16 Cunliffe 2001b, 126. Selkirk (1995, 183) had earlier conjured up the same figures to produce less than one day's sail from Caithness to Shetland, without evaluating what type of boat might have been used or what conditions were like in the Pentland Firth or Sumburgh Röst.
25 Cotterill (1993, 227-228) produces sensible figures for North Sea navigation in the late 3rd century AD, suggesting that at 2.5 knots and in favourable weather a crossing between Bruges and Kent would take at least fifty two hours, with at least sixteen hours for a return crossing at the Straits of Dover.
26 Str. Geog. 2.5.8. I see no inconsistency, pace Roseman (1994, 132) in Strabo's two separate statements that Thule was six days' voyage from Britain (Geog. 1.4.2) and that it was the most northerly of the British Isles. This collective data would corroborate the identification of Thule as Shetland.
29 Dion 1965, 445-446, 1977, 201-204. Dion bases his calculations on Ptolemy's map which was composed in the Hadrianic period, four and a half centuries after Pytheas. So how could Pytheas possibly have assumed a turning of Scotland, a distortion caused by Ptolemy's faulty correction of Marinus' coordinates (cf. Jones/Keillar, 1996, 48)? As Rivet (1977, 48) pointed out, no competent sea captain ever reported that the general trend of Scotland was from west to east instead of from south to north. In fact the earliest maps of Britain produce a triangular shape, as may be seen in Diodorus' account (Bibl. 5.21), derived from Eratosthenes. Pytheas' measurements would have provided the data for Eratosthenes' map (Ogilvie 1967, 165; cf. Jones/Mattingly 1990, 18, map 2:3). There is nothing in Tacitus' account (Agr. 10.3) or in any other previous account to suggest that Britain took a ninety degree turn along the Wear-Eden axis (cf. Strang's readjustments 1997, 12-13).
35 Hawkes 1975, 44. Roseman (1994, 149) points out that such a vessel is unlikely to have been chosen by a sensible Greek, planning to venture into completely unknown waters for an extended voyage of exploration.
42 Hawkes 1975, 29. Cunliffe (2001b, 77) assumes that Mictis and Ictis are different, locating the former between the Loire and the Gironde, and applying the latter to Mount Batten, neither of which is six days' sail inland (introrsum?) from Belerium. Rivet/Smith (1979, 487-9) produce a sound etymological case for the equation Ictis = Vectis. The geological possibility of a Solent causeway was proposed by Reid (1905, 284). But this has now been scientifically disproved. It is quite impossible to reconcile Diodorus' off shore island with the Pliny/Timaeus location where there is no reference to Belerium. The meaning of introrsum (up-channel?) is open to question. The possibility of confusion between the six days' sail from Britain to Thule and the six days' sail from Britain to the Isle of Wight does not negate the validity of the travelling time. If Pliny conflated his sources he merely erred in not substituting Belerio for Britannia when giving Ictis as the destination.
46 Cary/Warmington 1929, 37. I find it rather curious that Cary in his revised version (1964, 288) of Tozer's History of Ancient Geography retains Tozer's Thule = Mainland theory which he and Warmington were previously happy to rule out.
48 Carpenter 1966, 183. He argues (180) against Iceland on three accounts, a) there is no reason to suppose that Iceland was inhabited before the Christian era, b) there were no natives in Iceland with whom Pytheas could have communicated, c) there is no likelihood that Pytheas would have abandoned his investigation of the Atlantic trade route to steer a course in the opposite direction.
51 Cunliffe 2001a, 93. Pytheas certainly obtained information from others about areas he hadn't visited, as is clear from his reference to hearsay (Str. Geog. 2.4.1) and to Geminus of Rhodes who quotes Pytheas' observation that the barbarians have shown us where the sun goes to sleep...cf. Dion (1977, 204-207). Phillips (1969, 194) suggested that Pytheas may have obtained his information on the Arctic Ocean from Celtic or Germanic sources.
54Plin. HN. 4.104, sunt qui et alias prodant, Scandias, Dumnam, Bergos, maximamque omnium Berricen, ex qua in Tylen navigetur (some authorities mention other islands also, the Scandiae, Dumna, Bergi and Berrice, the largest of all, from which one sails to Thule). I find it curious that Pliny should give a British context to a group of islands which are essentially Scandinavian. Thomson's assumption (1948, 246) that Pliny thought these were British islands was questioned by Roseman (1994, 94) who talks about Pliny's disjointed account without explaining how a group of Scandinavian islands suddenly appears in a British context. Rivet/Smith (1979, 42) likewise offer no explanation. It was the British context which led Hawkes to assume a reference to the Outer Hebrides, where the Scandiae couldn't possibly be located. In this case Dumna is the odd one out and reflects Pliny's misinterpretation of different sources. It was this sort of confusion which led Pliny to assume that Scandinavia and Scandiae were not the same (cf. Schütte 1917, 136), since he was following Mela for the former, and a separate source for the latter. If a reference to Norway (Berrice, as I maintain) is assumed among this group, then the group is misplaced during textual transmission and belongs at the end of sections 96-97 where Pliny describes the area extending from the Vistula to the Ems. Once Pliny had mentioned Sweden he should have referred to Norway, misplaced after his account of Britain through confused collation of different sources. Pliny's geographical description of the extera Europae (the outer parts of Europe) goes anti-clockwise: the Polish coast, Denmark, Sweden, Norway, Germany, Gaul, Spain. The direction for mariners is explicitly stated, litus Oceani septentrionalis in laeva, donec perveniatur Gadis, legendum (You must coast along the shore of the northern Ocean towards the west until you reach Cadiz). The account of Britain and its adjacent islands (102-104) is a digression from the description of continental Europe where we find ourselves at the start of c. 105. The section sunt qui et alias. Cronium appellatur would fit very comfortably at the end of c.97.
55This would negate Dion's theory (1966, 211) of the probability of an extension of the ancient name for Norway to the Shetland archipelago. There is no reason to assume that Shetland and Norway ever had the same name, and no basis for claiming that Norway was called Thule, especially if it was called Berrice. One thing is certain, namely that Thule and Berrice are both Greek names, and they may both derive from Pytheas account, especially if, as Dion suggests, Pytheas landed in Norway. Berrice is not recorded by any other source. If it is Norway the vagueness of its location is understandable, since Norway was unexplored territory and consequently absent from Ptolemy's listings.
66 Rivet/Smith 1979, 241.Their failure to translate (76) the whole sentence in Mela is misleading. Roseman (1994, 90) likewise does not extend her phrase beyond contra Germaniam vectae. Burn (1969, 37) also fails to examine the entire sentence, assuming a confusion with Ptolemy's Aebudae. Since when were the Hebrides located in the Baltic?
70 The appearance of the Hebrides in Pliny (the first time they appear in literature), but not in Mela, may have been stimulated by reconnaissance of the western isles during the northern campaigns of Cerialis who may deserve more credit than he is usually given. Before Pliny's time Roman knowledge of the smaller islands to the west of Britain did not extend further north than the Isle of Anglesey (Mona Rivet/Smith 1979, 419). Pliny's earlier Greek sources, clearly responsible for naming the western isles (e.g. Andros) , may have required confirmation through Roman naval involvement.
71This would easily occur if Pliny was faced with two words of similar spelling, Haemodae and Haebudae (or Aemodae and Ptolemy's Aebudae, cf. note 66). Rivet/Smith (1979, 241) did not rule out the possibility that the two groups were the same. But their unquestioning acceptance of Ranstrand's reading in Mela was never a sound basis for this theory.
81 There is no doubt that the sentence is corrupt. Salmasius and Vossius read Vecta. But the Isle of Wight doesn't face Germany, and this reading is ruled out by ex iis which can only refer to a plural antecedent, i.e. Haemodae. Grienberger's exit (1921, 1199) is totally unnecessary, while Frick's suggestion of eximia merely duplicates antestat. Parthey began his next sentence with in illo sinu, but ex iis is not likely to appear in the middle of the clause. Ranstrand wasn't prepared to commit himself and placed ex iis between obeli. I don't see any MS reading Vecti, as claimed in RE Vll 2182.
84 Plin. HN. 4.96 crammed with islands. Pekkanen (1987, 552) claims that the sinus Codanus comprises the whole Baltic. But the whole Baltic is not crammed with islands. Silberman (1989, 579) points out that the sinus Codanus can only represent the south of Sweden and the area between the Skagerak and the bay of Kiel.
86 I find it strange that modern editors have not adopted the Vossian emendation, while translators (e.g. Rivet/Smith 1979. 42) persist in rendering the participle as opposite. Apposita means adjacent, adjoining, next to. Clearly editors have been misled by the incorrect definition in OLD, as meaning opposite which has no more justification than the assumption that the English words opposite and apposite are synonymous. TLL is correct in equating apposita with vicina. How can Thule, the furthest of islands, and six days' sail away from Britain, be adjacent to any coast?
94 Silberman 1988, 288 if Bergae is linked with Bergen, one is tempted to identify Thule with Shetland, opposite Bergen. This observation is inconsistent with his translation près du littoral des Belcae.
101 It is generally assumed that Dumna is Lewis and Harris in the Outer Hebrides. But Ptolemy's coordinates place it on the same longitude (30o) as Orkney and Shetland. Why do Pliny's sources place it in a Scandinavian context? The Ravenna Cosmography provides a list of the western isles, a list which includes Skye (?), Canna and Bute, but ominously makes no mention of Dumna, Bergi or Berrice. Strang (1997, 13), in referring to the turning of Scotland, says that Ptolemy dropped off the larger islands at convenient (?) places; Thule at 35o and the Orcades at 55o, trying to maintain some sort of visual integrity (?) in their relative positions above Scotland. But Lewis is larger than Shetland and surely merits the same consideration. Why does Shetland receive five sets of coordinates from Ptolemy, while Dumna gets only one? Why should Dumna be moved, when Thule and the Orkneys are not? Why can't Dumna maintain some sort of visual integrity and be identified with Hoy? Or why can't it be moved further to the east in Scandinavia where Pliny apparently locates it? Ptolemy makes no mention of Berrice or Bergi because exploration had not embraced Norway (cf. Geogr. 1.5), and such places were hearsay (sunt qui prodant..).
102 Whether Bergos is preferable to Vergos, the reading of the superior MS (A), is open to question. B and V are frequently interchanged in proper names during transmission of texts, cf. Batavorum/Vatavorum (Agr. 36.1), Boudicca/Voadicca (Agr.16.1). Stichtenoth (1959, 84) suggested a possible derivation for Berrice from the German bär (bear), assuming a connection with the Greek arktos (north) and the constellation of Ursa Major.
107 Selkirk's argument (1995, 183) that Norway is not an island is pointless. Sweden is not an island either, but Pliny believed it was. It is the ancient conception which matters, not the modern reality.
108 For the history and interrelation of the manuscripts, cf. Robinson 1935, 1-78, Lenchantin de Gubernatis 1945, v-xxii; De Saint-Denis 1956, xxiii-xxx; Mendell 1957, 257-293; Koestermann 1964, v-xxvi; Ogilvie 1967, 80-90; Till 1979, 7-10; Delz 1983, i-vii; Turner 1997, 587-589.
110 E is the Oxford abbreviation for the codex. Further refinements are added by Murgia (1977, 324 n.2). Till (1979, 7-10) uses H (Hersfeld) for the Caroline section and E for the fifteenth century transcriptions. Delz (1983) assumed that the Hersfeld and Aesinas were the same. That they were different was proposed by Mendell (1949, 134-135; 1957, 257-293) and Schaps (1979, 28-42) and supported by Winterbottom (1983, 411). This was challenged by Murgia and Rodgers (1984, 145-153, cf. Magnaldi 1997, 133). For the fate of the codex, cf. Schama (1995, 75-81) and Niutta (1996, 172-202).
114 Not a complete circumnavigation of the island, as suggested by Forni (1962, 125). It simply means sailed round. The Usipian mutineers had done the same in reverse direction (circumvecta Britanniam, Agr. 28.3), cf. Liv..H. 10.2.4; Plin. HN. 2.167). The suggestions of Ogilvie (1967, 282) and Durant (1969, 43) of a voyage round the west coast via the Irish Sea and the Channel to a southern base for the winter defies both logic and strategy. Nor is there any reason for the fleet to complete a circumnavigation which would merely take it back to its starting point in north-east Scotland. For an evaluation of the problems here, cf. Burn 1969, 59; Strobel 1987, 202 n.19; Hanson 1991a, 140-142. I find it inconceivable that after its intense campaigning along the east coast of Scotland and its voyage to the Orkneys and Shetland the fleet should be commissioned to sail all the way round Britain, a voyage which would not be completed before January, AD84, by which time most of the fleet would be at the bottom of the Ocean.
115 This is not strictly true, unless the previous exploratory mission of Demetrius involved only a few ships rather than a fleet. In any case assumptions that Demetrius' voyage involved the western isles (Ogilvie, 1968, 33) are unfounded. Plutarch (De Defect. Orac. 18, 419e) merely refers to deserted islands scattered around Britain and that Demetrius visited the nearest of the inhabited ones. The islands referred to may have been the Orkneys, not the Hebrides. Tacitus makes no mention of the Hebrides. They were irrelevant since they did not figure in Agricola's plans for conquest. But an exploratory voyage to the Orkney Islands may have been the prelude to their conquest. Demetrius' voyage should be linked to naval intelligence. He was only part of a team. I can't see Domitian commissioning him as a prototype of Columbus or Darwin. There is too much self-advertisement and fiction in Demetrius' story-telling to make it totally convincing, let alone put him in the western isles.
118 Stephenson 1894; Acheson 1938; Mattingly 1948: Burn 1969 (clearly sighted); Dion 1977 (en vue); Rivet/Smith 1979; Till 1979; Handford 1986; Spaltenstein 1990); Hanson 1991a; Clarke 2001; Cunliffe 2001b; Freeman 2001 (saw). Sighted would be conspecta est.
119 Seen/sighted/descried from far off/at a distance/in the distance. Orelli 1848; Church/Brodribb 1869; Andresen 1880; Davis 1892, Furneaux 1898; Pearce 1899; Anderson 1922: Gudeman 1928; Thomson 1948; Smith, 1987; Aujac 1988; Benario 1991; Wijsman 1998. Such sentences would be equated in Latin with procul visa/conspecta est.
125 There is no evidence to suggest that Marinus (and Ptolemy) obtained any coordinates on Thule directly from Pytheas, although, as will be suggested later, such data could have been available to Agricola. The necessary bearings for Thule would in fact be confirmed by Agricola's navigators, and subsequently made available to Marinus. Carpenter (1966, 183) makes a sweeping assertion that it is an unchallengeable inference that Ptolemy's data for the location of Thule must go back to Pytheas, since no one else in antiquity ever claimed to have visited that island. I hope that this assertion will soon be laid to rest.
136 Pytheas was probably funded by a Massilian merchant consortium which would certainly have been cautious in screening any information which Pytheas brought back. Any suggestion that Pytheas was sponsored by Alexander the Great does not consider the immediate needs of Massilia (cf. Roseman 1994, 154), but merits some consideration in so far as it might explain why the Carthaginians, in fear of Alexander, would allow Pytheas a passage through the Pillars of Hercules (if his travels were entirely by sea) into the Atlantic whose routes they monopolised. I would see Pytheas' mission as primarily commercial, then scientific. Pytheas would only have published what was in the parameters of his remit. Massilia and Alexander did not share the same objectives.
137 The protection of Massilian interests may be seen in Scipio Aemilianus' failure to obtain information about Britain from those he interviewed at Massilia (Str. Geog. 4.2.1 cf. Chevallier 1984, 91).
151 Set = sed. This interchange of consonants is quite common, cf. Agr. 18.1, 24.2, aut for haud; 18.2, 21.2, 22.4, 29.4, aput for apud; 23, 27.1, 28.3, 29.4, 30,3, 33.2, 34.1, 37.2, adque corrected to atque. A similar case involving sed is found at Agr. 10.3 where e reads unde & in universum fama est transgressis unde & universis fama sed. The awkward est before transgressis is clearly, as pointed out by Wölfflin (1867, 144) a metathesis of set (= sed). We have two conjunctions, when only one is required.
153 It has been suggested that campaigning summers are defined by the spring and autumnal equinoxes Raepsaet-Charlier 1991, 1843 n.165). But there is no need for such rigidity which implies a two-season campaigning year of winter and summer. This contradicts Tacitus' statement that summer was already over and winter was approaching. There must be a period between the end of summer and the approach of winter. The interpretation of winter quarters is a point in question, since troops could be sent to hibernacula at the height of summer (Ann. 2.23.1). Strobel (1987, 203) set the battle of Mons Graupius in AD84 (?) at the end of July or early August and the voyage of the fleet anywhere between mid-August and the end of September. If the battle occurred in the far north well beyond Moray, there is no reason why the fleet should not have started out in early August. This would leave ample time for operations in the Orkneys and Shetland before returning to a winter base. I would incline to the view that a campaigning summer would end whenever it suited the governor.
154 The Roman fleet would never have had the time to overrun more than a handful of some seventy islands. The conquest of offshore islands is given an aura of achievement scarcely merited. The conquest of the Isle of Wight is overplayed (Suet. VC. Vesp.4.2) as is the reduction of Anglesey (Tac. Ann. 14.290). But those two islands did not present the Romans with the type of obstacle which the Orkney brochs (over a hundred) would present to Agricola's marines.
155 Furneaux 1898, 149. Vires implies more than just naval strength. Manpower would be required for armed landings both on the Orkneys and on the north coast of Scotland. The fleet was only part of Agricola's resources (partem virium, Agr.25.1). Church /Brodribb (1869, 82) correctly translate as military force, while Sleeman (1914, 119) refers to 'troops' to make landings.
158 Burn (1969, 59) states that the translation on the face of it is not difficult and goes on to produce a version based not only on rejected readings (omni), but on a misinterpretation of the uninterpretable.
162 The end of summer does not necessarily imply an end to campaigning, cf. Agr. 18.2 quamquam transvecta aestas ire obviam discrimini statuit. Likewise Ostorius Scapula began his campaign in Britain coepta hieme (Ann. 12. 31.1).
163 Agr.25.1 the hardships of storms and waves. This triumph over the Ocean is significant in view of Tacitus' observation that nusquam latius dominari mare (nowhere does the sea rule more widely, Agr. 10.6).
164 I don't support the curious interpretation of Agricola's aims by Romm (1992, 149) that a boundary imposed by nature must forever be closed and divine anger would destroy any further expeditions sent out to explore it.
193 Hanson ibid. nor does reference to a base on the south coast of England seem entirely relevant to Tacitus' narrative at this point. Hind (1974, 288) says that it would not be unlikely that the fleet proceeded at a leisurely fashion southward around the Romanized part (proximo?) of the province to its place of origin. He doesn't take into consideration the season, the time factor, weather conditions, tides or naval strategy, apart from failure to suggest the place of origin. Was it to the place where they started the circumnavigation , or to some base on the south east coast?
212 Agr. 38.3. Ogilvie locates these new nations between Moray and Aberdeen. Breeze (1990, 59) suggests a return route through Glen Garry and Strathtay. Jones/Mattingly (1990, 76-77) note the mountain route taken by Edward l. Hanson (1991, 140) lists plausible alternatives for a marching route south.
221 Scholars have attempted to bridge the gap by linking unde, not with redierat, but with a participle. Furneaux, following Orelli and Ernesti, links unde with the unacceptable marginal reading lecto and says, the sense being 'quo' litore (sic) inde lecto, redierat. He justifies this with the comment that clearness is sacrificed to conciseness (1898, 150). Anderson similarly proposes that 'unde' is taken only with 'praelecto' (1922, 141). There is no parallel in Tacitus for such a use of unde in a subordinate clause. Ogilvie, unhappy with the suggestions of his Oxford predecessors, says that 'profecta' is to be understood with 'unde' (1967, 283), a concept taken up by Heubner (1984, 112) and Borzsák (1992, 101), while De Saint-Denis and Zúñiga actually introduce profecta into their texts, unde profecta proximo Britanniae latere lecto omni redierat, thereby producing two participles instead of one and resurrecting the long discarded readings of Furneaux. No one has yet been able to explain redierat. Madvig (1873, 569) thought this reading to be unsustainable (a view which I endorse) and recreated the text to read unde proximo anno latere lecto omni reditura erat. This is even more unsustainable. I see no point in introducing anno in reference to some future event outside the term of Agricola's governorship. The imperfect subjunctive, another of his suggestions, would imply that the fleet commander had been given instructions to return. What else was he supposed to do? Both before and after the discovery of E the problem has defied solution because scholars have insisted on retaining the concept of return. Attempts to relate redierat to the maritime events of AD81 (Burn 1969, 53) and to suggest that there were two separate half-voyages (an unnecessary convoluted explanation Hanson 1991a, 141) would require the insertion of biennio ante, and, as pointed out by Maxwell (1989, 70), the sense of the passage demands that a complete circumnavigation (of Scotland) should have been made in the final year of Agricola's governorship. The fact that no destination is given in this clause should have aroused suspicion. One doesn't return from without returning to. Redierat lacks a destination, and Andresen's comment (1880, 209) that qui portum tenet, rediit overlooks the fact that unde cannot imply a destination.
230 The middle verb equivalent praevehi (to sail past, on) illustrates the point, cf. Tac. Hist. 2.2.3, oram Achaeae et Asiae ac laeva maris praevectus, Rhodum..petebat. Titus had completed his coastal voyage and was now out in the open sea on his way to Rhodes.
232 Liv. H. 22.20.7, praelecta est ora (the coast was skirted). This was originally suggested by Madvig for the meaningless periectas oras of the earliest MS (codex Puteanus). He failed to note that Livy always uses legere. Weissenborn's later conjecture, praevecta est oram (sailed along the coast), perhaps based on praevecta est orâ (sed) is more convincing, since a middle verb, followed by an impersonal verb, in Ebusam insulam transmissum (a crossing was made to the isle of Ibiza) is followed shortly afterwards (22.20.10) by a similar pairing, inde flexa retro classis reditumque in citeriora provinciae (then the fleet turned back and returned to the nearer parts of the province).
233 Prop. Eleg. 1.8.19, praelecta Ceraunia. (Ceraunia skirted). But the earliest MS (codex Neapolitanus) reads praevecta, a reading which most editors have retained. Post victa, suggested by N. Heinsius (cf. Sil. Pun. 15.509-510), is preferable to post lecta (Mueller), unparalleled with this meaning. The host of suggestions, originally proposed for the true Propertian reading, should have alerted scholars to a reconsideration of E. One thing is certain; there is no evidence in the MSS for the maritime use of praelegere before the second cent. AD.
237 Agr. 29.2 the fleet had been sent ahead to cause terror. Onesicritus, the helmsman of Alexander he Great, is referred to by Seneca (Ben.7.2.5) as praemissus explorator commissioned to look for wars in the unknown sea, while exploring the route between the Persian Gulf and India.
238 Ogilvie 1967, 283. Lund (1981, 124) equates omnis with integra or incolumis. This would imply from the examples in the following note that Caesar was repeating himself. Omnis means all, not intact, as Hind (1974, 285) translates. If Tacitus had intended to say all the fleet he would have written omnis classis. Forni (1962, 224) cites Horace (Carm. 3.30.6), non omnis moriar? Does Horace really mean that he will not die intact? For the same reason Heubner's example from Quintilian is no parallel.
239 Caes. BG. 4.36.3, 5.23.6. In both instances omnes incolumes is plural and directly adjacent to its noun or pronoun. In Tacitus omnis appears to be singular, totally isolated from classis and used predicatively; quite un-Tacitean.
243 Ogilvie 1967, 283 (cf. Heubner 1984, 101). Omni should precede proximo (cf. omnis propior sinus, Agr. 23). But in any case it adds nothing to the sense of the translation by following the past participle. Burn (1969, 59) preferred omni in his translation, after the exploration of all the adjacent side of Britain. Exploration may well have been on the agenda of the fleet, but praelegere does not mean to explore, as Acheson (1938, 83) also translates, nor, as Benario (1991,51) suggests, to reconnoitre. An explanation of adjacent is also required.
244 Persson 1927, 83-84, die Flotte hatte die nördliche Seite Britanniens umsegelt, und zwar die ganze. His citation of the earlier reference to the voyage, hanc oram..circumvecta (Agr.10.4)shows that omni would in fact be superfluous.
247 Schoene (1889, 37) is the only editor, as far as I am aware, to assume that omnis required a noun next to it. He proposed omnis reditus erat, involving a shorthand version redit'erat. But this is virtually meaningless.
248 This solves Ogilvie's dilemma on the relevance of proximo. It is in the Sutherland-Caithness area that the battle of Mons Graupius should be located. Keppie (1980, 84) favours a fairly northerly site because it only made sense to send the fleet north, if the starting point was well up the NE coast, thus minimising its voyage. I would support this, although I prefer his reference (1986, 11) to Tacitus' implication that the battle site lay close to, and perhaps even in sight of, the sea, and far away in the very north of Britain.
249 Although the neuter plural omnia is regular usage in Tacitus for the subject and object form, omnis res (pl.) cannot be ruled out for the object form (the nominative omnes res is understandably avoided on grounds of euphony). I maintain that this phrase combines both res prosperas/secundas and res adversas. Statistically the neuter plural subject and object forms prospera/secunda and adversa are five times more common in Tacitus than their synonymous phrases.
252 Courtney (1996, 148) notes that this passage is very Tacitean in content and cites, nox clara et extrema Britanniae parte brevis (the night is bright and short in the most northerly part of Britain, Agr. 12). If the night is short in the most northerly part of Scotland, then Juvenal's superlative form minima must refer to the area even beyond that. There is no need to replace minima with nimia, as Martyn (1996, 80) proposes on the grounds of pointlessness of Britons being 'minima contentos nocte'. I see nothing in the Agricola which in fact supports 'nimia'(Martyn 1974, 345). There is no reason to suppose that Juvenal was being ironic. People living in the Shetland Islands are more than happy with the shortest nights and Juvenal was aware of it. The difference between brevis and minima is a matter of latitude; the further north, the shorter the nights (5 hours in Shetland).
253 Courtney, ibid. The evidence for a Roman presence at Drumanagh in the Irish Republic at the end of the first century has yet to be evaluated, as the site is currently sub iudice. From such a site and its links with the interior and with merchants crossing the Irish Sea would come the sources for Ptolemy's map. Gudeman (1898, 302-3 was convinced that Tacitus was referring to an invasion of Ireland at Agr. 24 on the basis that transgressus referred to the Irish Sea. This theory was ably dealt with by Haverfield (1899, 302-3). A link between Juvenal and Ireland in AD81 was suggested by McElderry (1922, 154) who proposed that the satirist may have served in northern Britain at the time. Higgins (1998, 407-8) offers some valid reasons why Domitian would not have sanctioned an invasion of Ireland.
255 For the dating of Agricola's governorship from AD77-83, cf. Büchner, 1960, 172 ff., Gallivan 1981, 189, A.R.Birley 1981, 73-81, Eck 1982, 299-300, Campbell 1986, 197-200, Maxwell 1990, 114-115, Hanson 1991a, 40-45, Raepsaet-Charlier 1991, 1842.
257 The bracketing of Ireland, the Orkneys and Shetland is repeated by Claudian, describing in the fourth century the activities of Count Theodosius in the far north, maduerunt Saxone fuso/Orcades; incaluit Pictorum sanguine Thyle;/ Scottorum cumulos flevit glacialis Hiverne (the Orkneys were drenched with Saxon slaughter, Shetland grew warm with the blood of the Picts, ice-bound Ireland wept for the heaps of the Scots). De lV Cons.Hon. 31-33, cf. Dion 1966, 210. In this parallel Thyle cannot be a general reference (pace Wijsman) to Britain which Claudian had already referred to in line 28. But it does illustrate how Roman military involvement in the far north becomes virtually a topos.
258 Britanni can be anyone living in the British Isles. It is regularly used by Tacitus to describe the Caledonians, while Diodorus (Bibl. 5.32.3) employs it to describe the Irish (cf. Freeman 2001, 35- 6) It is therefore equally applicable to the people of Shetland. Shetland is a British island. Its inhabitants are Britons.
260 This point needs to be qualified. Agricola was recalled to Rome at the end of AD83. But the Romans maintained a presence in Caledonia until AD87, and there is no reason why Agricola's successor should not finish off what Agricola began and complete the conquest of Shetland. In which case references in the Silver Age poets to the conquest of Thule under Domitian may well be true. Tacitus would obviously not mention the achievements of Agricola's successor.
262 Cf. in ultimos orbis Britannos (into the Britons, furthest in the world) Hor. Carm.1.35.29-30, divisos orbe Britannos (the Britons separated by the world) Virg. Ecl. 1.67, cum cetero orbe Vespasianus et Britanniam recuperavit (Vespasian recovered Britain also together with the rest of the world) Tac. Agr. 17.11, dierum spatium ultra nostri orbis mensuram (a length of days beyond the measure of our world) Tac. Agr. 12.3. Statius clearly differentiates between Britain and Shetland.
267 Momigliano 1950, 42. His views have greatly influenced later generations of scholars who have taken them as gospel. His argument was countered by E.Birley (1953, 19) who did not rule out the possibility of Vespasian being commissioned by Plautius to operate in the north. But with no reference to this in Tacitus, Suetonius or Dio the matter is still open to question. Dihle (1994, 176) notes Silius' reference to the conquest of Scotland under Domitian. But he doesn't explain the Silian context.
273 Lucan (BC. 6.68) is the first writer to record the Caledonian Britons. He is clearly distinguishing between the extremities of Britain, Richborough in the south and Caledonia in the north. But where did Lucan obtain the word Caledonian? Is he referring to the fact that the Romans had begun their invasion at Richborough in AD43 and at the time of writing they were involved in Caledonia? His Civil War epic was composed c.AD64, at the time when Trebellius was governor of Britain. It is hard to believe that in six years of governorship Trebellius was not involved in Brigantia. The Brigantes may well have capitalised on the governor's dilemma between reconstruction and the need for vigilance in the north. Perhaps Shotter's putative Neronian fort at Chester (2000, 37) comes into the equation here. Lucan's reference may be an echo of what was happening in Britain at the time. Where do the Caledonians fit into the scheme? Is it possible that Nero after a decade in power conceived the plan for the total conquest of the island (which at one time he had considered abandoning, Suet, Nero 18.1) thereby completing what his stepfather had begun (echoes of Augustus and Julius Caesar)? Perhaps Trebellius merits better publicity than that accorded by Tacitus. We know that Didius Gallus was involved with the Brigantian problem, and that the tribesmen received outside support, most probably from Scotland. The Romans must have maintained some war-footing in the north. The absence of information on Brigantia in Tacitus during the ensuing decade should not be taken to assume any inactivity under Venutius, nor that Trebellius was unaware of the support the Brigantes were getting. Might not the Caledonians be behind the pressure exerted on the lowland tribes to commit themselves to the nationalist cause? Lucan 's reference merits close attention.
276 The competitive nature between the two dynasties may be seen in Vespasian's dream (Suet. VC. Vesp. 35) of the balance, where Claudius and Nero appear in one side of the scales, while Vespasian and his two sons stand on the other.
285 Galimberti 1996, 74. This is not an entirely original theory, since Torrentius in 1574, in order to accommodate Pliny's location of Vectis in the Irish Sea, suggested that Vespasian had campaigned there or had stormed some other island further to the north, although Torrentius made no connexion with the Orkneys in citing Eutropius.
287 Eutr. BHR 7.13.3. Müller (1883, 104) charged Eutropius with wishful fantasy on the Orkney episode, while Bird (1993, 113) states that Eutropius' claim seems improbable. But neither stopped to consider where such an odd reference might have originated. Armit (1998, 104), in order to put the Orkneys in a southern context, suggests close links perhaps through marriage or military alliance between 'kings of Orkney' and tribes far to the south. Wainwright (1962, 65) claims that chieftains of the Orkneys made formal submission to Claudius. Chieftains of the Orkneys is a very vague expression, and there is no record of this in Eutropius. Do we assume that their names appeared on the arch of Claudius, whose reconstructed inscription by the antiquarian Gauges de Gozze included the colourful allusion to the Orkneys (Barrett 1991, 11-12) so as to read GENTESQU.E(XTREMARUM ORCHADUM). They probably wouldn't even know what was happening in the remote south, or even if they did, they wouldn't be bothered. This possibility is no more plausible than Stevens' theory (1951,8) that some adventurous sea-captain put into Camulodunum and made submission to Rome in the name of his people. The Caledonians, according to Calgacus (Agr. 30.2-3), had been confident that their remoteness would protect them from Roman domination as it had done for forty years, and I've no doubt that Agricola's interrogation of prisoners-of-war elicited such sentiments. But to assume that a delegation came all the way from Orkney to surrender to Claudius beggars belief. Why didn't the Caledonians send a delegation at the same time? Fitzpatrick (1989, 24-33) links the discovery of a specific type of amphora sherd at the Broch of Gurness with a Claudian clientela. But the link is tenuous and the evaluation of the literary evidence is unconvincing , relying too much on Momigliano, while the arguments against Maxwell are weak. I see no grounds for belief that the panegyrics provide inferential support for the Claudian submission of the Orkney Islands. Tacitus and Juvenal knew who was responsible for the conquest of the Orkneys and I don't believe they had at the back of their minds this earlier claim of annexation (Stevens, ibid.). Braund (1996, 150, 195) believes that a Claudian contact with the Orkneys was Flavian propaganda. I suggest that any link between the Orkneys and Claudius is the consequence of confusion between sources. This can be explained by textual analysis and comparison. Eutropius made a mistake, and Orosius and the others who followed him, fell into the same trap. Eutropius produced two separate accounts, one for Claudius, the other for Vespasian, when in fact the submission of Vectis (Wight), which he has taken from Suetonius, should bridge the roles of emperor and legatus legionis. It is worth noting that Eutropius gives Vespasian the initiative for the Vectis episode, without mentioning Claudius being in Britain, whereas Suetonius (Vit. Vesp. 4.2) makes the crucial qualification, partly under the leadership of Claudius. I am inclined to believe that Eutropius, following two different sources, one for each emperor (his source for Claudius produces a reference to Cn.Sentius Saturninus, not found elsewhere in a British context), has produced two different versions of the same event, the surrender of the Isle of Wight. Consequently any surrender of the Orkneys is false. Stevens (ibid.) tried to establish a Link between Claudius and the Orkneys via Mela on the basis that he is the first to mention the Orcades, whereas Diodorus and Strabo knew nothing about them. But the fact that the two Greeks don't mention these islands is no evidence for ignorance. Orcades is a Greek word and probably derives, like Thule, from Pytheas; it was available. Suetonius made no mention of Vectis in his Life of Claudius, preferring to reserve it for Vespasian, and no mention at all of the Orcades whose remoteness would surely have merited a reference in the biographer and in Cassius Dio and most probably in Seneca's Ludus de Morte Claudii. Suetonius and Orosius were correct in referring to only one instance of surrender. Historical reality and common sense must tell us that the reduction of the Isle of Wight is the natural consequence of campaigning in southern England, and that to link Claudius with the Orkneys is extreme. So how did Vectis come to be confused with the Orcades? The answer may lie in Agricola's conquest of those islands, a detailed report of which would have included the names of all the islands between the Caithness coast and Orkney. Tacitus was not interested in recording geographical trivia. But such details found their way to Marinus, and from him to Ptolemy. I believe that Vectis is the name given to one of these islands and was so recorded by Ptolemy whose text was subjected to numerous alterations before it reached the hands of the Ravenna Cosmographer and the scribes of the 13th and 14th centuries. Ptolemy's word for Wight is Ouektis (Geog. 2.3.14). But between Ptolemy's reference to Orcas promontory and the Orcades islands we find the mysterious Sketis, the reading of X (Vaticanus Graecus 191), which is assumed to be the Isle of Skye and is so identified in the RC whose sequence of islands defies any explanation. Yet Pliny doesn't mention Skye in his list of the western isles, and the location of Vectis in the Irish Sea by a man who knew Vespasian well has never been accounted for, while RC lists Vectis before Malaca ( Mull?). The belief in an island called Vectis located in the north appears to be a reasonable assumption. Its precise location is problematic. The reading of Ptolemy in U (Vaticanus Urbinas Graecus 82, which was unknown to Müller, which Fischer claimed to be the most reliable of all the MSS and which Rivet/Smith use whenever it suits their purpose) and in most MSS is Okitis. Rivet/Smith admit (1979, 129) that there is still no general agreement on the correct text, and to complicate matters they prefer (452) Scitis, a variant found in three Florentine MSS, (corruption from sigma C to O), presumably because it comes closer to its alleged derivative Skye. But there is an alternative possibility. A similar composite of X and U produces Oketis (restoration from sigma C to O), which is virtually a metathesis of Ouektis. Metathetical distortions are common in Ptolemy, especially where there is a 'k' or 'g' in the name concerned( Schütte, 1917, 22). So ket becomes ekt (Schütte, ibid., cites urg > rug, ukl > luk ). The result is that we have two 'Isles of Wight', a familiar one in the south and a duplicate in the north. The northern version would appear to be Pliny's Irish Sea Vectis ( the proper Latin name for Skye?). Ptolemy's northern Ouektis, whether or not it is Skye with a distorted location or some other island (Stroma?) off Caithness, was assumed by later historians to be one of the Orkneys because it was sandwiched by Ptolemy between Orcas and Orcades. Whatever variants we find in the medieval MSS of Ptolemy, I am inclined to believe that the original reading is Ouektis, that Ptolemy located it near Orcas promontory, and that this is what Eutropius' unnamed source assumed it to be part of the Orkneys. Duplications are common enough in Ptolemy (e.g. Camulodunum, Coria, Alauna, Derventio, Dumnonii, etc.). Agricola's marines could not have overrun Skye, since Skye is bigger than Thule or any of the Orkneys, and Tacitus would have mentioned any invasion of the Hebrides. Stroma was almost certainly overrun, but isn't mentioned by Tacitus because it was categorised among the Orkneys, as indeed Swona, likewise situated in the Pentland Firth, is today. Yet Stroma would be known to Agricola and may well be my suggested Vectis. In which case Dumna needs to be re-examined, especially since Pliny, as I maintain, puts it in a Scandinavian setting, where Stichtenoth (1959, 84) prefers to locate it.
291 Momigliano refers to senex as an indication by Silius that Vespasian's career builds up to his old age. But this career would appear to ignore the principate. I believe that the deified Vespasian and Titus are presented as they were at the time of their death and apotheosis, as principes senex et iuvenis, for enrolment among the members of the divine fraternity over which Jupiter presides. Why else would Silius describe Titus as iuvenis and then refer to his achievements of some decade earlier as primo in aevo (in his early years 606)?
292 Wistrand (1956,5) suggests that augebit is equivalent to addet and that the Flavian family will add its name to the pantheon which includes the Julio-Claudians. But nomen augere simply means to enhance the name, a common expression, cf. Cic.Dom. !9, auxit nomen populi Romani.
293 Agr. 17.1. I don't support the observation of McDermott/ Orentzel (1979, 51) that Flavian policy seems to have been against expansion and despite absence of comment it seems unlikely that Vespasian anticipated the campaigns in the north with which Agricola indulged himself. The literary and archaeological evidence suggests otherwise, and how can Agricola's campaigns be classified as self-indulgence? The authors' suggestions that Vespasian appointed Liberalis to keep a check on Agricola because the emperor had reservations about the governor need to be substantiated. The fact that Tacitus never mentions Liberalis is irrelevant. Tacitus' attitude towards Vespasian is distinctly favourable and reflects the debt which Agricola owed to the emperor.
294 Suet. VC. Vesp. 24.2 he was carrying out the functions of generalissimo, cf. Levick (1999, 258). It is quite likely that Vespasian relied heavily on Titus, his consular colleague, cf. Suet, VC. Titus 6.1. Jones (1984,150) noted that it would not be unreasonable to assume that in this, the last year of Vespasian's reign, Titus was even more closely involved with military planning and had fully approved of Agricola's plans. This would be enough reason for Titus to claim a fifteenth acclamation as imperator.
297 Titus assumed the title of Augustus within a week of his father's death, and the fifteenth imperatorial acclamation between July 1st AD79 and January 1st AD80 (Buttrey 1980, 25). Dio (RH. 66.20.3) does not link this acclamation to a victory, as Buttrey suggests. Tacitus makes no mention of any victory in AD79. But Dio's language (RH. 66.20.1), that Agricola in AD79 ravaged all the enemy's territory there corresponds with Tacitus' vastatis usque ad Taum nationibus (had ravaged nations as far as the Tay, Agr. 22.1) and repeated by percucurrerat (had overrun, Agr. 23.1) to describe incursions made beyond the isthmus. If Dio's statements are accurate, Agricola's advance to the Tay was in response to aggression by the tribes of southern Scotland, since it is likely that the war which had broken out again refers to a recrudescence of problems which Agricola's predecessors had failed to resolve. I believe that we have an 'ebb and flow' situation here with the both Romans and Caledonians seizing and losing the initiative. The Romans, however, must have controlled the sea for some years before Agricola took office, and consequently the Fife peninsula, the home of the supportive Venicones, might have taken on the strategic importance of a friendly right flank earlier than previously thought. If Vespasian's initial policy in AD69-70 was total conquest, any way of achieving that aim, whether by devastation or treaty, was acceptable. There were evidently two wars in the lowlands, no doubt punctuated by acts of aggression, spread out over a period of some ten years, the first one being the 'spill over' from the Brigantian war (AD69-71) which may well have been prolonged by Venutius (no reference to him being captured or killed) among the tribes of southern Scotland, in the same way as Caratacus had prolonged the resistance among the Welsh tribes. The second war, which Tacitus refers to as expeditiones, appears to have been initiated (if Dio is to be understood) by the natives, and was punished by devastation. Tacitus' attribution of Scottish initiatives to Agricola may well disguise the reality that Bolanus and Cerialis had led the first incursions into the area (cf. Breeze 1996, 34). In which case the blueprint for campaigning there was already in place and would facilitate and accelerate Agricola's campaigns The evidence for this will be examined later.
305 F, G, X (ed. Delz). These MSS, as may be seen from Delz's stemma, have as much, if not more, validity than the rest. Curiously this reading doesn't appear among the variants in the edition of Miniconi/ Duvallet.
306 The glorification of the warrior family is a poor imitation of Virgil's panegyric on the gens Iulia, cf. Aen.1.287 and 6.781. Silius proceeds from the rustic origins of the gens Flavia in Sabine territory to beyond the Ocean and into the furthest limits of the known world. This arrangement creates the illusion of how far and how quickly Flavian domination had extended. Salway (1993, 117) refers to the triumphal arch set up at Richborough to commemorate the completion of the conquest: Domitian would have had every incentive to mark as publicly as possible that it was under his auspices that the enormously prestigious expansion of Roman domination across the Ocean, so strongly associated with his own family, had been brought to a successful conclusion in his reign.
310 Ogilvie 1967, 33 especially with reference to Oceanus and Thetis. The theories of Fabre and Dion that Pytheas had received a commission from Alexander, whose apparent ambitions for conquest in the west were cut short by his untimely death, might gain some support if the Alexander- Pytheas-Agricola connexion is accepted.
311 Agr. 25.1 silvarum ac montium profunda (the depths of forests and ravines) , 26.2 nisi paludes ac silvae fugientes texissent (if the swamps and forests had not sheltered the fugitives), 33.5 evasisse silvas (to have emerged from forests), 34.2 silvas saltusque penetrantibus (penetrating the forests and ravines).
313 Hanson/Macinnes make valid points on the nature of the terrain facing Agricola, but are incorrect in stating that Silius Italicus' reference to Caledonian forests was written before Agricola's campaigns (1980, 102).
316 Wijsman 1998, 319, citing Sil. Pun. 17.417-7, caerulus haud aliter cum dimicat incola Thyles/agmina falcigero circumvenit arta covinno (just so the blue-painted native of Thule, when he fights, drives around the close-packed ranks in his scythe-bearing chariot). The arta agmina are Tacitus' densa agmina (Agr. 36.3). Caesar doesn't specifically mention that essedarii used scythed chariots. In fact the description he gives (BG. 4.33) would make any scythe attachment positively life-threatening to the drivers running along the pole. For the differences between essedarii and covinnarii, cf. Couissin 1932, 102.
317 Agr. 36.3 equitum turmae fugere (the squadrons of cavalry fled) is clear enough (cf. Burn 1953a, 154; Lund 1981, 84)). The British cavalry fled, and the covinnarii became involved in the infantry action. They were not driven off, as Dobson (1981, 10) suggests, and their alleged flight appears to have occurred before they were actually involved in the battle. It is hardly likely that Silius would use a simile suggestive of failure in order to describe a triumphant Masinissa.
319 Martial (Epig. 10.44.1) talks of his aged friend Quintus Ovidius leaving the comfort of his Sabine farm to visit Caledonios Britannos. There may be touch of irony here. But Caledonios was topical, and Martial's tenth book was published in the same year as the Agricola. The Flavian involvement in northern Britain was as much a topic for satirists as it was for the epic poets.
320 Stat. Silv. 5.2. 55-56. quantusque negantem/ fluctibus occiduis fesso usque Hyperione Thylen/ intrarit (with what impact he entered Thule which denies the western waves where the sun is ever weary). Just as Silius refers to the Caledonian as incola Thyles, Statius is merely elaborating on Bolanus' activities in Scotland. Claudian likewise (lll Cons. Hon.53-56), in describing thecampaigns of Count Theodosius in northern Britain, refers to ratibus impervia Thule (Thule inaccessible to ships, clearly ignoring Agricola's achievement), Pictos and the conquest of The Hyperborean waves. Wijsman's theory (1998, 323) that Thule in this instance is equated with Britain generally, instead of the far north of Britain, cannot be correct, and his quotation from Namatianus (De Red. 499-501) clearly distinguishes Thule from the ferox Britannus.
321 Stat. Silv. 5.2. 142. It is possible to regard this as somewhat exaggerated, but not necessarily inaccurate reporting, as Breeze (1996, 34) suggests. Despite the embroidered language there are elements of truth (cf. E.Birley 1953, 13; Dobson 1981, 4)).
322 Tacitus is vague about the activities of Vettius Bolanus. In Hist.3.45 we are told that the Brigantes received support from outside the tribe, probably from the Scottish lowlands (cf. Cunliffe 1991, 208). There were close ties between the Brigantes and the Selgovae and Novantae (cf. Frere 1987, 111) and it is possible that the Votadini at this time, as Frere suggested ( 2001, 291), may also have been part of the opposition. Such support was unlikely to go unpunished. The poet's advice to Crispinus, Bolanus' son, who is about to engage in a military career, is to learn from his father (disce patrem) what military glory is all about. Ogilvie's observations (1967, 156) on the extent of Bolanus' activities should be treated with caution; there may be some basis for Statius' embroidery. For details of Bolanus in Britain, cf. A.R. Birley (1981, 64-65; 1999, 68). For pre-Agricolan involvement in Scotland, cf. Woodhead 1947-8, 56; Hanson 1979, 16-17; 1991a, 55-56. For the most perceptive and prophetic observations, cf. E. Birley (1953, 13-14, 40-41).
323 Silv. 5.2.55-56. Statius has developed a Scottish mainland Thule from the sea-beaten island of Silv. 5.1.81. What does emerge among the Silver Age poets is that activities in Caledonia and the northern isles are seen as part of the same campaign. This is merely poetic licence for the more realistic account we find in Tacitus. There is no more likelihood of Bolanus campaigning in Shetland than there is of Shetlanders (incola Thyles) driving chariots at Mons Graupius.
324 Statius' statements must have some foundation in fact, however slender (Dobson 1981, 4). One wonders whether Agricola was overlooked by Bolanus during the Brigantian War. It was only when Cerialis became governor that Agricola's military career really blossomed : habuerunt virtutes spatium exemplorum (his qualities now had scope for display) Agr. 8.2.
325 Agr. 8.1; 16.5. The inertia erga hostes (inactivity towards the enemy) doesn't tie in with Statius' heroic depiction (cf. Wellesley 1972, 130). That Bolanus grabbed a breastplate from a British king (Stat. Silv. 5.2.149) may be less than grabbing the king himself, but suggests initiative rather than inaction. How can there be inertia when Bolanus is actually involved in a war? This may be the consequence of policy differences between Agricola and Bolanus or simply an attempt to boost the image of the magni duces whose pre-Flavian predecessors fare badly by comparison (cf. Agr. 20.1). Trebellius' policy, as Ogilvie (1967, 202) pointed out, was to curb the excesses committed by the army during the lull occasioned by the Civil War. But Tacitus doesn't tell us what the Brigantes or Trebellius had been doing between AD63-69 (cf. n. 273). He tars Trebellius with the same brush, segnior (lazier), as Bolanus (inertia) whose career suggests otherwise. The advent of the Civil War blunted the soldiers' fighting resolve. The army was adsuetus expeditionibus (accustomed to campaigning, Agr.16.3), and it might be asked where they had been campaigning for six years, even if Trebellius, like Didius Gallus before him, was acting per ministros (through subordinates). Tacitus does not link Bolanus specifically to any northern campaigns, whereas Statius does. Bolanus had possibly two campaigning seasons to deal with Venutius and the Brigantes, sufficient time for operations in southern Scotland, and may well deserve some of the credit which Tacitus gives to Cerialis, just as Cerialis deserves some of the credit given to Agricola. It would not suit Agricola's curriculum vitae to credit Bolanus with any initiatives. Statius refers to Bolanus as establishing watchtowers and forts (speculas castellaque, Silv. 5.146)) in Caledonia. The only watchtowers in Caledonia lie along the Gask Ridge. Are we to see here a Vitellian initiative at this time? In which case can we give any credence to Tacitus' account? Statius mentions Bolanus addressing his turmae (cavalry squadrons), while Tacitus,without mentioning Bolanus, refers to alae (cavalry wings). This makes no reference to legionaries (cf. E.Birley 1978, 244) and might preclude Agricola's involvement as commander of the twentieth legion, which could easily lead to frustration (cf. Agr. 8.1) and friction. The cohortes and alae which Bolanus had employed against the Brigantes (Tac. Hist. 3.45) were less than efficient and perhaps still undisciplined after their challenge to Trebellius (Tac. Hist. 1.60). But it is the legionaries, not the auxiliaries, whose indiscipline is stressed in the biography. Agricola's self-restraint under Bolanus (Agr. 8.1) is depicted by Tacitus in order to provide a contrast with Roscius Caelius' mutinous attitude under Trebellius. The charge of inertia cannot represent Bolanus' activities prior to Agricola's arrival (cf. A.R.Birley 1981, 64) and whether it is justified subsequently is debatable.
326 Hardie (1983, 69) refers to the absence of living viri militares in Statius' Silvae. He may well be correct in stating that the approval of the court may possibly be seen behind the rehabilitation of the pre-Agricolan governors like Bolanus. Since Statius published his tribute to Bolanus while Domitian was still alive, while Tacitus did not publish anything before the emperor's death, the latter's account may be seen as 'sour grapes' rather than setting the record right. There is always the risk that damnatio memoriae may create distortion rather than remove it.
328 It was not gout, but a slight fever (Dio RH. 66.17.1) which led to Vespasian's death. It is quite clear, however, from Suetonius (VC. Vesp.24) that from the time he was temptatus motiunculis levibus (attacked by a slight bowel disorder) to the time he died of more serious intestinal problems (dysentery) a number of weeks had elapsed during which he had been able to travel from Campania to Rome and from there to his Sabine homeland to carry on his duties as emperor. I suggest that it is during this period of his declining health that he received news of Agricola's advance to the Tay, and it is during this period that Valerius Flaccus composed his dedication.
329 Details of earlier contributions to the controversy on dating may be found in Scaffai 1994, 2368- 2373 (particularly Ussani's groundwork). See also Taylor 1994, 214 and Liberman 1997, xviii-xxiii.
333 Burmann 1724, 7. Flavian propaganda would emphasise Julius Caesar's failure in Britain (Agr. 13.1), and the wrecking of his fleet would be seen as the Ocean's revenge for Caesar's pilfering of its pearls (cf. Suet. VC. Jul.Caes., 47; Plin, HN. 9.116. Gaius' brief excursion into the Channel and back (Dio RH. 59.25.2), which Malloch (2001, 554) saw as symbolic of Gaius having extended his imperium over the sea would be interpreted otherwise by Flavian apologists and would be represented as mal de mer and a failure to conquer the Ocean, hardly compensated by a collection of sea shells (Suet. VC. Gaius 46; Dio RH. 59.25.2). Claudius' expedition, like Caesar's, found itself driven off course in the Channel (Dio RH. 60.19.4). None of the Julio-Claudians was enamoured of Oceanus, and vice versa. All this should be set against the triumph over the Caledonius Oceanus by Vespasian's navy under his legates. The conquest of the Ocean (victus Oceanus, Agr. 25.1) by Agricola's marines, who bragged about it in AD82, had been ongoing since AD70.
338 That Cerialis established a fort at Carlisle c.AD73 is now confirmed by dendrochronology (Frere 1990, 320) and supported by the numismatic evidence (Shotter 2001, 21-29). Naval operations through the Solway Firth towards Galloway and beyond would not be unreasonable.
339 Cf. note 70. Pliny in a well-known sentence (HN. 4.102, mistranslated by Rackham) writes, xxx prope iam annis notitiam eius (sc.Britanniae) Romanis armis non ultra vicinitatem silvae Calidoniae propagantibus (in almost thirty years now Roman arms have been spreading knowledge of Britain as far as the neighbourhood of the Calidonian forest. This can only refer to the period of expansion from the Claudian invasion to Pliny's own time. Dilke (1985, 69), following Rackham's impossible about thirty years ago (not, pace Hanson, possible linguistically) erroneously assigns Pliny's reference to the Claudian era, and then ironically accuses Pliny of lack of serious study. Bianchetti (1998, 143) is even more unfortunate in her observation that the reference seems to be to the expedition of Claudius in 43, when the "unexplored forest of Scotland" was not reached by the Roman army. The true and proper Caledonia was only reached by Agrippa in 83. Roseman (1994, 88-89) likewise mistranslates as it is almost thirty years now since acquaintance with (Britain) was extended by Roman arms and talks of exaggeration in relation to the Claudian conquest. I see no exaggeration on Pliny's part and nothing that would preclude the presence of Roman troops near the Caledonian Forest. Pliny's contentious statement is as loud as Tacitus' silence on the matter. The Historia Naturalis was dedicated to Titus in AD77. But this does not help us with date of the reference which could have been inserted either at the time when Book 4 was written, or added later, as the dedication was, in order to include the latest information. A date c. AD70-71, however, is in keeping with in almost thirty years, and would tie in with Statius' references to Bolanus' activities in the Caledonian plains. The location has always been open to question (cf. Rivet/ Smith 1979, 290; Hanson 1991a 56). If Bolanus was operating in the lowlands, then there is no good reason why Cerialis should not have capitalised on this by pushing up to the isthmus or even beyond. Agricola would not have made such rapid progress in his third campaign and been able to establish forts in the lowlands (Agr. 22.1), if his predecessors had not already prepared the ground, cf. Nesselhauf (1952, 222ff); Levick (1999, 159). Dobson's attribution (1981, 7) of Agricola's successful push to the lack of pugnacity of the Lowland tribes may well be correct if they had already experienced the might of the Roman army.
340 The political consequences of Roman control of the Ocean and the eastern seabord may be seen in the willingness of the Parisi, Votadini and Venicones to adopt a pro-Roman stance. Any arrangements made by the Romans to protect the land frontiers of these tribes required permanent garrisons. But control of the sea was always unassailable and reassuring . The effect on hostile tribes of the presence of the Roman fleet is well documented by Tacitus (Agr. 25.1-2, 29.2).
342 Köstlin 1889, 651. The link between Caledonius and Agricola's penetration of Scotland has lately been resurrected by Moreda (2000, 25). This must be correct, although I don't support his theory of a dedication made to a deceased emperor. Once again the failure to appreciate the revised dating of Agricola's governorship by putting the incursion into Caledonia in AD80, instead of AD79 has deprived Vespasian of the comfort of Agricola's success.
346 Syme 1929, 137. Wilkinson (1957, vi) made a valid point that it is hard to believe that if Book 1 was published under Domitian he would have been content to be ignominiously sandwiched between two appeals to his father.
360 Wijsman (1988, 319), following Ehler's chronology (AD71-79, cf. Liberman 1997, xxiii) for the composition of the Argonautica, says that time would not have allowed the author to have known the splendid results of Agricola. This claim is obviously based on the old dating (AD78-84) for Agricola's governorship, a dating which puts the death of Vespasian before Agricola's devastation of Caledonian territory.
363 Spaltenstein 1990, vol.1, 249-250 (following Miniconi/Devallet 1979, 158) his reign was marked by the submission of the Scots (?). He is incorrect in stating that Agricola's first contact with the Caledonians was in AD83, four years after the death of Vespasian. Agricola had penetrated Caledonia as far as the Tay in AD79. Hanson (1991a, 84-85) tentatively identifies the camps at Abernethy and Dunning with this campaign.
366 Mann/Breeze 1987, 87. Their claim that Agricola had not penetrated north of the Moray Firth would be invalidated by my interpretation of the MS. Henderson (1985, 328) drew attention to Ptolemy's list of putative forts in Moray, Tuesis and Pinnata Castra, noting that forts were invariably built in conquered territory in the year after the army had passed that way.
368 Agr. 25.1 civitates trans Bodotriam sitas (the states situated across the Forth), 25.3 Caledoniam incolentes populi ( the peoples inhabiting Caledonia), Hanson (1991a, 120) asks the question why Tacitus would refer only to the Caledonii in opposition to Rome and explains this as imprecision in matters geographical. On this basis the Boresti are anomalous. If Tacitus was imprecise, then Agricola would simply have taken his army down to the coast. Burn (1969, 38) says that Tacitus dexterously and deliberately avoids loading his narrative with unfamiliar names. But he doesn't explain why Tacitus mentions the Boresti, and his suggestion (1969, 58) that they were probably a component of the far-flung Caledonians is merely stating the obvious.
377 Agr. 10.1 tum primum (Britannia) perdomita est (then for the first time the conquest of Britain was completed). Keppie (1989, 61), commenting on the victory at Mons Graupius, aptly described the moment as the best opportunity of carrying through the conquest of the island to its logical end (cf. Robertson 1975, 5).
378 Agr. 27.1 the boundary of Britain should be found. It is somewhat ironic that Henderson (1985, 330) claims that the Boresti merited special attention because they inhabited the terminus Britanniae or finis terrarum. There is no doubt that tribes inhabited Caithness, but they were not fictional. Arguments for creating mythical locations, as Rivet/Smith did in the case of Trucculensis portus, because they are located in the furthest limits of the island, are insupportable.
382 Ogilvie 1967, 32. cf. Hanson 1991a, 22. It is precisely for this reason that the emendations of Richmond Anavam primum and of Oniga Novium primum (Agr. 25.1) for prima nave are really non-starters, although Richmond's suggestion still has its supporters, cf. Murgia 1978, 161-162. Likewise Maxwell's suggestion of Tamium (1984, 221-2) based upon E's unfavoured Tanaum (Agr. 22.1) cannot be supported, especially as Maxwell himself points out that the character of Agricola does not lead us to believe that he was concerned with such topographical precision. Henderson (1985, 319) was correct in his assumption that geographical details would bore or distract Tacitus' readers. In which case why mention the Boresti?
383 Gorrichon (1974, 196) drew attention to the emphasis on Agricola rather than on the island when she described the geographical imprecisions in Tacitus' account; somewhat ironic in view of the fact that she resurrects the long discarded Mont-Grampian and assumes that Tanaum (certainly inferior to Taum) refers to the Tyne, which de Saint-Denis (following Furneaux) believed to be the geographical limit of Agricola's third campaign. Ogilvie (1967, 31) was correct in his view that geographical information in the 'Agricola' is largely incidental. I am surprised that Rutledge (2000, 92 n.48) believes that such information is fundamental to the Agricola's ultimate purpose.
386 Amm. RG. 23.6.74. It is quite possible that Ammianus borrowed this adjective from Tacitus. Fletcher (1937, 390-392) lists thirty two similarities of vocabulary and phraseology between Ammianus and Tacitus, with five drawn from the Agricola. Blockley (1973, 67) attributes Ammianus' verbal borrowings from Tacitus to narrow, stylistic purposes. This need not be the case with individual words such as boreum/boreos, rare in prose writers. For a more recent study of the evidence for Ammianus' reading of Tacitus, cf. Barnes 1998, 194-195.
388 Ptol. Geog. 2.2. A link between Boresti and the Hyperboreans was suggested by Holder (1896, I.490) and supported by Henderson (1985, 330) with reference to Boreas and the adjectival boreus. Ogilvie (1967, 282) and Burn (1969, 58) discounted any connexion with modern Forres, an option which Frere (1987, 97) and Robertson (1976, 7) leave open. Other suggestions, which reflect, as Maxwell (1989, 70) aptly put it, the ingenuity of generations of scholars and amateur philologists, include Rhys (1904, 283-284) who proposed a connexion with the Late Latin floresta, Whatmore (1913, 215) who conjured up a link with Loch Ericht, Pitblado (1936, 108) who thought there was a connexion through 'phonetic change' with Brossi, equivalent to Ross or Rossia, the ancient name for Fifeshire, O'Rahilly (1946, 529) who imagined a corruption of Voretii (dwellers by the Forth?) and Walters (1899, 109) whose fantasies produced a connexion with the Borussi from Prussia. But all these suggestions are speculative, and lead away from the solution rather than towards it Hübner (1897, 731) at least had the sense to steer away from any identification.
398 Demetrius' imperial expedition (cf. Burn 1953a, 114-115; 1969, 53-54; Ogilvie 1967, 32-33) of inquiry and survey, which may also have been involved, as Wainwright (1962, 66) suggested, in gathering military intelligence (cf. note 115), would have produced a report written in Greek. It is further worth noting that the dedication, set up at York (RIB 662, 663), was written in Greek by a Scribonius Demetrius, surely to be identified with Plutarch's Demetrius (De Def. Orac. 410A, 434C) whose visit to York coincided with that of Agricola (cf. RIB. Addenda and Corrigenda 771).
403 Crawford 1949, 130-133; Burn 1953b, 127-133; 1968, 316; 1969, 56; Richmond 1958, 51-52; Ogilvie 1967, 65, 182; St. Joseph 1978, 279-285; Henderson 1984, 28; Keppie 1980, 79-88; Maxwell 1989, 58-59; 1990, 72-123; Jones/Mattingly 1990, 76-77; Hanson 1991a, 129-137; Jones 1993, 225 n.44; Fletcher 1999, 4. I had originally been seduced by St. Joseph's location of Mt Bennachie. But the camp at Durno might well be Severan. The question of a location is an irrelevance to Henig (2002, 67) whose suggestion that a pitched battle never occurred at all cannot be taken seriously.
404 A.R.Birley 1999, 85. His reference (88) to the possibility that the aestuaria (Agr. 33.5) could comprise the Moray and Dornoch Firths has much to commend it. The density of brochs north of the Dornoch Firth (cf. Jones/Mattingly 1990, 62) would suggest a considerable source of manpower for resistance against Agricola and a good enough reason why a final confrontation to defend the last remaining settlements should have occurred somewhere nearby. Frere's suggestion (1987, 94) that the broch-builders may have been friendly to Rome would be invalidated by the taking of hostages (Agr. 38.2) in the northern extremities. Richmond (1958, 51) noted that the broch area must have supplied a substantial part of the native force. Nearly a hundred brochs, so far discovered in the Shetlands imply close cultural and, perhaps, strategic links with the Orkneys and the Scottish mainland. Ogilvie (1967, 33) was misguided in relying on Solinus as proof that the Orkneys were deserted. The evidence from Clickhimin in the Shetlands suggests a tradition of maritime activity (Hamilton 1968, 78-79) which may have extended to the Scottish mainland. One can understand Agricola's need, despite Tacitus' silence, to sever the links and stop the Orkneys and Shetlands becoming, as Anglesey once did, a sanctuary for deserters (Ann. 14.29.3 receptaculum perfugarum). I see nothing to support Wainwright's theory (1962, 65) of hostile relations between the seafaring people and the mainland fort-dwellers which he has conjured up out of the alleged Claudian clientela. A.R.Birley's suggestion (1999, 90) that the Boresti should be located in Caithness supports my argument for the extreme north (fines borei). Burn (1953a, 156) located the coast of the 'Boresti' as well to the north without committing himself to a positive location beyond Moray which the terminus Britanniae must surely imply. Richmond's comment (1958, 52) that the Latin account indicates that after the battle Agricola did not advance farther can only be supported if he follows E2m,'s reducit, which all editors reject. Breeze (190, 55) says that the lack of place-names in Ptolemy's Geography north and west of the Great Glen suggests that Agricola did not penetrate into these areas and that therefore the battle of Mons Graupius was not fought beyond Inverness. No place-names means no army presence. This need not be so, since the lack of place-names need not preclude the existence of marching camps, indicative of penetration beyond Moray. Has anyone really looked for them?
406 Agr. 30.3 no people beyond us, nothing but tides and rocks. Burn (1953a, 150) says, fairly near the sea. The tendency of commentators to infer rhetorical exaggeration ignores the fact that the reality doesn't require distortion. Gsell's assertion (1894, 170 n.5) that it is impossible to draw any geographical indications from Tacitus' speeches fails to grasp the personal experi ences of Agricola in Britain, as relayed from father-in-law to son-in-law.
408 Grimal (1990, 68-71), on the basis of Tacitus' vocabulary, assumes that Tacitus was serving in Agricola's army during the Mons Graupius campaign and was an eye-witness to all that happened. His suggestion that Tacitus served under Agricola as proquaestor (?) is a figment of the imagination. There is no evidence that Tacitus was in Britain during the principate of Domitian. Despite the claims of Mann/Breeze (1987, 85) that Tacitus did not visit Britain a good case for a military tribunate under Vespasian and Titus has been made by A.R. Birley (1999, ix and 2000, 235).
409 Agr. 33.3 we hold with arms and encampments (so Rutledge 2000, 78). Hanson's translation (1991a, 129) we hold with arms and forts does not meet the definition of castra as laid down by Ogilvie (1967, 198) who assigns castra to the legionary fortresses, and castella to smaller units. I do not subscribe to Ogilvie's insistence that castra in English should be rendered by fortress (not camp). Which Latin word do we use for camp? Which legionary fortress ever graced finem Britanniae? Which governor would station his finest troops in Caithness? I believe that in this example castra means marching camps (so Robertson 1975, 7). This is implied by the use of in agmine in the next sentence. When the Romans set up camp in hostile territory it is an indication, however temporary, of conquest. It was the marching camp (castra)of the ninth legion which was attacked by the Caledonians in the sixth campaigning season (Agr. 26.1).
412 Curt. HA. 9.9.1 a desire to visit the Ocean and approach the ends of the earth. Fabre (1975, 44) noted a theory that Alexander was aware of the explorations of Pytheas, and that after the exploits of Nearchus he dreamed of conquering the British Isles and even Thule. But Fabre regarded this hypothesis as extremely risky. Dion (1977, 209) cites Seneca (Ben. 7.2.5) for the activities of Onesicritus (praemissus explorator), Alexander's pilot, in the western ocean.
Send email to the author: firstname.lastname@example.org
INTRODUCTION | PART ONE: Shetland: The Classical Geographical Context
Copyright © 2002 by Stan Wolfson. All rights reserved.