 1912.

There are several small cairns in the vicinity, and a little to the east e foundatioth s i n ovaa f lo n structure (ston d earthan e f abouo ) t

11 yardyards8 y sb , paced fore places i Th t v d nea e moutth r a f ho pass leading to near Baile nan Stiubhartach in Atholl. . Borenich.—18 The second for less i t s than hal mila f e eas Borenicf to h farmhouse birca n i ,h sout e wooth n hpublide o sidth f eo c road, about

Fig. 19. Queen's View Fort. (No. 19.) Schiehallion and Loch Tummel.

100 yards from the road, and visible therefrom. This, again, is a considerable ruin (figs. 15 and 16). It also shows about 4 feet of the original inside wall-face on the north for a distance of about 20 feet, and here also the wall is obscured by a great quantity of stones that have fallen inside the fort (figs. 17 and 18). The door faced west, and was narrow, a little over 3 feet, as I should judge. The fort seemed to have been not quite circular. Its greatest external diameter is 66 THE CIRCULAR FORT NORTF O S H PERTHSHIRE3 5 .

feet, with a wall of from 9 to 10 feet of fairly uniform thickness. externao n Somw sa e l I wall cairn t t wil I see e bu close . b l, sli n eby that this vers forwa yt similas companioit o t r n fort about three- quarter mila f eo s distant, except tha s situatioit t gooa s ni d deal lower. . Queen's19 View.—The for t Queen'a t s Vie wquits i e milea th o et east of No. 17, and is situated about 300 yards north of the public road and a little to the west of the dyke. A large rowan tree growing within it forms a distinctive landmark (fig. 19). The view from this poin vers i t y fine indeed e for s nearli Th t . y circular I mad. e eth external diameter 77 feet, and its wall about 10 feet thick. The door I coul s a face dr weste judgefa th do s , . Althoug place hbees th eha n mercilessly quarried e hugth , e stones that remain, especialle th n yo north side, form a striking sight, and give a strong idea of the power and solidity of the original structure (figs. 20 and 21). A glance at the map will show the relation of this fort to the pass from the east leading to the north side of Loch Tummel. On the opposite side of the river, and directly to the south, is Dun Teamhalach (Duntanlich), meaning Tumme lI regrette Fortd an , d that I could not inspect a place bearing a name so promising. I greatly regretted also that time and daylight failed me to examine the " castle " at the entrance to Fincastle Glen, though I passed near it. The whole of this glen deserves examination. The next grou f "o p castle s " that came unde observationy m r , though imperfectly, is below the great pass of Killiecrankie, near Pitlochrie. . Pitlochrie.—20 The e uppe th firs f thesn o ro t Pitlochris ei e golf vers coursewa y t largeinneI e . Th r. wal s perfectli l yo cleatw t a r points, which are fortunately directly opposite to each other, and the distance between these externapointe feet0 Th 8 s .i l diamete rmadI e 10 thicknese 7th feetwald e th teste an e ,fee f l3 b so 1 wher n ts di ca t ei 6e eas inchesth courset n w e masonrsidO fe th . e innea f so th f ryo

\\

THE CIECULAE FORT F NORTO S H PERTHSHIRE5 5 .

wall-fac stile positionar en i l viee wTh . extends e passoutth f so o ht Dunkeld. Two driving " tees " have been constructed on the wall, oae on the west side, the other on the east. The stones that remain largseverao n i s s t ea largee no l t ar othebu , r forts. 21. Edradour,—Eastwards of Pitlochrie, Pennant mentions the Black Castle of Edradour, 60 feet in diameter, with wall 8 feet thick. I had no opportunity of seeing what remains, if any, exist, but I talked wit a hmiddle-age kneo e ruinwh wth n . ma dPennan t also records thi f anotheo seminenclearneI t n a bu n sout e , ro it th d f ho eo t nothing. 22. Pitcastle.—Three miles below Pitlochrie, on the high ground above Moulinearn, there are the remains of.a circular fort on the farm of Pitcastle. The site is south of the public road, and a little S.W. of farmhousee th t planteds thicklyi no t I surrounded t . an bu , y db interioe walla e for Th fairls th i .tf o r y e walclearfacee th f lTh o s. visible ar e herthered t ean ther febu e , wear large stones visible, most of the facing stones having been removed. The form was fairly circular, with an internal diameter of about 68 feet. The wall measured approximately from not less than 8 feet to 13 feet. I havdescribew eno d such circular fortI observed s a s r knoo , o wt exist, in the basins of the Lyon, Upper Tay, and Tummel. Another group exists in the Amulree district—Glenquaich and Strathbraan. 23. Glenquaich.—Of these I saw only one, that on the south side of Loch Freuchie sligha n o , t eminence immediatel e soutth f o o ht y the public road, and a very little east of the U.F. Church. It is exceedingly dilapidated. The inner wall-face is practically obliterated. The external diameter is about 83 feet, but the thickness of the wall could not be determined with certainty at any point. The site commands a wide prospect eastwards. It is right in the mouth of e pasth s leading from Achnafre n Glei e n Almon o Glenquaicht d , whence one goes N.W. to Kenmore, or by way of Amulree and Glen Cochui o Aberfeldyt l r eas o y Strat, b t h Braa o Dunkeldt n . 56 PROCEEDINGS OF THE SOCIETY, DECEMBER 9, 1912.

This fort is called in Gaelic, Caisteal Dubh nan Crb, the black castle calledo s s i t t .i kno e peny no o th wfo r folds wh d so I . 24. After I had left the.district, I learned from the Rev. A. Dewar, ministe Amulreef o r , that ther remaine ear same th fora f f eo so t kind e faron th f Deanshaugh mo e moutth t f a Gle,ho n Cochuil e pasth , s whicy b publie hth c road goes from Amulre Aberfeldyo et . 25. Mr Dewar also informs me that there are remains of another n Glenquaichi , nea e foo f th Lairiro t g Mile Marcaich e pasth y , b s which runs the road from Kenmore to Glenquaich; while there are two mound e shepherd' th e eas littla sf th o t o et s hous t Garrowa e . Glenquaich, whic reputee har maro moredo t sitee tw kth f .so e structureTh s thus described resembl e brochth e n respeci s f o t thicknes possessinn i f wal] o sd an , g onle entranceyon . They differ from the brochs in possessing a much larger diameter. Whether they contained chambers is a point that can be settled only by excava- tion ; inspection reveals nothing certain. In view of the lowness of the remains, no trace of galleries can be expected, unless, indeed, the gallery started on the ground floor, as it does in the structures found by Dr Erskine Beveridge in Tiree and called by him " semi-brochs." On the other hand, the fort on Suidhe Cheanathaidk near Barrachander (Kilchrenan, Loch Awe), partly excavated by Dr Christison, showed neither chamber r galleryno s , thoug othen hi r respects quite broch- like. I saw this fort in 1911, and Mr Charles Macdonald, Barrachander, informe knowd fiftr thaha e fo y e dm t nh i t years seed , s thaha nit e h t walls about 1 note6e b fee gallerieso o t dn t s highi d thad t ha I an ,t .t i that the fort on the east side of Loch Lomond, two miles north of Balmaha, which was excavated by Mr David MacRitchie, possesses chamber broch-likd an s e walls thad an t ,calles i alst owa d Caisteal nam Fiann. Whethe galleried t appearha no t i rr so s uncertain. With regard to the situation of these circular forts, it has to be observed in the first instance that they are by no means hill forts. The true hill for thif o t s regio f Perthshirno regularls ei y placea n do THE CIRCULAR FORT F NORTO S H PERTHSHIRE7 5 .

commanding and, on one side, precipitous eminence, 1000 feet or more above the sea, and affording an extensive prospect. Such are the forts (1) on the south side of Loch Tay, S.E. of Killin, already mentioned ; (2) Dun Mac Tuathail, on Drummond Hill; (3) the e soutth f Aberfeld o ho t e Blac th n ) kDu (4 Castl y; f Balnaguareo d; e for(5 Killiehangin th )o t e Rock. e masonrTh 1 f thesyo e hill forts closely resemble se circula th tha f o t r forts, bot structurn hi n i d an e thickness of wall, and they may have been built by the same people. They would, however, requir a papee o themselvest r e circulaTh . r forts, with one exception, viz. No. 11 above, show no desire whatever for strength of situation. All of them are well within the limit of modern cultivation e greath d t majoritan , f the o ye quit mar e close e presenth o t t public road. Nor, again, hav e siteeth s been chosen with special view to a water-supply, as is so often found in the case r northeroou f e inferredb ny brochsma , t thereforeI . , that their occupants were content to trust to the strength of the wall, and did not expect to have to stand a siege. The possibility of a regular siege was, indeed, practically discounte face th t y thadb t these forts never stand isolated. The e regularlar y ye b so 'n placeca e don thay an t r morsupporteo o e tw neighbourin y db g forts, thud an 2siegea s o t , be effective, would have required a body of men sufficient to invest a

whole group. i: What, then, was the determining factor in the choice of site ? To this there can be only one answer. Every single fort, or combination of forts places i , d with referenc forte passa o Th t es . were meano t t guard the passes. It may be fairly said that invading bands were

Roce th Fonnvuick f fore kn o o tTh , below Killiecrankie toldalsos m i ,a f o ,I , 1 I cannothi t i sf o classt t speabu , k with Statistical d certaintyOl e Th . Account

(vol 475. ii . ) call t " si larga e round castle." 2 The case of Caisteal Braigh an BatUiaire, (No. 13) seems exceptional. But (1) it was quite close to a great hill fort; (2) there was a fort called Dun Ldcha, on the Locha t Killinyther a d hav y An ema e . been othe soutre fortth n hso sid f Loceo h Tay and at its western end. 58 PROCEEDINGS OF THE SOCIETY, DECEMBER 9, 1912.

not obliged to use the passes : they might come right over -the hills. . Here, however e objecth , f suco t h raids mus kepe b t viewn e i t Th . objectiv e raiderth s cattlef o ewhild wa s an e raider, eth s themselves might get in to Glen Lyon and elsewhere without using the passes, it passee th f o was y thas onlwa t y cattlyb e coul drivee db ne outTh . gam cattle-liftinf eo oldes gi r tha historyy nCeltan e Th s. were always fond of it, and the prestige of a chief who proved himself inactive in this respect was lowered. A Gaulish tribe in the pay of Attalus of Pergamos was called the Agosages, Steer-questers. In this connection e grouth fouf po r forts close together nea heae r th Glef do n Lyon, with one advanced outlying fort righ t tha t e passes, suggests that special danger was apprehended from the west, by way of Glen Orchy and the wester Glen Lochy (Gleann Lbc.ha Urchaidh as distinguished from the easter Glen Lochay at Killin, which is Gleann Lbcha Albannach).1 The fortifications in the neighbourhood of Pitlochrie and the pass of Killieorankie require further study, but it may be remarked that the Pitlochrie group appear to have been situated with regard both to e pasth s from G-len Briaracha e grea th Stralocd o ntan t pas d f o hsan Killiecrankie thad e generaan th .t l arrangement resembles that a t the head of Glen Lyon. The disposition of these circular forts might be thought to indicate the boundary of a province, but our present knowledge is too incom- plete to warrant such an inference. It is, however, safe to say that the arrangement is systematic and implies no mean degree of com- e binatiobuilders e th par th e mighf o tn On e inclineo nb .o t g o t d thay furtherdispositione sa th t d an e hea, th f Gle do t s a n Lyone th ,

strategie Th cjunctioe pointh t a tf Gle no n Orch d Gleyan n e Lochth y y(b 1 latte whicf o r enter e hon s Glen Dochar t Tyndruma t Dalmallys i ) . Dalmalls ywa fortified I hav t seet fortificationss eno bu nit , othee Th r. great pass leading into Glen Doohart from Loch Lomond sidd Inveraraean Gles yi n ITalloch e fortiTh -. fications of Glen Palloch were, so far as I can learn without visiting the spot, near Glen Falloch farm-house, where Lairig Airnein, e pasth s from Inveraray, meets Glen Falloch. THE CIRCULAR FORT NORTF O S H PERTHSHIRE9 5 .

foot of Strath Tummel, and even, perhaps, below Killiecrankie (if we suppose the forts in these groups to have been occupied simul- taneously) provided accommodation muc n excese i requireh th f o s - ments of any possible population in those districts, and have the appearanc a garrison f o e . Ther s nothini e e earlth yn gi historf yo Scotlan o precludt d e ide f th eorganisatioo a e d systemth an n t A . battle of Moris Graupius in 85 A.D. the native Caledonians turned out 30,000 strong under one leader, Calgacus, and whether they were defeate o signalls d s Tacitua y s would hav s believu e r noto e e ,th Romans advance furtheo dn r then 131n I . 4 A.D., King Robert Bruce, with a year's preparation, mustered less than 30,000 men for Bannock- burn. The possibility of a quasi-military occupation of the forts accords with tradition, which calls them Caisteilean nam Fiann, Castles of the Fiann. The Fiann, as Professor Kuno Meyer has lately pointed out,1 were band f warrioro s s unde leadea e rul th f r eo r chief o rd an , having some relation, thoug s exacit h t natur s difficuli e o definet t , centraa o t l authority institutioe Th . commos nwa Irelano nt d dan Scotland. It is notable that this is the only group of forts in Scotland, so far as I am aware, with regard to which there is a definite tradition. The district in which these fortifications occur formed the northern part of the old province of Fortrenn. The name Forttenn is derived, through Gaelic phonetics, from the tribe of the Verturiones, mentioned by Ammianus Marcellinus (ewe. 364 A.D.) as one of the two divisions of Pictse th . Principa Johr lSi n Rhy explaines sha name dth e Verturiones derivativa s a f Vertera,eo whice locativth n hi e plural form Verteris was the name of a place in Westmorland, now called Brough-under- Stanmore.2 Vertera is represented in Welsh by gwerthyr, a fortifica- tion. Verturiones therefore means " men of the strongholds," or " fortress-folk. n GaeliI " c phonetics Vertera would become Fortair, or, in modern Gaelic, Fartair. Now, the name Fortingal (accented on

Fianaigecht: Todd Lecture Series, vol. xvi.

1 2 Celtic Britain ; Additional Notes. 60 PROCEEDINGS OF T.HE SOCIETY, DECEMBER 9, 1912.

e firsth t syllable a lat s ei ) Englishe de Gaeli forth f mo c FartaircMll, substantially the same to-day as it was in the thirteenth century, when s spellewa t di Forterkil. e firsTh t par f Forter-kilo t , then, means " stronghold. lattee Th "r par s till,i t locativ f ceall,o e a church , a loan from Latin cella, parise cella Th .h churc f .Fortingalho , with its famous yew-tree, is of great antiquity as an ecclesiastical site, and ther reasos ei believo nt es immediat thait n i t e vicinity a ther s wa e pagan place of meeting and of judgment.1 .Further, it lies close to the rock on which stands the great fort of Dun Geal (No. 11), which may therefor e takeb e s havina n g supplied the differentiating part s namoit f f Forter-kileo r Fortress-churcho tere Th m. Forter occurs also in the name of the ancient castle of the Ogilvies in Glen Isla. e sit f whicth o es occupie wa ha prehistori y b d c fort. Philological evidence, therefore, points to the conclusion that the circular forts under consideration were the fortifications which caused the Cale- donians of the district between Forth and Tay to receive their distinctive name of Verturiones, or " men of the strongholds." For the photographic illustrations I am indebted to the Rev. M. N. Munro, Taynuilt.

1 Duneaves, on the opposite bank of the Lyon, is in Gaelic Tigh Neimhidh, e HousNemedth f o e . Nemed, Gaulish Nemeton, a paga s wa n meeting-place, analogou Norse th o est Thing.