¹ Tac. Agr. 29.2: ad montem Graupium pervenit, quem iam hostis insederat. The battle is described in chapters 29–38.

² Fraser, J.E., The Roman Conquest of Scotland; The Battle of Mons Graupius a.d. 84 (Stroud, 2005)Google Scholar, hereafter cited as ‘Fraser’. Though broadly favourable reviews seemed to endorse the author's thesis (Keppie, L., War in History 14 [2007], 113–14CrossRefGoogle Scholar; Jones, R., Britannia 38 [2007], 392–3)CrossRefGoogle Scholar, the book has not gone uncriticized (Woolliscroft, D.J. and Hoffmann, B., Rome's First Frontier [Stroud, 2006]Google Scholar, 223: ‘the Fraser hypothesis is anyway difficult to take seriously, on purely tactical grounds’), although the publisher's hyperbole (‘James Fraser is the first historian to identify the most likely site of this legendary battle’) has ensured wide take-up by internet sources.

³ Fraser, 69. The ‘several antiquaries’ are not identified, but Alexander Gordon (c. 1692–1754) may be taken as typical of his age. He laboured under the mistaken belief that, ‘after this Battle was fought ... Agricola led back his Army into the Country of the Horestii, or Angus, for as it is certain ... that Agricola ... had been there before, so it is natural to think, he led his Army to the Place where his Fleet was, which most probably was in the River Tay’ (Itinerarium Septentrionale [London, 1726], 38)Google Scholar.

⁴ Fraser, 71.

⁵ Fraser, 70.

⁶ Tac. Agr. 38.2: exacta iam aestate spargi bellum nequibat.

⁷ Fraser, 73. Precisely the same objection could be levelled at his own favoured location in Perthshire.

⁸ St Joseph, J.K., ‘The camp at Durno, Aberdeenshire, and the site of Mons Graupius’, Britannia 9 (1978), 271–87.CrossRefGoogle Scholar

⁹ Fraser, 72. However, regarding St Joseph's ‘110-acre series’ of temporary camps, he is mistaken in his belief that ‘investigations ... at the example from this series of campsites at Ythan Wells, underlain by a clavicular camp, revealed stratigraphy that supports the conclusion that the series were left behind by the Severan army in the third century’. On the contrary, all that the stratigraphy proves is that the smaller (‘30 acre’) camp predated the larger (‘110 acre’) camp, but not by how much: see Joseph, J.K. St, ‘Air reconnaissance in Britain, 1965–68’, JRS 59 (1969), 104–28Google Scholar, at 112–13. Both camps at Ythan Wells may (or may not) be Flavian, but evidence of Flavian occupation at Kintore (another ‘110 acre’ camp) would tend to carry both camps at Ythan Wells with it. Of course, Flavian need not mean ‘Agricolan’. For the dating of Kintore, see Jones, R.H., ‘Chasing the army: the problems of dating camps’, in Breeze, D.J., Thoms, L.M. and Hall, D.W. (edd.), First Contact: Rome and Northern Britain (Perth, 2009), 20–7Google Scholar, at 22.

¹⁰ Fraser, 72. The most important words here are ‘as outlined above’, as distance only becomes an issue if Tacitus' testimony of a battle in late summer is discounted in favour of an earlier date.

¹¹ Fraser, 72.

¹² Fraser, 73, developing a point originally made by Maxwell, G., A Battle Lost: Romans and Caledonians at Mons Graupius (Edinburgh, 1990), 116–18Google Scholar. Instead of crediting Maxwell, Fraser cites his source as ‘the Flavian survey’, which is his preferred method of citing the second-century Geography of Claudius Ptolemaeus (‘Ptolemy’); cf. Fraser, 33–7. It is generally assumed, from the level of detail in Ptolemy's description of northern Britain, that his source ‘was not earlier than the campaigns of Agricola in a.d. 80–84’ (Rivet, A.L.F. and Smith, C., The Place-Names of Roman Britain [London, 1979], 114)Google Scholar, while the omission of Hadrian's Wall seems to confirm a first-century date.

¹³ Fraser, 74.

¹⁴ For the problems of plotting Ptolemaic data, see (for Greece) Livieratos, E., Tsorlini, A., Boutoura, C. and Manoledakis, M., ‘Ptolemy's Geographia in digits’, e-Perimetron 3.1 (2008), 22–39Google Scholar; (for Spain) Tsorlini, A., ‘Higher order systematic effect in Ptolemy's Geographia coordinate description of Iberia’, e-Perimetron 4.2 (2009), 117–30Google Scholar; (for Italy) Tsorlini, A., ‘Spatial distribution of Ptolemy's Geographia coordinate differences in North Mediterranean eliminating systematic effects’, e-Perimetron 4.4 (2009), 247–66.Google Scholar

¹⁵ For the problems posed by the so-called ‘turning’ of Scotland, see Jones, B. and Keillar, I., ‘Marinus, Ptolemy, and the Turning of Scotland’, Britannia 27 (1996), 43–9CrossRefGoogle Scholar, citing earlier literature.

¹⁶ Previous suggestions include Strageath (Richmond, I.A., ‘Ancient geographical sources for Britain north of Cheviot’, in id. [ed.], Roman and Native in North Britain [Edinburgh, 1958], 131–55Google Scholar, at 140), Inchtuthil (Rivet and Smith [n. 12], 499), Dalginross (Frere, S.S., ‘Naming Roman Britain’, Britannia 11 [1980], 419–23, at 421)CrossRefGoogle Scholar, Fendoch (Strang, A., ‘Explaining Ptolemy's Roman Britain’, Britannia 28 [1997], 1–30CrossRefGoogle Scholar, at 22), and even Stracathro (Maxwell [n. 12], 122).

¹⁷ Fraser, 72.

¹⁸ Watson, W.J., The History of the Place-Names of Scotland (Edinburgh, 1926), 55–6Google Scholar. He does not indicate which text of the Agricola he consulted.

¹⁹ Savile, H., The Ende of Nero and Beginning of Galba. Fower Bookes of the Histories of Cornelius Tacitus: The Life of Agricola (London, 1598)Google Scholar, 255. Curiously, the erroneous spelling appears in the translation of Church, A.J. and Brodribb, W.J., The Agricola and Germany of Tacitus (London, 1877)Google Scholar, 28, although the same authors' Latin text has the correct spelling (The Agricola of Tacitus [London 1869], 20)Google Scholar.

²⁰ Merivale, C., A History of the Romans under the Empire, vol. 7 (London, 1862)Google Scholar, 86 (repeated in subsequent editions).

²¹ The derivation has now been disproved by Breeze, A., ‘Philology on Tacitus' Graupian Hill and Trucculan Harbour’, Proc. Soc. Antiq. Scot. 132 (2002), 305–11Google Scholar, at 305–6, pointing out that crwb is not an original Celtic word, but a loan word from Old French.

²² Watson (n. 18), 56, evidently referring to Skene, W.F. (ed.), Chronicles of the Picts, Chronicles of the Scots, and Other Early Memorials of Scottish History (Edinburgh, 1867)Google Scholar, cxliii (‘a battle ... at Duncrub’), with 10 (‘[bellum] ... super Dorsum Crup’). He is misrepresented by Hind, J.G.F., ‘Summers and winters in Tacitus' account of Agricola's campaigns in Britain’, Northern History 21 [1985], 1–18CrossRefGoogle Scholar, at 13, who states that ‘an attractive candidate for Mons Craupius was found by W. Watson’, for he had made it perfectly clear, firstly, that the identification with Duncrub rested on Skene's authority alone, and secondly, that there could well be more than one ‘hill of the hump’ in Scotland.

²³ Feachem, R., ‘Mons Craupius = Duncrub?’, Antiquity 44 (1970), 120–4CrossRefGoogle Scholar. It is irrelevant to our purpose that Feachem misinterpreted the camp at Dunning as encompassing 16 hectares (40 acres) behind ‘Stracathro-type’ gate defences, while it was shortly afterwards revealed to encompass 46 hectares (115 acres) behind titulus-guarded gateways: see Joseph, J.K. St, ‘Air reconnaissance in Britain, 1969–72’, JRS 63 (1973), 214–46Google Scholar, at 218–19. It is similarly irrelevant that Fraser erroneously calls it a ‘clavicular camp’ (73, 76).

²⁴ Quoted in Feachem (n. 23), 120. It is worth noting that Jackson restricted his comments to ‘Craupius’, and earlier stressed that ‘Graupius’ may not be a Celtic word at all (‘The Pictish language’, in Wainwright, F.T. [ed.], The Problem of the Picts [Edinburgh, 1955], 129–66Google Scholar, at 135).

²⁵ Feachem (n. 23), 120. See also Jackson, K., ‘On some Romano-British place-names’, JRS 38 (1948), 54–8Google Scholar, at 55: ‘confusion of c- and g- is of course common in Roman spellings of Celtic names’ (citing Clevo for Glevo and Vindogladia for Vindocladia); ‘Romano-British names in the Antonine Itinerary’, Britannia 1 (1970), 68–82Google Scholar, at 70: ‘Galleva shows the not uncommon scribal confusion of c- and g-’; ‘the frequent confusion in late Latin manuscripts between c- and g- (see Calleva and Clevo)’. It is striking that his three examples relate to the Antonine Itinerary, and we may suggest that the reason (which Jackson thought ‘rather unclear’) is simply the scribal misreading of capital C for G (and vice versa).

²⁶ Fraser, 74.

²⁷ Tac. Agr. 33.6: nec inglorium fuerit in ipso terrarum ac naturae fine cecidisse. I have quoted from the translation of Birley, A.R., Tacitus: Agricola and Germany (Oxford, 1999)Google Scholar, 24.