H. A. Manandian The Trade and Cities of Armenia in Relation to Ancient World Trade
Chapter 4.Trade and Trade Centres in Armenia
in the Roman-Parthian and the Byzantine-Sasanid Periods
Continued from Previous Page [90]
24. The Main Trade Routes of Armenia according to the Tabula Peutingeriana :
The main roads of Armenia in the Roman period are indicated on the geographical map known as the Tabula Peutingeriana which was made, in the opinion of the German scholar Konrad Miller, in the second half of [91] the IVth century A. D. by the Roman geographer Castorius. The map has been published with a detailed commentary by Miller in his extensive work, Itineraria Romana, Romische Reiswage an der Hand der Tabula Peutingeriana (60A). We shall see, however, that the roads passing through Armenia were most incorrectly discussed by him.The precise determination of the junction points and stations on the map under consideration is of importance first of all for the determination of the direction of international transit routes from China and Central Asia through Ecbatana and Artaxata to the southern, eastern, and northern ports of the Black Sea and also for the establishment of the line of connexions running from Artaxata to Tigranocerta and Harmozica. In addition the clarification of disputed points concerning these extremely important places in ancient Armenia is also important for the illumination of a whole series of other problems connected with the history of the wars which took place on Armenian territory.We must note that the Tabula Peutingeriana , in the part concerning Armenia, has not been sufficiently studied by Armenists. This is to be explained by the fact that the earlier, commented editions of the Tabula Peutingeriana were not available to the wider circle of local Armenian scholars. At present, however, sections X4-XII2 of this irreplaceable source have been almost completely deciphered in the historico-geographical wrorks published by me during the past ten years (70).
In order to simplify further study of the interesting information found in the Roman map of Castorius, we give below the routes of the roads crossing Armenia, all their stations, and the distances between them.
1. Artaxata-Satala
The highway from Artaxata to Satala is given in the following manner in the Tabula Peutingeriana:Artaxata 23 Paracata 33 Coloceia (Zotozeta, Ra) 24 Hariza 24 Raugonia (Ragauna, Ra) 24 Colchion 24 Chadas 17 Armanas 12 Andaga (Andacas, Ra) 26 Barantea 30 Adconfluentes 10 Datamisa 20 Tharsidarate 15 Autisparate 12 Calcidava (Chalchidara, Ra) 15 Sinara 22 Lucus Basaro 15 Aegea (Egea, Ra) 20 Darucinte 20 Salmalasso 20 Satala (71).In our text the station points printed in capitals are the ones near which two towers are to be found on the Tabula Peutingeriana (71A). These towers, in the opinion of Miller and Markwart, indicate that the points so designated were important cities or military posts (72). Ra indicates the Ravennatis Anonymi Cosmographia (72A) in which important variants in station [92] readings are to be found. The figures between the names of the stations indicate Roman miles (milia passus =1.4815 kilometers).We see that the total distance between Artaxata and Satala consists of 406 Roman miles or approximately 600 kilometers. This distance is perfectly realistic and does in fact correspond to the distance between these two points.
The ancient city of Satala, the present Satal, exists at the present time and is found near the Kelkit' river. As we have already shown, the location of Artaxata, the ancient capital of Armenia, is also known. On the basis of quite definite indications of Strabo and of the Armenian historian Movses Xorenac'i (72B), we have seen that this famous city was located at the confluence of the Mecamor and the Araxes not far from the modern Xor-Virap.
In addition to these two junction points the names and locations of nine other stations are given below by me: two (Coloceia, Hariza)--on the line Artaxata-Bagawan, and seven (Chadas, Armanas, Andaga, Barantea, Calcidava, Lucus Basaro and Darucinte)--on the line Bagawan-Satala. Thanks to the new evidence and considerations brought forth by me, the problem of the direction of the line under consideration receives, as we shall see, an altogether different elucidation, and the disputed routes can be established with more accuracy and foundation than has been done in earlier works.
The roads Satala-Artaxata and Tigranocerta-Artaxata joined together at the junction point Raugonia or Ragauna, as my be seen from the evidence of the Tabula Peutingeriana, and from there ran to Artaxata through the same stations--Hariza, Coloceia, Paracata. Consequently the location of Raugonia must first be clarified in order to determine the direction taken by these roads.
Konrad Miller, in the commentaries and explanatory map found in his Itineraria, supposed that the location of Raugonia-Ragauna was on the banks of the modern Suuk-su, north-east of Lake Van. As for the direction of the route from Satala to Artaxata, he determined it as follows: through Ashkale, Sankarich, Aghveran, Ahlat, Adeljivaz, that is to say, along the north shore of Lake Van (73).
Richard Kiepert, in his detailed map of Asia Minor, placed Raugonia in present-day Parnaut, which lies south of the Araxes between Koghb [Kulp] and Kalzwan [Kagizman]. For this reason the road from Satala to Artaxata is assumed by him to run through Bagarichh [Pekerij]--Ilija--K'eop'rik'eoy [Koprukoy]--Sarighamish--Kaghzwan [Kagisman]--Parnaut ---Koghb [Kulp] (71).
The attempts of both Kiepert and Miller to locate Raugonia cannot be accepted as successful. Markwart very rightly found that Raugonia or Ragauna [as well as Sagauana in Ptolemy] (74A) correspond to the ancient Armenian Bagawan (= Bagauna) (75). The reading Bagauna instead of Ragauna is very easily explainable on graphic grounds and the identification of Bagauna with the ancient Armenian Bagawan is entirely clear and evident. Therefore, the problem as to the location of this important junction may be considered as definitely solved. Bagawan, as we know, was not located on the bank of the Suuk-su not in modern Parnaut, but was an important center in the ancient Armenian province of Bagrewand and stood north-west of modern Diadin on the site of the present monastery of Surb-Karapet.
The identification of Bagauna with Bagawan gives us the possibility of determining the direction of the roads from Bagauna to Artaxata as well as from Bagauna to Satala and Bagauna to Tigranocerta. In the Tabula Peutingeriana the following stations are listed on the first road: Bagauna 24 Hariza 24 Coloceia (Ra Zotozeta) 33 Paracata 28 Artaxata (in all 104 Roman miles or 154 kilometers).
Of these stations, Paracata is to be identified, as was correctly noted by W. Belck (76), with the ancient Armenian P'arhaxot which is mentioned [94] in Movses Xorenac'i (77) and the Anonymous Source (78). After having carried out excavations on the spot, W. Belck succeeded in establishing that ancient P'arhaxot was located on the site of the ruins known under the name of Bulagh-Bash five kilometers south-east of the present day village of Dasburun [Tasburun].
In my opinion, the identification of Paracata with modern Bulagh-Bash also clarifies the position of the station of Coloceia (Ra Zotozeta), which was 33 Roman miles or 49 kilometers distant from Paracata and 33 Roman miles or 36 kilometers distant from Hariza. I propose to correct Coloceia (Ra Zotozeta) into Zolocerta (Colo--, Zoto = Zolouceia, zeta = certa). In this perfectly acceptable hypothesis, Zolocerta will correspond to the ancient Armenian Joghkert [Dzoghkert] which is called C'oghakert [Tsolakert] by Movses Xorenac'i (78A). The location of this station was incorrectly given by me in the first edition of the present work, as I showed in detail in my subsequent works (79). It is assumably about 26 kilometers north of the Caravansarai Pass, in the foothills south-west of Xosh-Xabar [Hosaber] and Nor Ch'aruxch'i. It seems to me that we may consider it more than probable that the Zogocaramentioned by Ptolemy (77° 15' long, and 42° 20' lat.), and whose position is given by him as being south-west of Artaxata, is to be identified with Zolocerta-Joghkert.
We must note that Belck (80), following the highly debatable indication of Movses Xorenac'i, and supposing that Jolkert-C'olakert was "at a distance of half a day's march for a man on foot" (81) from P'arhaxot, places C'olakert some 6-7 kilometers south-west of the present village of Dashburun and 12-14 kilometers west of ancient P'arhaxot. This explanation is evidently mistaken. The distance given by Xorenac'i is variously read by different scholars because of the alterations and variants in the manuscripts of Xorenac'i. Alishan, for example, argues this evidence differently and considers the same distance to be equal not to half but to a whole day's journey (82). This hypothesis also seems mistaken to me. In reality, in accordance with the evidence of Xorenac'i given above, C'olakert was distant from P'arhaxot "the space of one and a half day's forced march for a man on foot", as was shown by me in detail in my work, The Main Roads of Ancient Armenia (83). In this interpretation, the position of C'olakert is precisely located in the above-mentioned foothills north of Caravansarai.
The next station Hariza which lay 24 Roman miles or 36 kilometers south of Joghkert-C'olakert, is identified by me with the present-day locality of Aruch (84) located not far from modern Misun at a distance of approximately 10 kilometers south of the Caravansarai or Kujaghi Pass. The ancient name of modern Aruch was apparently "Haritsh", a name which is visibly identical with the indicated station of Hariza. The location proposed by me for Hariza is also confirmed by the fact that the distance [95] between Bagauna and Hariza shown on the Tabula, 24 Roman miles or 36 kilometers, coincides precisely with the distance between Bagawan (now Uc-k'ilise) and modern Aruch.
The clarification of the position of the stations of Zolocerta and Hariza solves the problem concerning the direction of the line Artaxata-Bagauna. This line ran not through Kalzwan nor through the Chingil Pass, as was supposed by Kiepert and Belck, but turned south from Zolocerta and descended to Bagawan through the Caravansarai or Kujaghi Pass which lay not far from the ancient Lake Gaylat, the present Balikgolu, some 10 kilometers north of Hariza-Aruch.
By determining the ancient mountain pass through which ran the route Artaxata-Bagauna, we can ascertain not only the invasion route of the Roman legions, of Pompey and other generals, into the province of Ayrarat, but also the probable route of the ancient Khaldians into the pre-Araxian plain and to ancient Armavir.
Through this very pass the transit route for international trade and the movement of caravans from China and Central Asia to the Black Sea
[96] must also be presumed to have crossed. The accuracy of this hypothesis may be proven if we bear in mind the position of cities in Armenia during the Roman period. It is of particular interest in this connexion that one of the most important commercial centres of Arsacid Armenia, the city of Erwandashat called "the great city" by Faustus of Byzantium, stood not far from a segment of the indicated transit route, not far from the confluence of the Axurean, the present Arpa-cayi, and the Araxes. South of the chain of Agri-dagi, and the Caravansarai Pass, not far from the junction station of Bagawan, was located the other city of Zarehawan. It seems to me that the location of Erwandashat and Zarehawan alone can serve as an indication that the highway from Artaxata to Bagawan did not go through Igdir, the Chingil Pass and Bayazid, but rather through the Caravansarai or Kujali pass, along the path indicated by us, from Joghkert through Aruch to Bagawan.
On the basis of a whole series of other indications, we may also assert with assurance that the region of the Caravansarai Pass, and the entire mountain district east of Lake Gaylat, had a pre-eminent strategic importance, and being the gateway of central Armenia, was protected by mighty fortresses. The ruins of these fortresses have survived to the present time and are described in detail by S. Haykuni in his interesting work Bagrewand (85).
In the mountain district east of Lake Gaylat, of which we spoke above, are also to be found the fortresses mentioned by Strabo, 'O lane (= castellum Volandum of Tacitus) andBaburas I indicated in detail in my work on The Main Roads of Ancient Armenia (85A). For this reason, the route of the Roman general Corbulo from central Armenia can also be determined: through K'eop'rik'eoy and Bagawan (Uc-kilise) and from there, through Aruch, into the mountains in the region of the Caravansarai Pass, where he besieged and captured the castle of Volandum and other fortresses. We may likewise suppose that the Babila of Ptolemy, whose location he gives as not far from Saguana (= Bagawan) and Sogosara. (= Zolocerta), is probably to be identified with the Babursa of Strabo.
According to the description of the Tabula Peutingeriana, the route from Artaxata divided at the junction point of Bagauna, one branch going to Satala and the other to Tigranocerta. The general direction of the first branch can be established with complete assurance from Bagawan, which lay north-west of modern Diadin, to the south of the Agri-dagi chain, at the point where the great caravan road now runs through Kara-kilise to Erzurum.
With the exception of Bagauna-Bagawan and Satala-Satagh only one of the stations listed above on this route has been identified correctly. This is Adconfluentes which lay at a distance of 133 Roman miles, or 197 [97] kilometers, from Bagawan and 169 Roman miles, or 250 kilometers, from Satala. Richard Kiepert associated Adconfluentes with the place where the Pasin-su flows into the Araxes near modern K'eop'rik'eoy. Indeed, judging from the distances given above, the approximate location of Adconfluentes near K'eop'rik'eoy is more than likely, as we shall see. It is confirmed by the recent clarifications, which we will give later, as to the location of the stations of Calcidava (Ra Chalchidara) and Armanas.
The following stations lay between Bagauna and Adconfluentes: Bagauna 24 Colchion 24 Chadas 17 Armanas 12 Andaga (Ra Andacas) 26 Barantea 30 Adconfluentes.
There are no indications as to the position of the station of Colchion, which lay at a distance of 24 Roman miles, or 36 kilometers, west of Bagawan. Its location is to be sought presumably not far from modern Kara-kilise. The next station Chadas (with a double tower), which lay at a distance of 48 Roman miles, or 72 kilometers, west of Bagawan, was unquestionably an important center on the plain of Alashkert. This station is to be associated in my opinion with the modern Xasdur (Hanzir, Chazdar) near which was found a cuneiform inscription of the king of Urartu, Menua. In the opinion of Lehmann-Haupt, Xasdur was the main center of the plain of Alashkert in the Khaldian period (86). From the indications of Haykuni (87), we see that in ancient times Xasdur was a fortress-town of which traces have survived until now. Markwart thinks it likely (88) that the name of the station of Chadas is to be associated with the name of Xad, bishop of Bagrewand, who was appointed to replace the kat'oghikos Nerses I (89). The accuracy of our identification of Chadas with Xasdur is confirmed by the fact that the distance between Bagauna and Chadas shown on the Tabula Peuiingeriana, that is 48 miles (=72 kilometers), corresponds to the distance between Bagauna and modern Xasdur.
The station after Chadas, Armanas, which stood 17 Roman miles (= 25 kilometers) from Chadas, is to be identified, as I have shown, and as was also correctly indicated by Markwart (90), with the village of Aramanay mentioned by the Armenian historian Ghazar P'arpec'i (90A). According to the account of Ghazar, Vardan Mamikonean, on his way from Persarmenia to Byzantine Armenia, stopped to rest in the village of Aramanay which lay in the province of Bagrewand not far from Basen and Twarcatap'. From this testimony, it is clearly visible that the village of Aramanay lay on the road from Ayrarat to Theodosiopolis, between modern Basen and the plain of Alashkert. Consequently ancient Aramanay coincided both in name and location with the station of Armanas which lay at a distance of 65 Roman miles (= 97 kilometers) from Bagawan.
The nearest station Andaga (Andacas according to the Ravennatis Anonymi Cosmographia), next to which there are two towers in the Tabula Peutingeriana, lay at a distance of 12 Roman miles (=18 kilometers) from [98] Armanas. In name it corresponds to the ancient Armenian Antak' which is mentioned in the Ayrarat of Alisan (90B). We must also note that, on the ten-verst map of the General Staff, there is the indication of a village of Endek near the Pass of Karaderbent and 6-7 kilometers to the south of it. The name of this village coincides almost exactly with Andaga. The position of ancient Antak' which corresponds to the Andaga of the Tabula is not precisely known.
Barantea, the station following Andaga (Ra Andacas), lay 26 Roman miles (=39 kilometers) from Andaga evidently in Upper Basen approximately 45 kilometers (=30 Roman miles) north-west of the present K'eo-p'rik'eoy (= Adconfluentes). According to the information of Movses Xorenac'i, the province of Upper Basen was also called Vanand in ancient times (90C). We may, therefore, suppose that the name of the station of Barantea, which is to be corrected to Banantea, corresponds to the ancient Armenian Vanand, of which the genitive is Vanandeay. This province was probably called Vanand after the name of the main centre of the district, Banantea, which has been preserved in the Tabula Peutingeriana.
According to the drawing of the Tabula Peutingeriana, what is meant by the station Adconfluentes, is not the actual point of meeting of the Pasin-su and the Araxes, as Kiepert mistakenly believed, but rather the preceding station. The location of this station may be supposed on the site of modern Hasanghala [Hasankale], as I have shown in detail in my work on The Main Roads of Ancient Armenia (90D). This may be seen from the fact that the distance indicated in the Tables between the stations of Chalchdiara (= Xaghtoy-arich) and Adconfluentes, namely 57 Roman miles (= 85 kilometers), corresponds almost exactly to the distance between Xaghtoy-arich (the modern Gaghtarich) and Hasankale.
The following stations are indicated on the next section of the road on the Tabula Peutingeriana: Adconfluentes 10 Datamisa 20 Tharsidarate 15 Autisparate 12 Calcidava (RaChalchidara) 15 Sinara 22 Lucus Basaro 15 Aegea (Ra Egea) 20 Darucinte 20 Salmalasso 20 Satala.
There are no indications in the ancient Armenian sources concerning three of these stations: Datamisa, Tharsidate and Autisparate. There can be no doubt, however, that the station following Autisparate, Calcidava (Ra Chalchidara) corresponds to the ancient Armenian Xaghtoy-arich which is identical with the modern Gaghtarich and is located near the Euphrates crossing near Ashkale.
Two large rivers are indicated in the Tabula Peutingeriana on the way from Bagawan to Satala: the Araxes, between the stations of Barantea and Adconfluentes, and the Euphrates, the present Kara-su, near the station of Calcidava or Ra Chalchidara. Konrad Miller, who ignores the indications of the Tabula as to the disposition of these rivers, identifies them [99] with the ancient Armenian Murc, now called Egri, or Araz, and with the Euphrates-Arsanias, the present Murad-cayi (91).
In accordance with the indication of the Tabula, the crossing of the Euphrates (Kara-su) took place not far from Calcidava. This information is of the greatest value. On the road from Satagh to K'eop'rik'-eoy, precisely at the crossing of the Euphrates, is still to he found the village of Gaghtarich which is identical with the ancient Armenian Xaghtoy-arich. This interesting coincidence has not been sufficiently appreciated by the investigators of the Tabula. Konrad Miller identifies the station of Calcidava (Ra Chalchidara) with the ruins of Arablarmezar, or modern Kara-kilise, which stands at a distance of 10-15 kilometers to the north of the Egri or Artaz river (92), while Richard Kiepert supposes that the same station is to be located on the site of the present Alaca which stands between the Kara-su and Erzurum. The latter supposition is incorrect if only because the position of the station of Calcidava is marked on the Tabula Peutingeriana on the left and not the right bank of the Euphrates-Kara-su. I propose that Calcidava, or Ra Chalchidara, be corrected into Chalchdiara, or Chalchdiaris, and I presume that this name corresponds to the Ancient Armenian Xaghtoy-arich which is to be identified with the modern Gaghtarich.
The nearest station following, Sinara, is correctly identified by Miller as the fortress of Sinoria (93), (Sinoria in Strabo, Sinora in Plutarch (93A), which is mentioned by Strabo and other ancient sources. The location of Sinoria has been studied in detail and clarified in my historico-geographical work on The Main Roads of Ancient Armenia (94). Ancient Sinoria or Sinara, which lay 15 Roman miles (= 22 kilometers) west of Chalchdiara, was the fortress and treasury of Mithradates Eupator. According to the account of Plutarch, Mithradates, pursued by Pompey, took refuge in this very fortress in which were kept all his treasures and from there he fled to Colchis (95). According to Strabo, Sinoria was an important fortress in the Pontic kingdom of Mithradates and "was close to the borders of Greater Armenia and this is why Theophanes changed its spelling to Synoria (96).
Judging by the last statement of Strabo, and bearing in mind the position of the station of Sinara on the bank of the Euphrates-Kara-su, we can determine precisely the frontier between the Pontic kingdom of Mithradates and Greater Armenia. On the road from Gaghtarich to Satagh, the location of the station of Sinara, standing 22 kilometers (=15 Roman miles) from Gaghtarich, could of course not be on the site of the present Sankarich near Erzurum, as was believed by Miller (97), nor near modern Alaca, as is indicated on Kiepert's map. It stood on the southern slopes of the Kop-dag between modern Ashkale and Lich. Consequently, the frontier between the Pontic kingdom and Greater Armenia is also [100] precisely determined as being in the district between Ashkale and the Kop-dag range.
The location of the following station, Lucus Basaro (= the grove of Easaro), lying 22 Roman miles (= 33 kilometers from Sinara) may be presumed to be in the plain of Bagarich mentioned by the ancient Armenian historian Asoghik (98). If we correct Basaro into Bagara (g instead of s), this station will correspond to the ancient Armenian Bagarich (cf. Chalchdiara = Xaghtoy-arich). The location of Lucus Basaro is to be found not in modern Bagarich itself but on the right bank of the Euphrates, as I have shown in detail in my work on The Main Roads of Ancient Armenia (98A).
In the first edition of the present work I identified the station following Lucus Basaro, Aegea (Ra Egea), in agreement with Miller (99), with the ancient Elegeia of Roman sources. This hypothesis has, however, proved to be mistaken. The location of Elegeia is to be presumed on the site of the modern village of Alaca standing not far from Erzurum and near the sources of the Euphrates, as has been correctly shown in historical works (100). Consequently, both the name and the location of the station of Aegea (Ra Egea), lying 15 Roman miles (= 22 kilometers) from Lucus Basaro, remain, as we see, still unelucidated.
We must note that the road Calcidava-Sinara-Satala did not run through Mamahat'un to Erzurum, as I pointed out in my work The Main Roads of Ancient Armenia (100A), but followed another path which is mentioned in the topographical sketches of V. Shtrekker who says that,
... between Erzurum and Erzincan there is also another, somewhat shorter road which used to be greatly used formerly since it wras often part of the caravan route from Erzurum to Constantinople (now abandoned with the organization of boat connexion between the latter city and Trebizond). This road runs through Alaca to Ashkale...One and a half hours from Ashkale begin the Shogan khans which stretch for the space of two hours along the right bank of the Euphrates (101).It seems to me that it is precisely along this ancient route and on the right bank of the Euphrates that we must look for the path of the road Calcidava-Sinara-Lucus Basaro, which then turned aside in the direction of Aegea.After the station of Aegea, the Tabula lists the stations of Darucinte 20 Salmalasso 20 Satala. Unfortunately neither Darucinte nor Salmalasso can be exactly localized as yet. Concerning Darucinte, I ventured a hypothesis with the reservation that the name of this station, which was evidently found in the ancient Armenian province of Derjan, might be related to the name Derxene-Derjan and need to be rectified as Darxinte (102).
2. Artaxata--Tigranocerta:
The following stations are listed by the Tabula Peutingeriana along the ancient road from Artaxata to Tigranocerta:Artaxata 23 Paracata 33 Coloceia (Ra Zotozeta) 24 Hariza 24 Raugonia (Ra Ragauna) 39 [Catispi 27 Sorue 24 Anteba 24 Nasabi 17] Isumbo 15 Flegoana 15 Dagnevana (Ra Dognavana) 26 Molchia 32 Vastauna (Ra Bastavena) 26 Patansana (Ra Patransana) 27 Dyzanas (Ra Dizanas) 22 Cymiza 20 Zanserio (Ra Zancerion) 30 Tigranocarten (103).The figures and names in brackets are also indicated in the Tabula on the adjacent road from Artaxata to Ecbatana. It is quite evident that the same stations were copied twice by mistake. These stations were, as I have shown in detail in my study of The Main Roads of Ancient Armenia (103A), the stations of the road, Artaxata-Gobdi-Paresaca. The Ravennatis Anonymi Cosmographia, which made use of an older and better copy of the Tabula, gives instead of these stations, Didyma, Indua and Arachia which, as we shall see, really were on the route, Artaxata-Tigranocerta. The distances between these stations are unfortunately not given by the Ravenna Anonymus.According to the description of the Tabula Peutingeriana, the road from Artaxata divided at the junction of Bagauna, as we have already seen, with one branch going to Satala and the other to Tigranocerta. Of the fourteen stations listed above on the second branch, only the first Bagawan and the last Tigranocerta were known to western scholars. Consequently the major part of the remaining stations has been studied in detail only in the last decade.
The road from Bagawan to Tigranocerta is indicated on Miller's explanatory map as running along the southern shore of Lake Van or through Parnaut, Melazkert [Manazkert], Mush, Bitlis. It is evidently impossible to accept the hypothesis of Miller. Not only has the location of Bagauna been given incorrectly by him, but the same is true of the position of the end point, Tigranocerta. Consequently not only the initial and final points but all of the stations of the route are incorrectly determined by Miller.
The second error of western scholars has been their identification of the stations of Dagnevana and Vastauna with Datvan (now Tadvan) and Vostan which stand on the shore of Lake Van (104). These identifications are in open contradiction to the distances indicated in the Tabula Peutingeriana, as was noted by Miller. Indeed Dagnevana is listed as the seventh station on the road from Tigranocerta to Bagawan and Vastauna [102]
as the fifth. The distance between Tigranocerta and Dagnevana is given as 183 Roman miles or 211 kilometers, while that between Tigranocerta and Vastauna is 125 Roman miles or 185 kilometers. In fact, however, modern Datvan, which was identified with Dagnevana, stands on the south-western shore of Lake Van and is closer to Tigranocerta than Vostan which stands on the south-eastern shore of Lake Van, opposite the island of Aght'amar. We may, therefore, conclude with certainty that the indicated correspondence of names is fortuitous and that this coincidence has certainly no significance in the determination of the route, Tigranocerta-Bagawan.As we have already seen, the location of Tigranocerta was clarified by the German scholars W. Belck and K. Lehmann-Haupt. After careful investigations carried out by them in the vicinity of modern Sgherd and Farkin [Miyafarkin], Tigranocerta was identified by them with modern Farkin and this opinion has now been accepted by all Armenists.
With the identification of Raugonia with Bagawan and of Tigranocerta with Farkin, the problem as to the direction of the route under [103] discussion is entirely clarified. This road, as we shall see, ran from the province of Bagrewand, in which was located the junction station Bagawan, through the provinces of Apahunik' and Hark' into the province of Taron and from there, through the modern Shinik and Nerjik, it came down to Farkin-Tigranocerta.
The stations which follow Bagauna in the Tabula Peutingeriana: Nasabi, Anteba, Sorue, and Catispi, were, as was already said, stations on the adjacent road from Artaxata to Ecbatana. On the line Artaxata-Tigranocerta, however, after Bagauna stood the stations of Didyma, Indua and Arachia, as is evidenced by the Ravennatis Anonymus. And, indeed, the first station after Bagauna, Didyma still exists at the present day, bearing its ancient name Didem, as was correctly shown by Xach'atrean (in the documentation to the report read by him at the Institute for Cultural History of the Armenian S. S. R.). The village of Didem, which is to be identified with Didyma, lies west of Bagauna near the left bank of the Arsanias.
Armenian sources give us no information about the name and position of the station of Indua. Howrever, the station of Arachia given by the Ravenna Anonymus, may, in my opinion, be related to the name of the province of Archene through which flowed the Euphrates-Arsanias (104A). In the opinion of Markwart, the province of Archene is identical with the ancient Armenian province of Hark' and included in antiquity, in addition to the Hark' of ancient Armenian sources, Apahunik' and other adjacent districts (105). If we admit the relation between Arachia and Archene, we may suppose that this province was named Archene after its main centre Arachia which has been preserved in the Tabula Peutingeriana. The difference in the endings of these names, Arch-ene and Arach-ia, is easily explicable if we bear in mind that ancient Com-mag-ene, Melit-ene and Mati-ene are correspondingly rendered as Com-mag-ia, Melit-ia and Mat-ia in the Greek inscription on the north wall of Farkin-Tigranocerta (106).
In addition to this, the explanation offered by me for the three nearest stations: Isumbo 15 Flegoana 15 Dagnevana (Ra Dognavana), is also of decisive importance for the determination of the direction of the road Artaxata-Tigranocerta through the provinces of Apahunik', Hark' and Taron given above. According to the distances between Tigranocerta and these stations, they should fall precisely in the provinces of Apahunik' and adjacent Hark' in which the ancient Armenian sources mention the localities of Cumb [Tsumb], Elegakan and Donewank' (107). The exact location of these localities is not known, but, since Cumb was in Apahunik' and Elegakan and Donewank' in Hark', it is perfectly evident that Cumb lay north-east of the other two stations and had the same relation to them as did Isumbo to Flegoana and Dagnevana.
[104] The station of Isumbo (with two towers), whose name I propose to correct into Tsumbo, evidently corresponds to the ancient Armenian settlement of Cumb which was located according to the evidence of Asoghik on a raised and fortified spot in the province of Apahunik' adjacent to the piovince of Hark' (108). The location of the station of Tsumbo may be supposed as being north-east of Manazkert, in the vicinity of modern Patnos or Zomak, as is indicated in my detailed study, The Main Roads of Ancient Armenia (109).
The following station Flegoana is identical with the old Armenian village of Elegakan which lay in the province of Hark', according to the indication of Asoghik (110). According to the evidence of Miller, there is a crack in the parchment between the o and the a of the word Flegoana (110A). We may, therefore, suppose that there was a letter between o and awhich can no longer be deciphered. On the basis of the correspondence between Flegoana and the Armenian Elegakan which I have proposed, I suppose that Flegoana is to be corrected into Elegacana.
We must note that professor Markwart read Flegosana for Flegoana and suggested that Flegosana be altered into Elegosine. He related this name with the Armenian eghegashen "built from reeds", or eghegnacin "reed bearing". Since Markwart identified Dagnevana with present-day Datwan, he supposed that the location of Flegoana-Elegosine was 15 Roman miles from Datwan, on the north shore of Lake Van (111). Both the suppositions of Markwart, as to the name as well as the position of Flegoana, are, as we see, mistaken.
With the station of Dagnevana (Ra Dognavana), which lay 15 Roman miles (=22 kilometers) from Flegoana, we must evidently identify the above mentioned Donewank' (= probably Dadnevana) which was also to be found in the province of Hark'. We know that the ancient Armenian provinces of Apahunik' and Hark', through which ran the road from Bagawan to Tigranocerta, were situated, the first north of Manazkert, along the course of the Arsanias, and the second to the west of it. Consequently, the direction of the route under consideration can be established, not along the southern or northern shores of Lake Van, as was supposed by Miller and Markwart, but through Bagrewand, Apahunik', Hark', Taron and the chain of the Taurus.
With the elucidation of the locations of the stations of Didyma, Arachia, Isumbo, Flegoana, Dagnevana, we can also determine, in my opinion, the route of the retreat of Xenophon and the Ten Thousand (112), as well as the route of Corbulo, who went from Artaxata to Taron through the ancient Armenian provinces in which these localities were found (113). Most scholars have supposed up to now that Corbulo, going from the vicinity of Artaxata to Taron and Tigranocerta, sustained the attack of the Mardians in the province of Mardastan which lay east of Lake Van (114). In [105] this hypothesis, the route taken by Corbuto is determined as the one through Berkri, Vostan and the southern shore of Lake Van to Taron and from there to Tigranocerta. Professor Markwart has shown quite correctly the incredibility and unsoundness of the presumption that Corbulo moved from Bagawan to Tigranocerta through the long and round-about route through Berkri and Vostan to Taron and thence to Tigranocerta (115). Markwart himself is of the opinion that the province of the Mardians crossed by the Roman legions of Corbulo must be sought to the north of Taron. The hypothesis of Markwart agrees approximately with the evidence presented by us as to the direction of the road Bagawan-Tigranocerta. On the basis of this new evidence, it seems to me that we may seek the country of the Mardians, mentioned by Tacitus, in the mountainous district of the present-day Ala-dagi.
The station following immediately Dagnevana, Molchia, cannot be explained. The next station, however, Vastauna (Ra Bastavena), which lies at a distance of 32 Roman miles (= 48 kilometers) from Molchia was located, as we can easily see, on the site of modern Bostak'end which was evidently called in antiquity Bast-awan (= Basauna or Bastavana). The village of Bast, which has the same name, still exists at the present time in the Sparkert district of the vilayet of Bitlis (116).
The next station Patansana (Ra Patransana), which lay at a distance of 26 Roman miles (= 39 kilometers) from Vastauna, corresponds to ancient Armenian Patrans-awan. For this reason the name of this station should be corrected to Patansauna (cf. Bagauna--Bagawran). The location of Patan-sauna is to be sought not far from Mush. A village of the same name, Patras, exists at present in the Sparkert district of the vilayet of Bitlis (117).
After Patasana, the Tabula lists the stations of Dyzanas (Ra Dizanas) 22 Cymiza 20 Zanserio (Ra Zancerion) 30 Tigranocerta.
In my detailed study, The Main Roads of Ancient Armenia, I suggested that the name of the station of Dyzanas (Ra Dizanas) was probably a deformation of Arzanas (= Arzan, syr. Arzon) (117A). This supposition is, to be sure, merely a guess and must still be confirmed by some evidence. There is, however, no doubt that the next station Cymiza, 22 Roman miles (= 33 kilometers) from Dyzanas, is identical with modern Kildiz which is noted on the ten-verst map of the former General staff as being on the road from the plain of Mush to Farkin-Tigranocerta, north of the village of Nerjik. I propose that the disfigured name of the station be corrected from [Greek alphabet] CYLDIZA to CYMIZA (LD = M).
The next station Zanserio (Ra Zancerion), at a distance of 20 Roman miles (= 30 kilometers) from Cymiza-Kildiz, may be associated with the Zoniqart mentioned by Barhebraeus (118). It is entirely possible that the Zanserio or Zancerion given above is a deformation of Zancerta. The location of Zoniqart is given by Barhebraeus as being at Hesna de-Zajid, near [106] the confluence of the Deba and the Arsinos. On the basis of this evidence, Markwart has supposed the location of Zoniqart as being near the Euphrates-Arsanias, opposite modern Pistek (119). It seems more likely to me that by the Arsinos is meant not the Euphrates-Arsanias but the present Kulp-su which is found in the ancient province of Arzanene (= Syr. Arzon).
The association of the stations of Cymiza and Zanserio with modern Kildiz and Zoniqart fully confirms the identification of ancient Tigranocerta with modern Farkin which had already been throughly demonstrated in the works of Belck and Lehmann-Haupt. Indeed the position of Farkin-Tigranocerta, 50 Roman miles (= 75 kilometers) from Cyldiza, is precisely at the required distance south of this station which corresponds to modern Kildiz.
We must note that the direction of the ancient route from the Plain of Mush to Tigranocerta was correctly given in the works of the German scholars Karbe and Eckhardt. Karbe gives the route of Xenophon and the retreating Ten Thousand along the route indicated by us through modern Nerjik, Kilgiz, and Shinik (120). Similarly Eckhardt supposes that the most convenient route for the movement of the army of Lucullus from Corduene to the plain of Mush was the road through Nerjik and Shinik (121).
Thus, in accordance with the new data given above, the road from Bagauna to Tigranocerta can be clearly determined through modern Didem, Zomek, Bostak'end into the Plain of Mush, and from there along the above-mentioned route to Farkin-Tigranocerta.
Continued on Next Page
Footnotes
Page opens in a separate window.
Classical Antiquity Menu
Historical Sources Menu
History Workshop Menu
No comments:
Post a Comment