44 PROCEEDINGS OF THE SOCIETY, 1957-58.

III.

E ORIGITH N CENTRE PICTISTH F EO H SYMBOL STONES.

BY ISABEL M. HENDERSON, M.A.

In this paper I use the classification of the stones given by Romilly Alien s descriptivhi n i e earle th lis f yo t monument Scotlanf so d c1 Clas . sI Undressed boulders with incised symbols. Clas . sII Dresse d slabs with symbols, cros figurd san e sculptur reliefn ei . Class III. Similar sculptured slabs but Avithout symbols. Sinc publicatioe eth Romillf no y Alien's lis 190n i t 3 additional symbol stones have been found with more or less regularity. His statistics Table,2 therefore, requires some revision. Apart from numerical correction there are other ways in which the Table can be made more useful. There is no doubt that his inclusion of monument statistics for the non-Pictish area makes the lack of symbol-bearing monuments there an impressive blank on the Table, which shows clearly the national character of the symbolism. unfortunates i t I , however, that all- figures have been carried over by other writers into discussion concernin e developmength purele th f o yt Pictish monuments. The presence in the statistics of numbers relating to, say, West Highland Crosses, can only be misleading. Alien gives statistics for each modern county. The Table could perhaps be given more significanc e lanth df i eunit s were taken fro e earliesmth t relevant land survey. Thi e Poppletos founi sth n i d n MS.treatisa n i 3 e beginning De Situ Albanie . . . composed probably in the reign of William the Lion (1165—1214). The survey omits and so presumably refers to Pictise th h area prio 843o rt perioa ; d very relevan symboe th o t t l stones. Accordin e surveth dividelane s o gt yth dwa d anciently into seven parts: "Enegus cum Moerne" (Angus and the Mearns); "Adtheodle et Gouerin" ( and ); "Sradeern cum Mejieted" ( and ); "Fif cum Fothreue" ( and Kinross); " cum " (Mar and Buchan); "Muref et Ross" ( and Ross); "Cathanesia" (). Most of these names are used as district names in present day Scotland, whether they still represent exactl same yth e territorial area less si s certain. 1 The Early Christian Monuments of Scotland (Edinburgh, 1903), n, 3. The introduction is by Joseph Anderson (Rhind Lectures, 1892). In this paper numbers after a place-name refer to Alien's enumeration of the stones at any one site. 1 . Ibid.,10 , II 3 Bib. Nat., Latin 4126. The treatise is translated in A. O. Anderson, Early Sources of Scottish History (Edinburgh, 1922), I, cxv-cxix. This MS. contains several Scottish pieces including a list of the kings of the and a list of the kings of the Scots. E ORIGITH N CENTRE PICTISTH F O HE SYMBOL STONES5 4 .

A. second survee samth en yi treatise define e lanth s d division naturay sb l feature thid san help determino st e just wha meant. s uniti te Th s which

seem to have departed most radicall1 y from the modern areas are those e MounthBuchand th an N f r o . Ma ,; Mora Rosd Caithnessd yan san r Fo . example, "Marr cum Buchan" must bear some relationship to modern Mar and Buchan but these areas as we know them are not the equivalent of the area defined by the second survey: "... from the Dee to the great and •wonderful river tha s callei t e Spey.dth " However, inexact thouge hth interpretatio e surveyth f no se impressio musth , e be th t f o n t whicge e hw territorial division certais si n enoug mako ht e them useful. Her intere eth - pretatio f "Marno Buchanm cu r " e secongiveth n ni d surve accepteds yi

with the exception of Strathspey itself whic2 h is treated as part of Morayshire. secone Th d survey give topographicao sn l clues abou aree th t a "Muret e f Ros." However we can safely assume that Moray comprises modern Nairn, Moray and Inverness-shire and that Ross means . Caithness is secone t mentioneth no n i d l surveyal t da . Watson define s Caithnesa t i s s south-easd an t .3 The following Table will give statistic lane th dr unitsfo s e giveth n ni treatise, omit all non-Pictish statistics and include corrections of Alien's discoveried ad tabl d ean s published later.* Clas I figuresII omittee ar s s da they would require considerabl et relevan re-workine no th e ar o t t d an g present discussion. Statistics form the bases of arguments for both Alien and Anderson, particularly in their discussions on dating and origin. The question of origin, alone, wil treatee lb d here. It is clear that the custom of erecting stones with incised symbols began in the N., that is to say N. of the mountain range known as The . The numerical superiority of the Class I stones in this area is such as to make thi certaintysa . Alien believes thapossibls i t i t locato et origie eth n centre even more specificall witd mucys an h a h certainty. frequenc"Ie fth e th f yo occurrence of the monuments is any criterion of their origin then the table that has been given clearly points to as the home of the stones belonging to Class I . . . ." 6 This interpretation of the statistics is shared by Joseph Anderson introductioe th n i Alien'o nt s list. Alien recognizee sth

t includeI s Argyll however Poppletoe th sevee n th I als. lisa f n.o s to i nMS son f Cruithneo s , geographica1 l eponyms of the districts in Pictland. All but two can be equated with the units in D fula e r situ.l discussioFo e surveyth f no . Ws v . WatsonJ . Historye Th , Celtic e oth f Place Namesf o Scotland (Edinburgh, 1926), 107 ff. 2 In some ways the area covered by the old bishopric of might be taken to represent "Marr cum Buchan," but such an equation ignoring as it does the evidence of the second survey would require further investigation. interpretatioe th r Fo 3 Moraf no Rosd Caithnesd yan san . Watsonsv cit.,. op , 115-17. 4 For Class I additions and corrections see Appendix. : 5 B.C.M.S., n, 13. He suggests that might be the centre from which the habit spread northwards and southwards. 46 PROCEEDINGS OF THE SOCIETY, 1957-58.

SYMBOL-BEARING MONUMENT PICTLANDN I S .

Areas Class I ClasI sI Angus 10 32 The Mearns . 6 1

Total 16 33

Atholl 1 2 Gowrie 4 14

Total 5 16

Strathearn 2 3 Menteith

Total 2 3

Fife .... 3 2 Kinross

Total 3 2

Mar 1 . 49 6 Buchan 5

Total 54 6

Moray .... 34 3 Ross .... 8 5

Total 42 8

Caithness 5 3 South-east Sutherland 16 1

Total 21 4

Western Isles 2 5 I 1 . 4 1

1 Mar is interpreted as Aberdeenshire excluding Buchan, and all Banffshire excluding Strathspey. The provinc proper Ma f reo represent mucsa h smaller area. Westere Th 2 n Isles, Shetlan Orkned t appea situ.de an no D o yn d ri E ORIGITH N CENTR EE PICTISOTH F H SYMBOL STONES7 4 .

assumption which lies behind this conclusion: area of prevalence signifies are f origino a . Anderson believes that this must stane db unlesn ca t i s shown that there were special reason prevensuggeste o t se H on . s it ta s such reason, the sudden conquest by an alien race, which might compel the people employing the symbols to remove to another area, and there to reduplicate their symbols. But, Anderson adds, since all classes of stones, I, II and III are represented in Aberdeenshire there is no suggestion of a disturbing influenc thaf o e tbelievee h sort o s ; s that throwe "..ar e nw . back on the unavoidable conclusion that the type of Class I originated and was developed in Aberdeenshire . . . ." 1 The following points can be made in criticism of Anderson's argument: 1. The fact, even if it were allowed, that there is normal continuity from AberdeenshirClas n i ClasClaso o t t I I sI II sI a t no s ei valid proof that the centre of origin of Class I did not suffer a disturbing influence. A centre of origin other than Aberdeen- shire might stil postulatee lb d which suffered durin earliese gth t period of the use of symbol stones "a disturbing influence" sufficient to cause a transfer of settlement to Aberdeenshire and multiplicatioe th symbole th f no s there. 2. Anderso thus ni s left with onl argumens yhi t base statisticn do o st support his suggestion that Aberdeenshire is unavoidably the centre of origin. Now while the numbers of Class I stones there must remain impressive s possibli t i , offeo et a statisticar l argumena r fo t more northerly origin centre. The impressiveness of Aberdeen's numerical lead depreciates considerably whe e considew n r statistics relevant to the more ancient land divisions. Alien attributed forty-one stone Aberdeenshiro t s e wit s neareshit t rival Sutherland with fifteen. The new relevant figures are: Mar and Buchan fifty-four; Moray and Ross forty-two. More- over sixteen of the twenty-one examples in "Cathanesia" are in SB. Sutherland. If then, high numbers, together with normal representation in other classes are a criterion of the origin centre, then we can look as justly to the coastal strips round the Moray and Dornoch Firths as to Aberdeenshire. 3. Anderson writes elsewher discussions hi n ei : "Wtherefory ema e conclude with probability that lonno tg d Clasbeeha nsI intro- duced into Forfarshire change s whewa t ni d inte th o y Clasb , sII advancement of the art, but that the advanced type took some time to spread into Aberdeenshire where the primitive type

continued prevailo t . . " . Thi. s admission greatly reducee sth 2 1 Alien, op. cit., I, cv. 2 Alien, op. cit., I, civ. 48 PROCEEDINGS OF THE SOCIETY, 1957-58.

significance of Aberdeenshire's large numbers as evidence for the origin centre. Aberdeenshire's slightly inflated numbers may be simpl failurresule e yth th backwara f f o teo d are accepo at t the advanced type. Andersen's further point, that the disuse of the symbols meant virtually the disuse of the monuments in Aberdeenshire, might argue that the Aberdeenshire stone masons were solely interested in carving the symbols because these belonged especially to Aberdeenshire by right of invention; but it might also argue once again the conservatism of a backwater area. We ca facn ni t suggest tha abundance th t f Claseo sI stone simpls si ya indication a t corollarno lacd e n f Clas th an ko tha o I yt I s t s Claswa I s invented in Aberdeenshire. The matter of course, could only be one of subjective interpretatio statistice th f no s unless there were positive indica- tions that some of the Class I stones which make up the total for Aberdeen- shire showed signs of distinct lateness in the Class I series as ,a whole. Is ther evidency ean thir efo s ? Some of the Class I symbol stones bear Ogams. It is known that the Ogams of Pictland for the most part belong, at the earliest, to the century,h 8t tha o e presencs th t Oga n a a Clas f n meo o I sston e would

indicat1 e a date of that order.2 Aberdeenshire has two Ogam-bearing symbol stones; Logic Elphinston Brandsbuttd an e2 botn i ; h caselikels i t si y that inscriptioe th contemporaneous ni s wit symbolse hth thao s , t these examples ca safele nb y considere t belonginno vers e da th yo g t earlies t period. This, however, accounts for a very small number of the Aberdeenshire examples. Mrs Cecil Curie writing of the Class I stones comments: ". . . there is considerable variation in the quality of the design and workmanship; the best example foune sar extremOrknee n di th d yan e nortmainlande th f ho , while many further south, particularl Aberdeenshirn yi vere ear y debased." 3 It seems fair to suppose that a symbol which is debased is a late one. Mrs Curie however, doe t attempsno describo t natur e Aberdeene eth th f eo - shire debasement, sayo that s ,i t wha qualitp to t f breao y p desigku s nha occurred.4 R. B. K. Stevenson has recently analysed the variation in the design of interioe th r decoratio symbol—the on f no e crescent. crescene Th 5 t allows

this kind of analysis because the decorative element is sufficiently elaborate. . JacksonK 1 Probleme Th , Pictse oth f (1955), 139. 2 The Auquhollie stone is an exception v. ibid., 139. Alien believed that the stone bore no symbols. Diack, P.S.A.S. Lix (1924-5) point t thaou s t ther smale ar e l symbols f OgaI .symbold man s were put on the stone at the same tune then the dating of the Ogam would give valuable evidence for the firse datth tf eappearanco symbolse th f eo . 3 P.S.A.S., LXXIV (1939-40), 65. 4 The debasement of course in no way affects the stereotype quality of the symbols whose essential

form remain matte a qualit e s samee representatioe i th sth t th f I o rf y.o symbole th f no . 5 The Problem of the Picts (1955), 104 ff. THE ORIGIN CENTRE OF THE PICTISH SYMBOL STONES. 49

a GOLSPIE b C MONIFIETKINTORE H

d CLYNEMILTON 6 KINTORE MONIFIETH

g WHITECLEUCH K ANWOTH 1 MONIFIETH

j GOLSPIE K I LARGO UU tn WHITECLEUCH tl MILL OF NEWTON 0 ABERLEMNO Kg. 1. VOL. XCI. 50 PROCEEDINGS OF THE SOCIETY, 1957-58.

Any results which his analysis yields is particularly helpful for the crescent commonese th f o e ison t symbols. Stevenson shows tha representatione tth s incision i crescene th f no t symbo dividee b n ca ld into three classes according theio t r main features: Clas , thosA s e wit ha centra l pelta-shaped figure either way up (fig. la); Class B, those with two spirals; Class C, those with a central dom d winan e g shapee sideth t s a s (fig. 16)fourtA . h class i s typica f Claso l I representationI s e crescentth f o s ; here scrol peltd an l a patterns are abandoned and any variety of decorative infilling is used (fig. Ic). overlappine Th firse th tf go thre e classe desigf so n suggest scommoa n origin thid founs an si d most naturall mose th tn y i comple x pelta—"t i r fo . . . contains a majority of the details found in the others, details which it would be hard to combine but which could have separated during simplification.'' This most complex example is an example of Class A and it is found on a ston t Golspieea , Sutherland desig e Golspie th Th f .n o e crescen therefore, tis , the prototype design and in Stevenson's words—"... it follows that all the decorated crescents are later than a design of that type."

The particular result thif so s analysis bear t Curie'wels sou Mr l s observa- 1 tion. Aberdeenshir representatioo n s eha e clas th f crescenf so no t nearest to the prototype design. It has, however, a number of very debased crescent designs, notably at Logie Elphinstone 2 and 3—including the stone already classed as late because of its Ogam; Kinellar; ; Kintore (fig. 16); Daviot o suggest .t Thino t s i stha l Aberdeenshiral t e crescent-bearing symbol stones are late, but Mr Stevenson's Table does show, that many representations of the crescent in this area show late-looking forms. If it is agreed that the origin centre of crescent-bearing symbol stones is most likely •where b o t e thergreatese th s ei t preponderanc correctlf eo y drawn designs, Sutherlando t , theN. musr e nfa w te , looth Caithnes o kt Orkneyd san . Nopossibls i wt i exteno et analysin da e crescen th f o sinclud o t t e eth terminals of its characteristic modifying v-rod. As one would expect, a surve terminale th f yo s shows tha "correcte th t " form foune ar s d regularly on the crescents belonging to Mr Stevenson's Class A. These are a blunt shara fish-taid pan arrod en l w hea (figd den . ~Ld). These terminal forme sar for the most part carefully maintained throughout the corpus of Class I stones. "With Class II there is a considerable breakdown. Terminals of representations of the crescent in Class II may show such anomalies as identical terminals (fig. I/) wilr o , l borro flame wth e terminals reserver dfo z-rod Clasn si s I. This breakdow terminae th f no l syste naturae th ms i l

complemen abandonmene th o t 2 scrole th f o tl infillindesige e th r th nf fo g o

1 Accordin Stevensoo gt Golspie nth e design itsel relates i f hanging-bowo dt l patterns. 2 The classic form of the z-rod seems to involve an arrow end and a flaming end (flg. Ig). The identical terminals on the z-rod at Anwoth (flg. 1A) suggests that the symbols there are not evidence for early Pictish occupation of . Identical terminals are a typical Class II perversion (flg. li). The penetration of the "spectacles" by the rod is also a late feature being an elaboration natural to relief sculpture. Thi alss si o incisefoune th n di d representatio t Anwothna . THE ORIGIN CENTRE OF THE PICTISH SYMBOL STONES. 51

crescent itself exampler fo , stone th , t Hiltoea Cadbolf Monifiete no th d lan h plaqu ee lat (figth e f ./)d crescenO I.can t design f Aberdeenshiro s t ea least one example has late terminals, namely Kintore (fig. le). A comment by Romilly Aliestona n no e from Little Ferry Links mile2 , s wes f Golspieo t , is relevant here: "Although a mere fragment, this stone is valuable as affording one of the most perfect and beautiful examples of the ornamental terminatio v-shapee t founth f ye n o Scotland n d di dro artistie Th . c feeling exhibited in the drawing of the subtle curves is unquestionably great." x Once again a superlative example is located in the far N. Th ecalleo s natur desige e dth th f swimminf eo no g elephant allowo t t si underg analysin oa s simila thao t r t whic r StevensohM n providee th r fo d crescent. If we examine all the incised elephant symbols we see that the designs descend frocommoe mon n formul a highl f o a y complicated kind. That complicated formula appears on the stone at Golspie, that is to say e Golspith e elephant (fig. lg) contain e featureth l al s s whic e founar h d dispersedl designe th n y i othef so r incised elephants. These featuree th n i s words which Stevenson employ prototype th f so e crescent, "woul hare db d to combine but . . . could have separated during simplification." Thus the Golspie elephant shoul regardee d b e prototypth s l otheda al f ero incised elephant thid san primac confirmes yi face th t y thadb t appearti e th n so same stone as Mr Stevenson's prototype crescent. Its features are: the long flowing line of the back, running down without a break to the tip of the nose; heae tuckine thath o th df s t o runi t n gi s parallel wit limbse hth innen a ; r line articulatin bodye gth , endin larga n g i insid ehine e lobth th df n e o o leg , foreleg e joinscrole ana th th f n d i o t secon a la ; d inner line running parallel to the front of the foreleg from the spiral foot to end in a lobe on the chin; extrn a a cur botn forhino e ld hth ean d leg doubla ; e line lappet following backe line th f th e.o An analysis of the elephant is especially useful in an examination of the nature of Class I stones in Aberdeenshire, for the Aberdeen area is, in quantity, particularl e elephante homth yth f o e . Whee th looe t nw ka Aberdeen elephants we see that examples close to the prototype do exist, notabl t Crichieya . Ther howevee ear numbea r latf o r e looking designs: for example, the elephant at Dyce with its legs dropping in different directions, s turneit snoutp du s clearlit , y demarked foreheasquins it d tan d runnin g foreleg lineelephane th ; t Kintora t e wit s anglehit d protruding forehead, its turned up snoxit and absence of the leg curl. Of the two examples of elephants at Rhynie one has something of the greyhound quality of the Golspie elephant t omitbu , s entirel articulatine yth g inner line scrollsd san , the other, a fragment has the tilted snout of Dyce entirely at odds with the tucked in smoothly flowing head of the other example. Clatt has a good flowing design, but it is simplified by the reduction of the lappet to a single 1 E.C.M.S., HI, 47. 52 PROCEEDING E SOCIETYTH F O S , 1957-58.

line and of the interior scroll to a single short open lobe. The head has no mout d onl n embryonian ha y c eye. Among these "Ma d Buchanan r " examples are three particularly late seeming designs: the elephant at Mortlach shows the head supported on a very thin neck, the head itself is large and has the snout separated off as a kind of beak, there are no interior lappee spiralth s reduced i t an sa singl o dt e linee elephanth ; t Fyvia t e (fig. 1/e) is squat and ungainly in design, with heavy projecting brow and short body lege th ,s dro differenn pi t direction line th e d articulatinsan e gth body meanders meaninglessly acros e body th e selephan th ; t Logia t e Elphinston heaa s dha quite3 e misproportionat heavs i bode d th yyan o e t browed. This stone at Logie Elphinstone, as we have seen, also bears a debased crescent. analysie th elephane f th I f assumptione so th correc s i tf i d an t , deviation equals latenes time,n si corrects i , the more nrepresentatioy th ean n deviates fro prototype mth more eth t eshouli d approximat representatione th o et n so Class II monuments which, it is agreed are later than Class I. In the case of the infilling of the crescent this argument was impossible as Class II coincided with the abandonment of Class I patterns. We did see however thae debaseth t d terminal f Claso s I sapproximate typicao t d l ClasI I s representations e sam Th s truei . f debaseeo d elephants. Typicae th f o l Clas I elephansI beae th ks i t treatmen snoute th upturnine f o th t; e th f go snout; a bulging forehead; misapplied interior lines; and above all a rigid squat qualit whole th n yei design quite unlik flowine eth g bese gracth tf eo Clas sexampleI s most perfectly expresse t Golspieda . Example Clasf so I sI elephants illustrating this point are at Meigle 5, Scoonie, Brodie, Ulbster d Largan o (fig. 1Z). greaa t Thertno mane ear y representation symboe th f so l knowe th s na notche possible b dy rectanglderivo ema t t i t ecorreca ebu t symbol design. The most complex pattern is that of the curious design found on the terminal e chaith rin f ngo from Whitecleuch, Lanark notcheo (figtw . , 1m).it s n I lefoccupp to t hane y th middl e d th thir d edan righ t hand third respectively. thire Th d notc takes hi n e rectanglefro th fooe mf th o t , formin legso gtw . An exact reproduction of this highly specific formula is found at Clynemilton, Sutherland; Birnie, Elgin; Arndilly, Strathspey stona n t O Inveralleea . n the design has been made symmetrical by the placing of the side notches on either side of the cross bar of the modifying z-rod. In Aberdeenshire, Tyrie presents a correct form, while Mill of Newton (fig. \ri) simplifies the design in the obvious way by placing the side notches opposite each other, and a further simplification appear rectangle th e Clasn si th I f AberlemnoI seo , Churchyard, Angus (fig. lo). The analysis would seem to bear out the northern prerogativ mose th r t efo accurat e representation greatese th n si t number woule t wison dt no pres o ht sbu smal e viese th thi th f w lo n s i numbe r of representations. E ORIGITH N CENTRE PICTISTH F HEO SYMBOL STONES3 5 .

There is no doubt that the animal symbols had a correct representation sense th en i tha have w t e been using extraordinare th ; y duplicatinge th f o Burghead a stereotypbul n i l e manner alone would suggesa t thid an s "correct" duc eagld kan e could probabl isolatede yb . Howeve mosn i r t cases the number of representations of each animal is so few that an evolu- tionary series possiblwa s t i e o presensuccrescentt e th s a hr d fo t an s elephants canno e constructedb t e mak.w Noen ca comparisonr s with Clas I animalsI s ther quite fo , yar e different animalss a indeed r an fa , s da teln wca el animal Clasn i s I losI s e their symbolic significance. Whas i t clear howeve tha s i ranimale th Roste Morad th san n si y are designee aar d wit e samhth e masterly assuranc e Golspith s ea e elephant. Moreovern i , animalelephane e th th d vere san th t y same articulating interior lined san lobe e found e samar s Th e. sculptor coul responsible db e Golspith r efo e elephan Ardrose th d an ts wolf. The area centrin Invernesn go s lonha sg been recognise speciae th s da l homanimae th f eo l symbols analysie crescene Th Stevensor th . f M so y b t n analysee th d an screscene herth f eo t terminals elephane th f o , t symbod an l of the notched rectangle suggest that the system of symbolism belonged in entirets it thiyo t s region. These result suppord s an bea t interpretae rtth ou - tion given earliestatisticae th f o r l table. bees ha n t I observed glanca e distributio th d t an a e, makep nma t i s clear, tha symboe th t l stone sfertile keeth o ept river valleys, notably that of the Spey, and of the Don and Urie. It is a striking fact however, that the stones tend to occur at the tops of rivers. There are no stones on the rivers Findhorn, Spey, Deveron, Nort Soutd han h Tyrie, Ugie, Ythad nan Dee, until a minimum of 10 miles upriver. Equally the centre of concentra- tion on the Don-Urie valley does not begin until 12 miles inland from the Don mouth. This tely faclma tu s something abou kine th f sitet do e sth Pict equatine sth liket livbu do gt , occurrencee oin fth symbof eo l stones with populated areas is unjustifiable for as Stevenson has rightly remarked, ". . . the north-east corner and are curiously bare;" both these desirable regions must have supported a population. It may be thin possibli se featur se distributioe o eth t f eo somp nma e further evidence e origifoth r n centr d somean e indicatio e routeth f no s along whice th h spreasymbole th f do s took place. The Clas I se extremstone th n i s e N.—Sutherland, Caithness, Orkney d Shetlanan e almosar d t without exception, coastal. Immediatelf o . yS Sutherland the picture changes and the stones show a distinct preference for inland greasitese Th t. jutting peninsula betwee e Dornocnth d han e Cromartth y Firths r examplefo , s completeli , y ignored fino e Tw incise. d animals are 5 miles from the shore of the Cromarty Firth. While the majority of the Ross examples are tucked in at the foot of the firth, the especially fertile is ignored, and examples do not occur 54 PROCEEDINGS OF THE SOCIETY, 1957-58.

at all on the shores of the Moray Firth until the mouth of the Ness. With the exception of Burghead and two associated stones there are no Clas sI monument s alon e coastgth north-easf so t Scotland from Aberdeen Invernesso t , thereforeis t inlane I th . o dt , routes tha muse w t t look for the spread of the symbol stones. These inland cultural corridors are fairly evident. They are provided in the main by the valleys of the Spey and Don-Urie have W e .note d that Romilly Alien suggests that Inverurie might be the origin centre from which the stones spread N. and S. Now if the stones originate r dow fa e valle s nth da s Inverurieya s difficuli e t se i , o t t why, in face of the low-lying fertile country, the stones do not continue to the> coast and up along the fertile strips crossed by the Ythan and the Ugie— the area which is curiously bare. The natural extension south from Inverurie is over the Mounth passes; the Cairnamounth, from Kincardine O'Neill to e Crynth ; e Corse Mounth, from Durri o Glenberviet s e th ; El sick Mounth from Culter to Netherley.1 That this natural drop did take place can be seen in the presence of stones on the routes to these passes: Nether Corski Craigmyld ean Cairnamountho et Keith'd an , s Muir, Drumoae th o kt Cryne Corse Mounth clustee Th stonef . ro t Dinnacaisa r which stande th t sa Netherlee enth f do onle y th pas ye examplesar s of .Clas sI symbo l stones in Kincardine. impulse Th e from Inverurie doe t seesno mhavo t e been strong

enoug carro ht symboe yth 2 l stone habi furthey an t r intcountye oth n A . alternative centre nea Don—Urie rth e valley would seeRhyniee b mo t . Rhynie itself has five very fine Class I stones covering a wide range of symbols. That strategithoughf s o e wa b t i o t c importanc t somea e earlpoine th yn ti period is indicate impressive th y db O'Nortp e Ta for t a t h clos . Thaeby aree th ta lona d g ha traditio settlemenf no stono suggestes i tw t ee circleth y dt b a s Rhynie itself and the presence of four others within a radius of 2 miles. From Rhyni have ew ecompleta . W d e sprea an , N. . E. S , stonef e do th o st . downE Rhynie Don—Uriee e th th o t valley: Percylieu; Clatt; Ardlair; Newbigging Leslie; ; Logie Elphinstone; Pitcaple; Drimmies; Brandsbutt; Keith Hall; Inverurie; Crichie; Kintore; Kinellar; Dyce. Rhynie to the N. up the Deveron: Leys of Dummuies; ; Tillytar- mont; North Redhill; . Rhynie to the S. down the Capel Mounth: Glen Muick to Clova; ; Mill of Newton; . Rhynie to the NW. by Balhinny to the upper reaches of the Deveron, across Glen Fidaich to the Spey and thence up Glen Rothes to the valley of the Lossie .- Upper Manbean; Birnie; Basterton of Roseisle; Burghead. This route to Moray crosses the Spey just at the point where the symbol stones start. From there they go up the valley: Arndilly; Knockando; Inveravon; showinp ma a Mounte gr th 1 Fo . Simpson D h . passe ProvinceW e e Th ,s se (Aberdeenr oMa f , 1943), facing p. 130. 2 These stones however migh relatede b t settlemeno t nearbe th t ta y Dunnottar. (Annals of Ulster s.a. 680 Obsessio duin Foither.) THE ORIGIN CENTRE OF THE PICTISH SYMBOL STONES. 55

Advie; Grantown; Congash; Findlarig; Inverallen; Lynchurn; Dun- nachton. The Class I stones S. of the Mounth are interestingly placed. The northerl ye mountai th example f o d nen spasses e stanth t da . Baggerton and Aberlemno stand close to the South Esk, the valley of which joins the Glen Muic Clovo kt a pass. Bruceto rivee closs ni th ro e t Isl a which leads into GleSpittae th n o Glenshef Islt lCairo e p au th nd welean l pass. Strowan stands at the beginning of Glen Garry which leads up to the pass of Drum- Spee ochteth yd corridorran . thenWe e se summary n i , , tha Clase th t sI stone Mounte sth Nf o .e har distributed coastally in Sutherland, Caithness, Orkney and Shetland. The only heavy concentratio founns i d nea miler2 1 Golspie coastlinf e so th n o , e between Little Ferry and Kintradwell. The stones to the S. of this are placed along the inland routes from the Moray Firth down the Spey valley dowd an e routnth e fro e vallee Lossie Don—Urimth th th f yo o et e valley via Rhynie. From these routes there are extensions down the Great Glen and down the Mounth passes. The significant distribution of the northern exampleMounte th f o . hsS confirm alreade th s y very safe assumption that the stones spread from N. to S. (see fig. 2). The bareness of the north-east corner and the feeble impression made on Kincardine speak against Aberdeenshire as an origin centre. If Aberdeen- shir s favouredi e , however a sit, e further west than Inverurie, (Alien's choice) woul preferablee db . Rhyni bees eha n suggested. Ther however, eis , nothin e distributioth n gi p whicma n h speaks agains e Golspith t e area e origith s na centr d indeean ee firsth d t class qualit f exampleo y n o s the Orkney mainland, where there are three Class A crescents and another at close by, suggests that they must lie close to the origin centre, that is, to an area more northerly than Aberdeenshire. Rhynie is well placed for receiving and disseminating influences from the Moray Firth obvious it o s area d s importancan , distributioe th n ei mighp nma t thee nb origie t thath f no no t centr t rathe egatewabu a s a r y whereb habie yth f o t erecting symbol stones reache rese Pictlanf dth o t . d N fro e mth It is exceedingly difficult to imagine what sort of historical circumstances could initiate a symbolism so exact, and so rigorously observed from Pabbay Forth-Clyde th W.e o o Shetlant T o th t , d n N. e i an S .th e n d i linth n ei give the symbolism the prestige of a national system requires an origin centre of first importanc leadea d ean r wielding wide authority. Aberdeenshire seem havo st e bee aren n a littlf ao e interes Irise th ho t chronicler thao s t losa knot o st a e wear wha politicae th t l importanc thif eo o st ares awa the Picts. For the importance of the Moray Firth area we have the evidence of the Irish annals, Adamnan and Bede for the activities of Brude son of Maelcho . 555—84)n(c . Accordin Adamna o gt domusa d ha e regianh some- where near the mouth of the Ness. On stylistic grounds some writers would 56 PROCEEDINGS OF THE SOCIETY, 1957-58.

• CLAS STONES1 S . » CLUSTERN I ii S MOUNTAIN PASSES.

Fig. 2. THE ORIGIN CENTRE OF THE PICTISH SYMBOL STONES. 57

conside 6th-centura r ydifficultie e stonee dat th earlo th t er to ysbu e fo ar s not insurmountable. In many ways Brude fills the role of initiator of the

symbolism "well. He1 was a "rex potentissimus" whose check of the Irish must have give immensm nhi e prestige. reignee H tima 2 t da e whee nw know that the Picts controlled the Western Isles. He had considerable authorit n Orkneyi y f coursbelieve O w . f i e e. Anderso O wit . hA n that Brude was king of Fortrin, then he can have no especial significance for

symbolisa m which starte Ne .th n dHowevei 3 r Adamna Bedd nan e writf eo northera n i m nhi contex therd t an soms ei e evidenc killes ea thawa n di e h t battle agains southere th t n Picts.4 t seemI s the ncounte thath n statisticaf o tso l distribution, geographical distribution, qualit representatiof yo perhapd nan historicaf so l circumstances much can be said for the shores of the Moray and Dornoch Firths as the origin centre Pictisth f eo h symbo justifiee lar stonese w do s perhap: n i s questioning Joseph Anderson's "unavoidable conclusion" that Class I originated in Aberdeenshire.

APPENDIX. Romilly Alien gives a list of the monuments in Part II of the Early Christian Monuments ostatisticaA f ff5 Scotland. . pp t la summary, use boty db h Joseph Anderson and Alien as the bases of arguments, is given on p. 10. These lists do not always tally with the descriptive list which makes up Part III. Here correction made sar Clasr efo sI example stoned san havo st knowe e m bee y nb n found since the publication of Alien's book are listed under the modern land divisions.

Mrs Curie, and for other and better defined reasons R. B. K. Stevenson, think that the elephant symbo1 l canno datee b t d earlier tha . 700n c r Stevenso M . n believes thae elephanth t t symbos wa l suggeste fantastie Picte th th y sb o dt c animal Hibemo-Saxon so n metal wor kHunterstoe sucth s ha n Brooch. In general it seems peculiar that from a series of Hiberno-Saxon beasts, no two of which are identical Picte ,th s should have derived their elephant, which self evidently fro momene mth inventiof to n becomes frozen in a single unvariable form. As it happens, the animals such as those of the Hunterston Brooch, with their bulging foreheads and separated beaks, are more satisfactory analogies for the debased elephants than foprototypee rth design foun essentias it t Golspie dl a al n I l stylisti. c feature Golspie sth e elephant has close analogies in the bottom left and top right lacertines in the vertical sections on f. 174 vo. Booe o th f outee Durrof ko th n r i 650 . animalwd (c ornamentee an ) th n si d plaque beneat centrae hth l hingSuttoe th f eo n Hoo purse (burie 650). dc . Non thesf eo e animal lappetes s i eagle e e th th t n so d bu Sutton Hoo purse and throughout the hoard have fully fledged lappets. Given these Sutton Hoo animals the Pictish elephant could have been invented. Alternatively and more probably, the Sutton Hoo animalelephane th d san t share some common ancestor, either Romano-Britis Britishr ho . 1 hope to discuss on another occasion the whole problem of the relationship of the Pictish animals and the elephant to Sutton Hoo and to Hiberno-Saxon metal work and MSS. notably to the Echternach Gospels, Corpus Christ! College Cambridge MS. 197 and the Book of Durrow. See also for early dating: . ClaphamAW . , Antiquity, vin (1934). RadfordB . A ,. 43-5C , d Antiquity,7an xvi (1942), 1-18. 2 v. Annals of Tigernach, c. 560 and probably c. 574. Scottish Historical Review (1948), 25-47. 3 4 T. F. O'Rahilly, Early Irish History and Mythology (Dublin, 1946), 508. 58 PROCEEDINGS OF THE SOCIETY, 1957-58. ABERDEENSHIRE. Corrections : Add Brandsbutt Ardlair.d an These stones appea Appendie th n ri x (III, 505) t include no statistice bue th ar tn d i s give. II n i Add Craigmyle. This stone is described in III, 158 it is omitted in the statistics given in II. Add Clatt 2. This stone is described in III, 158 it is omitted from the statistics in II. Add Logic Elphinstone . 4 Thi s ston s describeei s havinda g been destroyed (III, 175) t seemI . s fairly certain howeve shoulo s rexis d d thadi d an tt i t appear in the statistics given in II. Add Turriff. This ston describes ei omittes i t i IIIn d7 i d,18 fro statistice mth s in II. Omit Rhynie 3. There is no symbol on this stone. It is included in all the describes i t statisticI . IIIn di II ,n si 182 . Additions: Clatt . 3 Horsesho elephantd ean , P.S.A.S., XLIV (1909-10). East Balhaggardy. Part of spectacles and z-rod, P.S.A.S., XLIX (1914-15). Nether Corskie. Mirror, mirror cascombd ean , P.S.A.S., XLIX (1914-15). Newton of Lewesk. Double crescent, rectangle and rod, mirror case, P.S.A.S., I (1915-16). Tillytarmont . 3 Crescen v-rodd an t , spectacle z-rodd san . Tillytarmont 4. Crescent and spectacles(?). Bot P.S.A.S.,n hi LXXXVIII (1954-6). Total: 52.

BANFFSHIBB. Additions: Advie. Crescen v-rod mirrod an t dan r case, P.S.A.S., (1905—6)L X . Mortlach. Elephant, P.S.A.S., LX (1925-6). Total: 8.

ELGIN. Corrections: Add Easterton of Roseisle. This stone is describe IIIn di , omittes 124i t I . d from the statistics given in II. Total: 13.

INVERNESS-SHIRE. Corrections: Omit Balblair. There is no symbol on this stone. It is included in the statistics in II. It is described in III, 95 and illustrated in III, 517. THE ORIGIN CENTRE OF THE PICTISH SYMBOL STONES. 59

Additions: Fiscavaig. Spectacles and z-rod, crescent and z-rod. Dunvegan. Crescent and spectacles(P). Tote. Crescent spectacle and mirror and comb symbols. All given in P.S.A.S., LXI (1926-7). Invereen. Spectacles and z-rod, crescent and v-rod fragment of a circle crosse perpendiculaa y db r line, P.S.A.S., LXVIII (1933-4). Inverness Museum. Recent acquisition, mirror and comb, fragment of notched rectangle with z-rod and top of horseshoe. Total: 20.

SUTHERLAND. Corrections : Clynemilton whic2 describes hi gives i IIIn d0 i 4 n, mistakenl Craigtons ya n i 2 II, 5 and 80. Additions: Golspie. Crescent and v-rod, elephant, mirror and comb, P.S.A.S., LXXVII (1942-3). Total: 16.

Ross. Additions : . Salmon and eagle, P.S.A.S., L.XXXVI (1951-2). Nonikiln. Poor example. Cast only survives primitivo tw , e spectacles with two triangles set apex to apex, P.S.A.S., LXV (1930-1). Total: 8.

CAITHNESS. Corrections : Add Links of Keiss Bay. This stone is described in III, 28. It is omitted in the statistics give. II n i Additions: Ackergill. Rectangle, P.S.A.S., LX (1925-6). Total: 5.

OBKNEY. Additions : Greens. Crescent mirro mirrod ran r case, P.S.A.S., LVIII (1923-4). Knowe of Burrian, Birsay, P.S.A.S., LXXIV (1939-40) XIIL P . .I d Total: 4. 60 PROCEEDING E SOCIETYTH F O S , 1957-58. KINCARDINE. Corrections: Omit 6. It is doubtful whether there is a symbol on this stone. Auquollied Ad spectacles(P) unnotice Alieny db , P.S.A.S., L.IX (1924—5). Total: 7.

ANGUS. Corrections : Ad d. 2 GlamisThes d an e 1 stone s were originally werClasd an e sI re-usen di the Class II period. They should therefore appear in the Class I statistics. The describee yar IIIn di , 221. Additions: Kinblethmont. Crescen v-rodd mirroan a t f ,combd o elephanan p r to , d an t P.S.A.S., LXXXV (1950-1). Total: 10.

PERTHSHIRE . Corrections : Omit Dunkeld. There is no symbol on this stone. It is described in III, 284. Additions : Farm, Gleneagles. Goose and mirror case, P.S.A.S., LXXXI (1945-7). Inchyra. Salmon spectacle Ogamd san . Perth museum. Total: 6.

FIFE. Corrections: Add Walton. This stone is described in III, 345. It does not appear in the statistics give. II n ni Additions: East Lomond. Bull, P.S.A.S., LX (1925-6). Total: 3.