Short URL for this page:

[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

This webpage reproduces an article in the
English Historical Review
Vol. 2 (1887), pp100‑103

The text is in the public domain:
Thomas Hodgkin died in 1913.

This page has been carefully proofread
and I believe it to be free of errors.
If you find a mistake though,
please let me know!

 p100  The Roman Province of Dacia

A question of historical geography which, as it seems to me, deserves more attention than it has yet received, is this: What were the limits of the Roman province of Dacia added by Trajan to the empire? I propose here to recapitulate some of the arguments on this subject adduced by M. de la Berge ('Essai sur le Règne de Trajan,' 55‑62), adding a few of my own. Most geographers have considered themselves bound by the authority of Ptolemy (III.8.4) to accept as the boundaries of Trajan's province the Tibiscus (Theiss?) on the west, the Carpathian mountains on the north, the Tyras or Dniester on the east, and the Danube on the south.​1 This demarcation gives to the province of Dacia the eastern half of Hungary, the Banat, Transylvania, Wallachia, Moldavia, and Bessarabia, forming an aggregate of at least 70,000 square miles.

Even on the face of the ordinary classical atlas there are some objections to such a demarcation as this. The interval between the Danube (when it is flowing from north to south) and the Theiss is so long and narrow that it is difficult to suppose that a strategist like Trajan would leave such a wedge between Pannonia and Dacia to be occupied by the Jazyges Metanastae, to whom, on the authority of Ptolemy, it is assigned. Again, on the north-eastern frontier of the province it is almost inconceivable that the Romans would abandon the splendid natural defence afforded by the Carpathians, and choose such a comparatively feeble defence against the wandering hordes of Scythia as might be afforded by the river Dniester. The chief argument, however, brought forward by M. de la Berge is derived from Eutropius,​a who estimates the whole circumference of the province out of Dacia at 1,000 Roman miles; ea provincia decies centena millia passuum in circuitu tenuit. For the Dacia of the maps this figure is decidedly insufficient.​2 And though Eutropius is certainly  p101 not a first-rate authority, it is to be observed that he had no reason for minimising, but rather for magnifying, the extent of Trajan's conquests. As M. de la Berge remarks, this number is found in all the MSS. of Eutropius, is confirmed by his brother abbreviator Sextus Rufus,​3 and may by probably have been borrowed from some official record to which Eutropius had access.

Let us then for a moment, relying on this passage of Eutropius, admit the possibility that Ptolemy was speaking, not of the Roman province of Dacia, but of a very different matter, the geographical extension of the Dacian people; and then let us consider what size we should be disposed to attribute to the Dacian province, judging from the best of all evidence, the undoubted traces of Roman occupation. Thus considering the question, we shall, it is submitted, be almost compelled to reduce the area of Dacia to that of Transylvania and Little Wallachia (or Wallachia west of the river Aluta) with the eastern half of the Banat.

Take the Roman roads as given in the 'Tabula Peutingeriana,' and as explained, for instance, in the preface to Smith's 'Atlas of Ancient Geography.' There is a little difficulty about the identification of a few of the sites, but there is no doubt that they were all in Transylvania, Eastern Banat, and Western Wallachia. The Peutinger table itself shows the roads running up into the roots of the mountains (Alpes Bastarnicae apparently being the Carpathian mountains), but never crossing them.

Still more striking is the argument which we may derive from a study of the inscriptions in vol. IV of the 'Corpus Inscriptionum Latinarum' (edited by Mommsen). We there find that the Latin inscriptions for the province of Dacia exist in overwhelming preponderance in Transylvania, chiefly at Apulum (Karlsburg), Napoca (Klausenburg), Polaissa (Torda?), and Sarmisegetusa (near Varhely). A few are found in Eastern Banat, and one or two, far fewer than might have been expected, in Little Wallachia, but none at all — as far as the 'Corpus' bears testimony — in Moldavia or Wallachia east of the Aluta. It is true that German settlers in Siebenburgen (Transylvania) are probably better finders and reporters of Latin inscriptions than their Roman and Slavonic neighbours; still that fact alone will hardly account for so enormous a difference.

Another weighty argument may be derived from the comparative smallness of the Roman army of occupation in Dacia. According to Mommsen ('Corpus,' IV.160) this consisted only of the thirteenth legion (Gemina) possibly increased under Septimius Severus by the fifth (Macedonica). When we remember that three legions were the minimum of the army of occupation for Britain, can we suppose that only two would have been entrusted with the defence of the immense tract of territory between the Theiss and the Dniester, intersected by the great Carpathian chain, which if not used as a bulwark would immensely increase the difficulty of holding it?

 p102  Another argument, to which, have, I do not attach so much importance, is that when the true Dacia, north of the Danube, was abandoned, and when Aurelian formed the new province of Dacia out of Western Moesia, its northern frontier was formed by the Danube between Singidunum and a point a little below Ratiaria. It thus stood nearly fronting what I believe to have been the old province of Dacia, and was not far from its equivalent in size. There is no such correspondence at all between the Dacia of the maps and the new province of Aurelian.

With reference to the western frontier of the province, it seems to be admitted by the general (but not unanimous) consent of map‑makers that this was not the river Theiss, but the Vallum (of which there appear still to be traces), which runs from a point north of Temesvar southwards to the Danube, which it touches at Viminacium. This certainly makes the narrow slip of territory left to the Jazyges Metanastae look somewhat less absurd. We must suppose that the desire not to occupy too large an extent of territory prevented the emperor from pushing his frontier, as we might naturally have expected him to do, up to the eastern border of Pannonia. But is it conceivable that while thus cautious on the western side he would have pushed his eastern frontier over the Carpathians into the limitless Scythian wilderness?

As to the geographical extent of the lesser Dacia for which I am contending, its perimeter is thus calculated by M. de la Berge:

Roman miles
From Viminacium to the mouth of the Aluta 243
Length of the Aluta 190
From the source of the Aluta to Porolissum (Dees?) 120
Porolissum to Viminacium 285

This result, as some of the distances have been taken as the crow flies, corresponds nearly enough with the 1,000 Roman miles of Eutropius.

It is clear from the language of D'Anville (I.262, Eng. transl. 1810) that Transylvania was in his time considered to be pretty nearly conterminous with Dacia, and I suspect that it is chiefly on his authority that the latter name has been extended to include also Wallachia and Moldavia.

In recent times philologers finding the Roumanian language spoken on both sides of the Carpathians, and believing that this was a legacy from the Roman occupation of Dacia, have fallen easily into the same view. But this argument from language proves far too much, since Roumanian is spoken in Thrace, in Macedonia, and even in Thessaly, and I suppose it will now be generally admitted that it is not safe to found upon the limits of the diffusion of Roumanian speech any argument as to the official boundaries of Trajan's province of Dacia.

Possibly I may be arguing for a proposition which scholars have already silently accepted; but if so, our school and college maps certainly require reconstruction. Inscriptions found in large numbers and south of the Carpathians might easily upset all that has been here advanced. My chief interest in the subject — on account of which I should be grateful even to a hostile critic who would give me some nearer approach to certainty on the point — is that this romanised Dacia, whatever  p103 were its limits, seems to have been the chief dwelling-place of Goths (rather, however, of the Visigoths than the Ostrogoths) during the hundred years which elapsed between Aurelian and Valens.

Thos. Hodgkin

The Author's Notes:

1 Ptol. 3.8.1 (ed. Müller). Ἡ Δακία περιορίζεται ἀπὸ μὲν ἄρκτων μέρει τῆς Σαρματίας τῆς ἐν Εὐρώπῃ τῷ ἀπὸ τοῦ Καρπάτου ὄρους μέχρι πέρατος τῆς εἰρημένης ἐπιστροφῆς τοῦ Τύρα ποταμοῦ . . . ἀπὸ δὲ δύσεως τοῖς Ἰάζυξι τοῖς Μετανάσταις κατὰ τὸν Τίβισκον ποταμόν. ἀπὸ δὲ μεσημβρίας μέρει τοῦ Δανουβίου ποταμοῦ τῷ ἀπὸ τῆς ἐκτροπῆς τοῦ Τιβίσκου ποταμοῦ μέχρις Ἀξιουπόλεως ἀφ’ ἧς ἤδη καλεῖται ὁ μέχρι τοῦ Πόντου καὶ τῶν ἐκβολῶν Δανούβιος Ἴστρος. There is some doubt whether the Tibiscus is meant for the Theiss or the Temes. Axiopolis is generally identified with Rassova.

2 Though I do not think M. de la Berge can be right in saying that the Theiss alone is 1,400 kilometers (875 miles) in length. From the map 500 kilometers looks more like the distance.

3 De Victoriis, cap. 7.

Thayer's Note:

a Breviarium, VIII.2.

[image ALT: Valid HTML 4.01.]

Page updated: 3 Dec 19