It is generally accepted that Tacitus' Thule1 is Shetland, and that the latter is the Thule of Ptolemy2, derived through Marinus of Tyre from Agricola's naval reconnaissance in AD83, although the Ptolemaic scale of latitude on 630N is still two to three degrees off the modern equivalent. I suggest that Thule is Shetland in all the classical geographical sources from the time of Pytheas (c.330BC), who was the first to record the name, right through to the time of Tacitus. The comment by Bianchetti3 that the Thule of Agricola had supplanted the Thule of Pytheas needs to be substantiated, while the point made by Aujac4 that Shetland was certainly not Thule is arbitrary and fails to examine the implications of the time-distance factor. The observation of Rivet/Smith5 that the application of the name Thule to Shetland by Tacitus and Ptolemy is not evidence that Shetland really bore the name implies that Shetland had some alternative name, a point which I propose to reject. Roger Dion6 was correct in his assumption that when Pytheas reached Shetland he found the name of Thule attached. Müller7 in his notes on Ptolemy also preferred Pytheas' Thule as Shetland. I believe that Agricola, as will be suggested later, knew all about Pytheas and what and where Thule was, before he commissioned his naval expedition. I don't see Thule as part of a fictional and schematic mapping out of the world8 nor do I support Hawkes's comment9 that Agricola's sailors would have been told not to expect any Shetlands there. Thule was a reality, not a myth, and the reality was attainable.
Pliny10 and Strabo11 noted Pytheas' comment that Thule was six days' voyaging northwards from Britain. This would be the time it would take to reach Shetland for a native craft, not sail-powered, but paddled or rowed, as Roseman12 plausibly suggested, travelling perhaps from Gills Bay in Caithness to Lerwick, negotiating the mare pigrum et grave13, stopping off at Stroma, Orkney and Fair Isle and tackling the formidable Sumburgh Röst. It is in the framework of Pytheas' six days, as Jones/ Keillar14 stressed, that Thule must be located. Ogilvie15 suggested that Thule was Iceland on the basis that six days' sail would cover some six hundred miles. This is supported by Cunliffe16 who thinks it quite reasonable for boats to achieve some four knots sustained over a period of twenty four hours in Atlantic conditions. This is totally unrealistic. Casson17 refers to only two possibilities: Iceland and Norway, preferring the latter because Iceland is too far north for oats, while Thomson18 ruled out Iceland because it had no bees for the honey drink. Iceland is out of the question19 . Any evaluation of voyaging in what the Romans called Oceanus needs to be made on the basis of contemporary record of the experience. A crossing from the European mainland to the promontory of Belerium in Cornwall, according to Diodorus' sources20, took four days, not one day, as Cunliffe21 imagines. The distance from Land's End to the nearest point of Brittany is some 100 miles, not 400 miles! Caesar's crossing from the vicinity of Boulogne to Dover in 54BC22 took some eighteen hours to cover less than thirty miles, while Pytheas23 said that it was several days' sail from Kent to Gaul. Even in Mediterranean conditions, as Casson24 pointed out, the speed of a fleet could vary from one to three knots according to wind direction. Allowing for twelve hours of continuous rowing or paddling (in Atlantic conditions) at 2.5 knots25 , it would take nearly three weeks to reach Iceland, which no more fits the description of Thule as the most northerly of the British isles26 or the furthest of the islands around Britain27 than it does today. The statement by Rivet/Smith28 that all that can be said with certainty is that Thule was not one of the British isles is one certainty which may be discounted. The most northerly of the British isles is Shetland, and this is as far north as any commissioned navigator was likely to travel. Dion29 tried to resolve the problem of time and distance by suggesting that Thule was Shetland and that Pytheas covered a distance of 550 miles, starting from Galloway, while Bianchetti30, following the same starting point, offers a Norwegian fiord as her Thule. But any crossing to Shetland must surely be calculated by the direct route from the vicinity of Orcas promuntorium (Tarve(du)num, Dunnet Head). It is simply impossible for native boats to cover Dion's estimated 130-150 kms. per day. But at least they both, quite rightly, eliminate Iceland from the equation. Where Norway fits into the scheme will be examined later. Whitaker31 was too quick to dismiss Shetland on the grounds that it is not six days' sail from Britain, without giving any reasons, while André/Baslez32 state authoritatively that Thule is undoubtedly one of the Faeroes or Iceland. Where does such conviction come from? Pytheas was correct in his time allocation which accords well with a Shetland destination. Hawkes's suggestion33 that Pytheas made a five day voyage to Iceland from Westray (?) in the Orkney group is pure speculation. His insistence that Thule is Iceland (based on the migration of whooper swans), that Pytheas heard of it first in North Uist34 and that he sailed in a penteconter35 , is equally speculative and not supported by the requirements of Atlantic tidal conditions. It is curious that Hawkes cites Servius (on Virg. Georg. 1.30), who locates Thule in the north-west, as evidence for Iceland, when Servius clearly locates it iuxta Orcades (close by the Orkneys). Cunliffe36 recognised Pytheas' achievement as outstanding, even if only to the Elbe and Shetland, using local ships. Dilke's comment37 that later writers were misled into assuming Thule to be Shetland, whereas it may have referred to the Faeroes is fanciful, since a voyage from 'Cape Wrath' would require more than six days to cover 300-350 miles, and, in any case, an expedition to Orkney would start from the north-eastern extremity of Scotland, not from Cape Wrath. Strobel's suggestion38 of Fair Isle may also be discounted for reasons which will later become apparent. Burn39 was incorrect in assuming that Foula was not six days' sail north of Britain, especially since he refers to the virtual immobility of boats in the Pentland Firth40 . In asserting the claim for Iceland or Norway he overestimated the capabilities of seafarers in tidal oceans. Are we to assume that Timaeus' island of Ictis41, six days' voyage inwards from Britain, was some six hundred miles away? Geographically impossible. I find it curious that Hawkes42 claims it would take only one day to cross from Ushant to Belerium, yet is happy to allow a six days' voyage from Belerium to the Isle of Wight (on the assumption that this is Diodorus' Ictis).
The passage in Strabo43 where Pytheas is criticised is described by Roseman44 as key in most attempts to identify Thule. She refers to the Broch at Clickhimin (near Lerwick) and to the excavations conducted there by Hamilton45, and she suggests that Pytheas did report details that were certainly present in areas he could have observed. I believe that the possibility is a fact and that Pytheas' Thule (pace Cary/Warmington46) is Mainland, Shetland. Roseman virtually confirms this when she talks of the reasonable certainty that Pytheas got as far as 620 N (Lerwick is 600) and says that his observations could very well have been made in the Shetlands. Dicks47 had already found it improbable that Pytheas went much beyond latitudes where nineteen hours would be the longest day (about 610N). Carpenter 48 says, as far north as the Shetland Islands, and no farther, while Dion 49 observes that in the thought of the ancient geographers Pytheas' voyages did not go beyond the latitude of 61-620. He provides a route from the Outer Hebrides (?) to Shetland and then on towards Norway. This is much more logical, since a circumnavigation of the British isles with a continental destination in the eastern Baltic is hardly likely to include Iceland, for which, pace Cunliffe, no sound case can be made. References to Arctic conditions need not imply that Pytheas travelled that far north and Aujac50 notes that it is probable that Pytheas did not get as far as the frozen sea. One should distinguish between what Pytheas gleaned from British sailors51, about conditions in the far north, what he conjured up out of his own imagination and what he experienced personally. If Pytheas saw native dwellings on Shetland, perhaps the precursors of the brochs, then the more imposing structures, such as the broch at Clickhimin, would not have escaped the notice of Agricola's fleet commander who would have encountered them already in Caithness and Orkney.
To identify Thule as Mainland, Shetland, requires the elimination of all other possible candidates for the honour. Roseman52 , following Müllenhoff53 , proposed that Pliny's Berrice could be another name for Shetland. But Pliny's unspecified sources54 claim that Berrice was the embarkation point for a voyage to Thule, and since any voyage to my postulated Shetland must originate from the south (Scotland and Orkney) or from the east (Norway), I propose that Berrice is Norway55 on the basis that an unattested name of an island, conspicuous for its size and grouped with the Scandiae, requires a Scandinavian, not British, context. Consequently this would be the first recorded contact between Norway and Shetland. Rivet/Smith do not rule out the possibility that Pliny's Bergi and Berrice are Scandinavian, while Stichtenoth56 believes that Berrice is the original Scandinavia, contra Hawkes57 who makes no valid case for identifying Berrice with North Uist in the Outer Hebrides.
Pliny's Acmodae58 and Mela's Haemodae59 are clearly identical and have been erroneously suggested as the Shetland Islands by Harduin60 , Hergt61 , Burton62 , Thomson63 , Hawkes64 , Rackham65 , Rivet/ Smith66 , Cunliffe67 , Breeze68 and Freeman69 . Such views are based on a reluctance to accept Tacitus' interpretation of Thule as part of an ongoing tradition. Mela locates his Haemodae (Acmodae) facing Germany, and this should have given pause for thought. Mela, Pliny's source, lists his island group in the sequence: Orcades (Orkney Islands), Haemodae, and makes no mention of the Hebudes (Hebrides)70. Pliny's sequence is Orcades, Acmodae, Hebudes. Pliny clearly used a different source for the Hebrides and the other islands off the west coast of Britain which are not mentioned by Mela, and through loss of concentration he confused his sources71, and consequently his direction. But Mela, as the earlier source, must be right. It is noticeable that whereas Pliny starts off with the Orkneys Islands, moving south-west through the Irish Sea into the English Channel, thereby eliminating, by reason of direction, the Acmodae as the Shetland Islands, Mela moves south-east across the North Sea towards Denmark, Sweden and the Baltic where his seven Haemodae are now to be located. Hawkes's assertion72 that the Haemodae were north of the Orkney Islands is rather odd, considering that Pliny's direction puts them to the south.
Any interpretation of the work of Pomponius Mela depends ultimately on a ninth century MS of dubious quality73 , generating even more dubious texts, and there is no firm guarantee that one interpretation is better than another. But the most recent editions, commentaries and translations, those of Silberman74 , Brodersen75 , Romer76 and Berry77 follow Vossius78 and Müller79 in punctuating the text to put the Haemodae in the western end of the Baltic Sea, contra Germaniam vectae in illo sinu quem Codanum diximus. Ex iis (S)codanovia magnitudine antestat80. But they don't explain Pliny's anomalous sequence of islands. In weighing the options I prefer the Vossian tradition, although not without reservations81. Mela describes his Scodanovia as the largest of the Haemodae, while Pliny 82 described it as the most famous of the islands in the Codan Gulf, of undiscovered size, without realising that Scadinavia, Scandiae and Acmodae are part and parcel of the same group. This is the consequence of inadequate source collation. Ptolemy fails to mention the Haemodae and (S)codanovia (Pliny's Scadinavia), using the term Scandiae83 instead, and listing only four islands (including south Sweden) in a bay which Pliny describes as refertus insulis84. The location and number of the islands between the Bay of Kiel and Halsingborg suggests that seven is more apt than the four, which include Sweden, and that Mela and Pliny were closer to the correct number. The name Haemodae had become obsolete by Ptolemy's time, just as the Scythians, Pliny's prisca appellatio, are replaced by the Sarmatians. There is nothing to connect the Haemodae with Shetland, Ptolemy's Thule.
The Vatican codex of Mela, which reads that Thule is adjacent (apposita) to the coast of the Belgae is seriously at fault here, since an isolated Thule cannot be adjacent to any coastline, especially one situated south of the Rhine. Vossius 85 correctly emended it to opposita86, which gave Thule its requisite isolation and correct bearing. But no adequate explanation was given for Belgarum which has no geographical relationship to Thule. Abraham Gronovius87 conjectured Belcarum, since Mela had already referred to Scythians called Belcae88. But the Scythians did not live in north-west Europe. Silberman's map is distorted in order to accommodate this conjecture, with the result that Thule and the Orcades appear in the same latitude. The MS error involves only one letter and, although Belgae and Belcae are attested forms, they are geographically flawed. Bergarum, conjectured by Tzschucke89, would fit neatly, and Hawkes90, Rivet/Smith91 and Silberman92 noted it as a possibility. It might refer to the natives of Bergi or Berrice, located in Norway. Shetland would be Bergarum litori opposita because it lay on exactly the same line of latitude as Bergen (600N) at a distance of 220 miles. Dion93 was sure about this and talks of the Bergae as partie du littoral norvégien où se trouve aujourd'hui Bergen (indication précieuse en ceci qu'elle confirme l'identification de Thulé avec les Shetland). Silberman also was right to explore the possibility of a link here, commenting that si l'on rapproche Bergae de Bergen, on est tenté d'identifier Thulé avec les Shetland, en face de Bergen94. But his opposite in this commentary note contradicts the next to in his translation of apposita, and I don't see the point of his reference to Jordanes95 , whose Bergio is used in connexion with the Goths and should relate to Sweden (Scadinavia). I do not subscribe to the view of Whatmore96 and Dion97 that Berrice is a corruption of Nerigos (not a MS reading, but a conjecture of earlier editors), to be linked with Norge, although the end-product is the same. Equally speculative are Hawkes's attempts to link the inferior MS variants, Vergon and Verigon with Rerigonius sinus (Loch Ryan)98. Roseman's suggestion99 that Berrice might be one of the Bergi is reasonable, insofar as there could be a link between the two, but her proposal that Bergi could be an earlier name for the Shetlands doesn't tie in with her Baltic location for the Scandiae which appear in the same sentence. Ninck100 fancifully claimed that Berrice was Pomona (Mainland, Orkney). But this is pure speculation. How can Pomona be the largest island of all, when Dumna (allegedly Lewis101 ?), listed in the same group, is larger than any in the Orcadian archipelago ?
Any account of what is now called Scandinavia should contain some reference to Norway. Ptolemy's failure to produce any names or co-ordinates for Norwegian locations simply reflects the uncertainties about the area in earlier writers, due possibly to inadequate or incomplete exploration (as in the case of Sweden), the reluctance of merchants to reveal their lucrative trade routes, and the absence of Roman military and naval involvement there.
A link with Bergen (Norse Bjørgvin, hill pasture, cf. OE Beorg and modern borough and broch) should not be ruled out. Bergen, like Lerwick with which it now has close maritime links, has a conspicuous harbour, geared to the needs of the fishing industry which would have been a distinctive feature of Norwegian and Shetlandic lifestyles in the Iron Age. The Greek version, Berrice (perhaps a duplication of Bergi102) may well reflect the mountainous Norwegian landscape, as described by bilingual traders operating via the Skagerrak between the Graeco-Roman and Celtic worlds. Yeames103 refers to an archaic Greek statuette, allegedly excavated at Bergen, and suggests that it reached Bergen as a result of the trade in amber. He goes on to say that from Jutland to the south of Norway is not an incredible voyage for a sailor in the sixth century. Roman imports into Hordaland may well have come in through Bergen104 . Dion105 proposed that Pytheas on the final stage of his journey proceeded from Shetland to Norway, where he landed not far from Bergen. This might explain the origin of the claim put out by Pliny's unnamed sources who would have assumed that if one (Pytheas ?) could travel from Shetland to Norway, then the reverse was also possible. But Dion's theory of two Thules, a smaller one, Shetland, and a larger one, Norway, can hardly pass muster 106 Are we to assume that when Pytheas reached Norway, he found the name of Thule attached there as well ?
If Sweden can be treated as an island, why can't Norway (Berrice) be regarded by Pliny as the largest of all (maximam omnium)107? The points I wish to emphasise are that Thule and Berrice cannot both refer to Shetland, that neither the Haemodae nor the Bergi should be identified with the Shetland group, and that Iceland cannot be listed among the British Isles any more than Cyprus can be called the furthest of the islands around Greece.
Send email to the author: email@example.com
INTRODUCTION | PART ONE: Shetland: The Classical Geographical Context
Copyright © 2002 by Stan Wolfson. All rights reserved.