Chapter VI (2)

The two most important, and to which the greater part of the population of the others was drawn as early as the time of the successors of Alexander, were Alexandria and New Ilium; and these continued to be the chief towns under the Roman emperors. Alexandria has preserved considerable remains to this day. Of New Ilium only the foundations of the walls with a few other fragments are to be seen.

As Hamaxitus, Larissa, and Colonæ, were from their proximity to Alexandria absorbed by that city at the time of its foundation[424], we are not surprised that no remains of them have been remarked by travellers. Some circumstances, however, mentioned by Strabo[425], are sufficient very nearly to fix their positions. Hamaxitus in particular is determined by the salt-works of Tragasæ, which are still in a state of operation on the sea-coast near the mouth of the river of Tuzla. This river (perhaps the ancient Satnioeis) does not, however, take its name, which means _salt_, from the maritime salt-works alone: there are other salt-works at some very copious sources of hot salt water, at a considerable distance from the sea, on the northern side of the valley, where is a village called Tuzla, and where the neighbouring hills are composed of rock salt. This curious fact accounts for the name Halesium, anciently applied to the district[426].

As it appears from Strabo that Cebrenia bordered on the territories of Antandrus, Hamaxitus, Neandria, New Ilium, and Scepsis[427], and that the Scepsia was on the Æsepus[428], consequently on the eastern side of the summit of Ida,—Cebrenia seems to have occupied the higher region of Ida on the western side, and its city very probably stood at Kushunlú Tepe, not far from Bairamitsh, where Dr. E. D. Clarke, proceeding from the latter place towards the sources of the Mendere and the summit of Ida, found very considerable remains of antiquity. The fine valley which extends from thence to the modern town of Ene, seems to answer in its upper part to the level country of Cebrenia, mentioned by Strabo[429]; and in its lower or western to the plain called Samonium, which belonged to Neandria[430]: for Neandria being described by the geographer as inland from Hamaxitus towards New Ilium, and as 130 stades distant from the latter[431], corresponds exactly in position with Ene.

In the plain of Troy, or region watered by the lower course of the Mendere and its branches, the only positions proved to be ancient sites, by remains of buildings existing in their original places, are—

1. That of New Ilium on a hill which rises to the eastward of the villages of Kum-Kiúi and Kalafátli, about 5 miles to the S.E. of Kum-Kalé or the lower castle of the Dardanells, and three miles from the nearest shore. The vestiges of the walls of the citadel are to be traced on the summit of the height; and some of the buildings of the town, on the western slope and at the foot of the hill: but very little now remains in its place, the site being resorted to (as it probably has been ever since its abandonment), as to a stone-quarry, for the materials of modern constructions—whence we find all the villages, farms, and particularly the Turkish cemeteries of the surrounding country, full of the inscribed or decorated marbles of New Ilium. 2. Paleó Aktshi Kiúi. This, by its direction and distance from New Ilium, corresponds exactly with the Ἰλιέων κώμη, or village of the Ilienses, described by Strabo[432] as being 30 stades eastward of New Ilium towards Ida and Dardania. 3. Paleó-Kastro, near the Turkish village of It-ghelmés, on a height overlooking the Bosphorus. This is probably the site of the town Rhœteium, on a part of the sea-shore of which was the Æanteium or tomb of Ajax[433], still existing. 4. Yenishehr, the ancient Sigeium. 5. Another Paleó-Kastro, near the mouth of the small river which receives the canal derived from the river of Bunárbashi. This has been supposed, with great probability, to have been a small town and port called Agameia[434]. 6. The hill which rises above the less or lower Bunárbashi to the S.E., and which is bounded in the same direction by the deep valley of the Mendere. This, it is not improbable, was the site of Scamandria; for it may be presumed that Scamandria being named by Pliny together with New Ilium[435], was in some part of the lower plain of the Scamander, near that river; and there is no site on the Mendere so remarkable as that of Bunárbashi. Pliny describes Scamandria as a _small_ town: but it seems from an extant inscription to have been of sufficient importance to make a recorded treaty with New Ilium concerning the sale of corn[436].

The same heights are by many persons supposed to have been in an earlier age the position of the renowned capital of Ilus and his successors: indeed, so many of the most intelligent _travellers_[437] in the Troas are agreed in placing the Homeric Ilium at Bunárbashi, that I should have been satisfied on the present occasion with stating my concurrence with their opinion, and with referring to the arguments of such of them as have supported it by their publications, had not some adverse systems been recently maintained with great learning and ingenuity; though chiefly, it must be admitted, by those who have considered the question in the closet only. I shall here offer, therefore, a few observations on this subject; first stating what appear to me to be the strongest grounds for thinking that Bunárbashi was the site of Troy, and then the principal objections that have been made to that opinion, together with the arguments which occur in reply to them[438].

As even the identity of the country on the Asiatic side of the entrance of the Hellespontine strait with the scene of the Ilias has been doubted, it may not be useless to premise, that if the war of Troy was a real event, having reference to a real topography (and to doubt it would shake the whole fabric of profane history), no district has yet been shown that will combine even a few of the requisite features of the plain of Troy, except that between Kum-Kalé and Bunárbashi: whereas in that district, and in the surrounding country by land and by water, we find the seas and mountains and islands in the positions which the poet indicates, and many of them with the same or nearly the same names. The features which do not accord so well with his description are those which are the most liable to change in the lapse of ages,—the course and size of the rivers, and the extent and direction of the low coast where these waters join the sea. Instead of a river with two large branches, which Homer seems to describe, we find on one side of the plain a broad torrent, reduced in the dry season to a slender brook, and a few stagnant pools; and on the other side a small perennial stream, which instead of joining the former is diverted into an artificial channel, and is thus carried to a different part of the coast. But the diminutive size of some of the most celebrated rivers of antiquity is well known to those who have travelled in Greece; and it must be considered that a poet writing of a real scene is obliged to magnify those features, which without exaggeration would be beneath the dignity of his verse. In regard to the course of the streams, it seems sufficient still to find, at the end of three thousand years, two rivers which, if they do not now unite, evidently did so at a former period of time: and for the sources of that stream which Homer describes as rising under the walls of Troy, to find some very remarkable springs, not very different in their peculiarities from the poet’s description, and rising at the foot of a commanding height on the edge of the plain.

For poetry this coincidence appears sufficient: and in regard to the position of Troy itself, it seems enough to find a hill rising above the sources just mentioned, not only agreeing in all particulars with the kind of position which the Greeks[439] usually chose for their towns, but the only situation in this region which will combine all the requisites they sought for; namely, a height overlooking a fertile maritime plain,—situated at a sufficient distance from the sea to be secure from the attacks of pirates, and furnished with a copious and perennial supply of water,—presenting a very strong and healthy position for the city; and for the citadel a hill beyond the reach of bowshot from the neighbouring heights, defended at the back by steep rocks and precipices, surrounded by a deep valley and broad torrent, and backed beyond the river by mountains which supplied timber and fuel. That it was precisely such a situation as the inhabitants of Greece and Asia in remote ages preferred, might be shown by a great variety of examples: and it can hardly be doubted that a person totally unacquainted with the Ilias, but accustomed to observe the positions of ancient Greek towns, would fix on Bunárbashi for the site of the chief place of the surrounding country.

It is a necessary consequence of placing Troy on the heights to the S.E. of Bunárbashi, that the river flowing from the sources which give that village its name (meaning Spring-head), is the Scamander of Homer: that the large torrent which flows through a deep ravine on the eastern side of the heights, is the Simoeis: and that notwithstanding the much greater magnitude of the bed of the latter and occasionally of that stream itself, the united river after the junction in the plain was called by the name of the former, Scamander. In support of this opinion, it has been justly observed by Lechevalier, that Homer’s description, allowance being made for poetical exaggeration, is correct, both as to the springs themselves, and as to the very different character of the two rivers: nor can it be denied that the two hills, that of Bunárbashi and the higher eminence behind it, correspond to the mention by Homer of Ilium and its citadel Pergamus. The termination of the slope towards the springs accords also with the idea which we receive from the poet of the extent of the city on that side, and of the position of the gate Scææ or Dardaniæ, which was near the sources of the Scamander, and was the principal outlet towards the plain[440]. But if these assumptions are not unreasonable, it cannot be denied on the other hand that in attempting to identify such objects as the tombs of Ilus, Myrinna, and Æsyetes, Lechevalier has exposed himself to reasonable objections from his opponents, and has rather injured than strengthened his cause. For it is not certain that all the monuments mentioned by Homer were tumuli; and it is very possible that if they were, several of them have been obliterated by time. Nothing can be more likely than that the real history of the monuments should have been forgotten in the interval between the destruction of Troy and the foundation of New Ilium, and that names should have been ascribed to them by the inhabitants of the latter place, suited to their own system of Trojan topography, and favourable to the pretensions which they held, that their city stood upon the ancient site. With regard to the existing barrows, it seems incontrovertible only that those which stand in conspicuous situations on either side of the mouth of the Scamander, are the tumuli, supposed in the time of the Romans, and probably with reason, to have been the sepulchres of Ajax, Achilles, and some other chieftains; and these monuments are so far important, as they prove the identity of the plain of the Mendere with the scene of the Ilias[441].

It is objected to the springs of Bunárbashi, that instead of being only two,—one hot and the other cold, as described by Homer[442],—they are in one place so numerous as to have received from the Turks the name of Kirk-Ghiuz, (the Forty Fountains), and that they are all of the same temperature.

But viewing them as the springs of a river, they may in poetical language, or even in common speech, be considered as two, since they arise in two places, distant from each other about 200 yards: in one the water appears in a deep basin, which is noted among the natives for being often covered with a thick vapour like smoke: in the other place, there are numerous rills issuing from the rocks, into a broad shallow piece of water, terminating in a stream which is joined by that from the smoking spring. As to the temperature of the water, the observations of travellers give various results. Some have observed a difference: according to others, it would appear that being all deep-seated springs, their temperature is the same at all seasons, or about 60° of Fahrenheit at their eruption from the ground; consequently that they will feel cold when the air is at 70° or 80°, and warm when it is at 40° or 50°[443]. But even in this case it is obvious that there will be a real difference between the heat of the shallow recipient of the springs called the Forty Fountains, and that of the single deep pool. It seems sufficient to justify Homer’s expression, that a difference of temperature was believed, and that an occasional appearance of vapour over one source was often observed by the natives: for the poet would probably flatter the local prejudices, even if he had examined the fountains so attentively as to be convinced that the warmth of all the sources was the same.

Another and a more weighty objection to the placing of Troy on the heights of Bunárbashi, is that the much greater magnitude of the river, which flows on the east side of those heights, concurs with its modern name Mendere in showing it to be the Scamander of Homer; and that such was evidently the opinion of several authors of antiquity, particularly of Demetrius, a native of Scepsis in the Troas, from whom Strabo principally derived his information on the geography of this district. In fact there can be no doubt, that in the time of Demetrius, who wrote in the second century before Christ[444], the Mendere from its source in Mount Kazdagh to its junction with the sea was called Scamander. But was it so in the time of the Trojan war? In this inquiry we have nothing to do with any authority but that of the Ilias itself: for it is evident from the remarks of Demetrius and Strabo, that the topography of the poem and the site of Troy were as much a subject of doubt and dispute in their time as they are at present. Nor is this surprising. The result of the Trojan war was the subversion of Ilium and the extinction (with the exception of a single branch of the royal family) of the colony which had settled in this part of Phrygia[445]. Strabo repeatedly remarks that the revolutions following the Trojan war were the great cause of the difficulty which he experienced in adjusting the Homeric chorography. The barbarous people of Thrace, called Treres, who then established themselves in the Troas, could not have taken much interest in any thing relating to the former colony, to whose language they were strangers, and whose history was recorded only in the songs of an Ionian stranger. It was not till long afterwards that the Æolian Greeks of Lesbus extended their settlements into the Troas. It was not even by them that New Ilium was founded, but by a Lydian, and consequently a semibarbarous colony[446], about the eighth century before Christ; and it was not till a taste for the poems of Homer having begun to prevail in European Greece, and the Athenians having taken possession of Sigeium[447] and a part of the Chersonesus, that their enlightened sovereigns Pisistratus and his sons[448], if they were not the first to collect, arrange, and edit the Ilias,—were at least the first to bring it into notice among the most lettered of the European Greeks[449]. We cannot wonder that the Homeric topography should at that time have become subject to the same kind of uncertainty now found to prevail in regard to such places as Athens, Rome, Jerusalem, Alexandria of Egypt, and even many cities much more modern.

For the New Ilium founded by the Lydians, colonized afterwards by the Æolians, and augmented and first fortified with a circuit of forty stades by Lysimachus[450], a situation was chosen which, being nearer to the sea than that of the ancient city, was better adapted to the more advanced state of commerce and civilization[451]. It was very natural that its inhabitants the Ilienses[452] should pretend that their town stood on the site of the ancient city[453]; and no less so, that a historian of a neighbouring and kindred race should flatter them by concurring in their opinion[454]. That the conquerors of Asia likewise, and so many other illustrious visitors of Ilium from Xerxes to the Cæsars, when they found it useful to their purposes or grateful to their vanity to sacrifice to Minerva Ilias, should have willingly followed the guidance of the priests to the temple in New Ilium, and should have admitted without inquiry that it stood on the site of the Pergamus of Priam—is nothing more than we should expect under such circumstances. But we know that the claim of the Ilienses was strongly contested during the whole period in which their city flourished. Demetrius of Scepsis and Hestiæa of Alexandria Troas opposed it about the time of the Antiochian war, and Strabo subscribed to their opinion in the Augustan age[455].

Although Demetrius found it impossible to assent to the claim of the Ilienses in this respect, and seems to have been far from implicitly believing in the identity of all the Homeric places pointed out by them[456]; he appears never to have suspected that the Scamander was any other than the large torrent, to which he found that name then applied from its mouth in the Hellespont to its distant source in the summit of Ida called Cotylus[457]. It was a necessary consequence (as all those who have concurred in the same belief have experienced) to identify the Simoeis with one of the branches of the Mendere flowing from the eastward. The Ghiumbrek-su, the most important of the Trojan streams after the Mendere and Bunárbashi river, seems to have been that which Strabo (probably following Demetrius[458]) supposed to be the Simoeis, as may be inferred from his observation that the site of Troy, which he places at the Pagus Iliensium (Paleó Aktshi), was near the river Thymbrius; and that the temple of Apollo Thymbræus at the junction of this river with the Scamander, was 50 stades from New Ilium[459]; for these data concur in showing that the Kamára-su[460] was the Thymbrius, and consequently that the Ghiumbrek-su was the Simoeis of the geographer.

But although a site had been found for Troy at Pagus by those who did not subscribe to the claims of the Ilienses in favour of their own site, neither Demetrius nor Strabo was able to discover any springs corresponding to the Scamandrian sources of Homer. Demetrius, having observed how utterly irreconcileable the single source of the Scamander in the distant summit of Mount Ida is with Homer’s description of the Scamandrian springs, was under the awkward necessity of imagining that those fountains, wherever they might be, were called the springs of Scamander, not as being in reality the sources, but only because they were near the Scamander, or because they afforded a stream which joined that river[461]. And as the valley and river of Ghiumbrek do not unite with the plain and river of the Mendere till very near the sea, Demetrius distinguishes the Simoeisian from the Scamandrian plain[462]—a distinction, it may be observed, which no where occurs in Homer, and is in fact inconsistent with his topography.

There seems no other mode of obviating these difficulties, inevitably attendant upon taking the Mendere in its whole course for the Homeric Scamander, but to suppose that the river of Bunárbashi was the _ancient_ Scamander, that it gave name to the united stream, and that the part of the Mendere above the junction was the Simoeis. The latter name appears to have become obsolete during the ages in which the events of the war of Troy had been almost forgotten on the scene itself, and in the time of Demetrius and Strabo to have been known only to antiquaries inquiring into the topography of the Ilias. The name of Scamander on the other hand, being the more illustrious of the two, and a name apparently of familiar import in Asia Minor[463], was retained in use: but as the river of Bunárbashi had lost much of its local importance, and had now become of inferior consideration, the name of Scamander before attached to the united stream and to the Bunárbashi-su, was after the revival of New Ilium by Lysimachus (and perhaps long before that time) applied to the united stream and to the whole course of the Mendere.

In some of the preceding pages we have had occasion to remark in the instances of the Sangarius, Mæander, and Hermus, how easily the names of two branches of a river are confounded with one another or with the united stream, and how readily they are transferred from the one to the other. In addition to these examples, it may be observed that a similar transmutation of name in two branches of the same river, under circumstances which cannot so easily be accounted for as in the Trojan rivers, is to be found in Thessaly, where the river called by Herodotus and Thucydides Apidanus, is undoubtedly the same as the Enipeus of later writers, whose Apidanus is at twelve miles distance, and joins the other branch not far from the confluence of the united stream with the Peneus.

The principal causes of the obscurity into which the Homeric Scamander (or river of Bunárbashi) had fallen at the time of Demetrius, are sufficiently manifest. When Troy stood at Bunárbashi, it was natural that the river which had its sources under the walls should be one of the _deified_ rivers of the district. In the climate of Greece a perennial fountain, however small, was held in at least equal honour with a large torrent affording only water that was either turbid or stagnant: and we find many proofs in ancient history, and upon ancient monuments, especially coins, of the importance often attached to streams, however diminutive, which flow near the sites of large cities. It is not surprising, therefore, that the river, which from the position of its sources and from its utility was more peculiarly the river of Troy, should, while Troy flourished, have had a preference over the broad torrent in giving name to the united stream; or that its local importance should have ceased when the capital of the district was removed to a situation nearer the sea.

But besides these accidental causes, there were others arising from physical changes which tended to destroy the importance of the river of Bunárbashi. The Mendere and its tributary streams, which flow from Aktshi-Kiúi, from the Kamára valley, from Tshiblak and from Ghiumbrek, being all torrents descending from lofty mountains, bring down with them a great quantity of stones, earth, and other matter: while the Bunárbashi stream, deriving all its water from pure deep-seated veins, has little or no deposit. Hence during the ages which have elapsed since the Trojan war, the eastern side of the plain has been gradually rising; the course of the Mendere has been gradually receding from that side[464], and the western side has become more and more marshy; until at length the Bunárbashi, instead of uniting with the Mendere about the middle of the plain, as in the time of the Trojan war, is now forced to find its way through the marshes on the western side, and from those marshes into the Mendere by two exits not far from Kum-Kale, or towards the ancient Sigeium. Its waters in the plain have been still further diminished by a canal, which carries off a large portion of them into another stream, which joins not the Hellespont, but the Ægæan, at a part of the coast situated not less than seven miles from the ancient mouth of the Scamander. Whether this canal is the remains of an ancient work made for the purpose of draining the plain, when it became marshy by the operation of the causes above stated, or whether it was formed by the Turks merely for its present use, of turning some mills, may be doubtful: its effect has been to cut off in summer all communication between the Bunárbashi springs and the marshy ground on the western side of the plain; so that it is only in rainy seasons that the old bed of the river, which is still very traceable, is now filled with water. I shall here take occasion to remark, that the manner in which the alluvion collects in this plain, as already described, will account for an apparent difficulty in regard to those passages of the Ilias which shew that the Scamander (the united stream) flowed on the left of the Grecian encampment, or toward Rhœteium[465], instead of towards Sigeium, as might be inferred from Strabo[466] and present appearances: for it is evident from the causes mentioned, that the altered course of the river would be to the westward of the former course; and consequently that when there was a bay at the mouth of the Scamander, the river probably issued into that bay, not towards its western, but towards its eastern side[467]. No appearance of a bay indeed is now visible; but its former existence is undoubted, as well from the testimony of Homer as from the physical structure of the land. Instead of two promontories with a beach between them, as described by the poet, there is now only one low point of land, which has been formed between the two ancient capes by the soil brought down from the upper country by the river, and deposited at its mouth in the course of ages. The rate at which the new land has accumulated may be inferred from Strabo and Pliny, from whom it appears that in their time New Ilium was distant about a Roman mile and a half from the nearest shore[468]. Now it appears from the existing vestiges of New Ilium, by those of its citadel on the summit of the hill of Paleó Kastro, which rises behind Kalafatli, and Kum Kiui, and by other remains on the western slope of that hill,—that the lower part of the town reached nearly to the position of Kum Kiui, which is three miles from the shore, or more than double the distance assigned by Strabo. Allowing therefore the same rate of accumulation between the Trojan war and the Augustan age, as since that period, it becomes probable that in the former age the sea reached to about half a mile below the position of Kum Kiui: and consequently that Hestiæa of Alexandria was nearly correct in supposing that all the plain below the hill of New Ilium had been gained from the sea since the time of the Trojan war[469],—the sandy ground at the extremity of the slope of that hill, which gives name to Kum Kiui (Sand-village), marks perhaps what was at one period the sea beach. To those who may think this formation of new land over-rated[470], it is to be observed, that in every instance in which the history of Greece has left us the means of comparison, the same phenomenon has occurred in the maritime plains; and that in the instances of the Spercheius and Mæander, but particularly of the latter, the soil has been formed in the same period of time with a much greater rapidity.

From all these considerations, therefore, it seems highly probable that the mouth of the Scamander in the time of the Trojan war was not far from the situation now occupied by the village of Kum Kiui, and that the river of Bunárbashi or Scamander, instead of then creeping along the foot of the southern and western heights, crossed the plain from near Erkessi in the direction of Kum Kiui, and that it joined the Mendere or Simoeis towards the middle of the plain, perhaps not far from the present village of Kalafatli. The passages of the Ilias in which the πόρος, or ford of the Scamander is mentioned, tend to show that such must have been the course of the river, if Troy stood at Bunárbashi; and we have seen that the nature of the plain, and the manner in which the alluvion has been accumulated, render such a state of the river in ancient times highly probable.

A third objection to Bunárbashi as the site of Troy is, that its distance from the Grecian station at the mouth of the Scamander is so great as to render impossible some of the events of the Ilias. In considering this distance, however, we must first deduct from the actual distance of Bunárbashi from the nearest shore, the new land formed since the Trojan war, together with the _depth_ of the Grecian encampment, which in _length_ extended from the foot of the hill of Achilleium on the right, to the mouth of the Scamander on the left. The new land we have already seen to have been nearly all that which now lies below Kum Kiui. The following are the only circumstances upon which we may build a judgement as to the extent of the Grecian encampment.

According to the poet, the bay was too narrow to contain the whole fleet, which was therefore arranged in several lines[471]. Although nothing but necessity could have made the Greeks submit to having any of their vessels at a distance from the sea, and that we may therefore suppose the number of lines to have been as few as possible, the poet’s expression will hardly allow the supposition that there were fewer than four or five lines. And this number agrees very well with the dimensions of the ground: for if we allow 25 feet for the breadth of each ship, added to the interval between it and the next, we shall find that about one-fifth of 1200, which is the amount of Homer’s enumeration[472], would have been sufficient to occupy the space of one mile and a quarter, to which the rear of the Greek encampment was confined by the hill of Achilleium on the right, and by the river on the left, supposing its mouth to have been near Kum Kiui[473].

For the breadth or depth of the encampment it would not be necessary to assign more than three or four hundred yards, if it were measured only by the length of the ships, added to the necessary interval between the rows: but it is obvious that a large space must, either in the length or depth of the encampment, have been required for the tents of the leaders, for the chariots and horses, for the market, and for the places to contain the cattle and other commodities which the Greeks collected for provisions, or to be exchanged for wine[474]. In short, for a permanent encampment of between 50,000 and 100,000 men[475], with a front of a mile and a quarter, a depth of not less than half a mile would be necessary. Such a space would not be greater than was required by the Romans for their encampments[476]; in which, although there was ample accommodation for the several departments of the army, there was no necessity for the space required in the camp before Troy, for the ships, and for some of the other incumbrances incidental to its permanence. On the one hand we can hardly restrict the Greek camp to a smaller space than I have mentioned, because it would have been insufficient to contain the ships and tents: on the other, a much larger can hardly be assigned; because the inconvenience of having any of the ships at a distance from the sea-shore would be a powerful motive for contracting the space towards the plain, and because the poet expressly states that the army was crowded[477].

In considering, therefore, the transactions of the Ilias, the present distance of Bunárbashi from the mouth of the Scamander must be diminished about three miles and a half, in order to give the distance between Troy and the Grecian rampart, which will thus be reduced to about six miles.

The events which have been considered most inconsistent with the distance of Bunárbashi from the Hellespont, are those occurring on the days called by Pope the 23d and 28th; the former day occupies the 2d, 3d, 4th, 5th, 6th, and the greater part of the 7th books of the poem; the 28th day extends from the beginning of the 11th to the middle of the 18th book.

On the 23d day the Greeks are drawn out, after their forenoon’s repast, in the plain lying between the rampart and the Scamander; and from thence they advance to the city, where, after the duel between Menelaus and Paris, the armies join battle with alternate success. At one time the Trojans have so far prevailed as to have approached the Greek camp[478]; and at another, the Greeks are again near the city[479]. Hector then rallies his army; a duel ensues between him and Ajax, which is put an end to by the approach of night[480], and the Greeks retire to their encampment. It does not seem necessary to suppose that the ground passed over by the Greeks on this day is more than 20 or 22 miles; six of which were performed after the close of day.

On the 28th day the two armies drawn out in the plain before the Greek encampment, fought only with the light troops until the hour of the woodman’s meal[481], which, to judge by modern customs, was about 9 or 10 o’clock in the forenoon. The charioteers of the two armies having then come to action, the Greeks had the superiority, and beat back the Trojans quite to the walls of Troy[482], where Agamemnon being wounded, Hector in turn leads the victorious Trojans to the Grecian rampart, forces it, and fights at the ships. Patroclus then advances to battle in the armour of Achilles, and drives the Trojans back to the city. Here he is slain, and the Trojans again advance near to the Greek camp before the day closes[483]. As the movements on this day carry the parties quite up to the hostile fortifications, the distance passed over is in so much, but no more, greater than on the 23d day; and 24 miles seems to be the utmost distance that we are obliged to suppose the Greeks to have passed over on this day.

In considering the probability of these exploits, we must take into consideration that whatever may have been the proportion of the infantry to the chariots, the extreme distances appear to have been performed only by the latter; for Homer, in all the great movements from the Greek camp to Troy, and from Troy to the Greek camp, as well as in all the principal actions, notices the chariots only. Even in the assault of the wall, in the beginning of the 12th book, Hector descends from his chariot; and all the other Trojans, adds the poet, follow his example.

Not much argument, however, seems necessary against objections which, when allowed in their fullest force, are founded only on the exaggerations of a poet, to whom, however accurate as a geographer and historian when it was his object to be so, we cannot refuse the usual poetical liberties in some of the most animated descriptions which his work contains. If the labours of the Trojan and Grecian heroes in the two days the events of which are thought to disprove the position of Troy at Bunárbashi, were too great for ordinary men; they were not beyond the power of heroes who could hurl such rocks as two men in the time of the poet were unable even to lift[484]; who could make their voices heard from the centre to either extremity[485], or even from the one end to the other[486] of an encampment of sixty or eighty thousand men; and who could see so clearly, that Helen is able from the walls of Troy to point out and minutely describe all the leaders of the Grecian host, when the whole Trojan army lay between[487]. It is evident that these are fictions which the Muse allows and encourages; and instances of them are so frequent throughout the poem, that it cannot be necessary to make any more particular reference to them. At one time the poet found it convenient to magnify beyond probability, or even beyond possibility, the common occurrences of war; at another, to bring together the actions of an extensive field, in order to present them to view in one continued scene.

A fourth objection which has been made against the site of Bunárbashi is, that in this position it would have been impossible for Achilles to have pursued Hector three times round the walls of Troy, as Homer relates. But does Homer really so relate? It cannot be denied that many interpreters, ancient and modern, have understood the poet in this sense; and it is perhaps the most obvious meaning to a cursory reader, who does not particularly consider the fact described, or who has not, by a view of the site of Troy, been convinced of its extreme improbability. Virgil, however, who in the latter part of the 12th book of the Æneis, has very closely imitated every part of Homer’s description of the encounter between Achilles and Hector, seems to have understood his prototype very differently. He does not represent Turnus as pursued by his adversary _round_ the walls of Laurentum, but as forming a circle in a plain which was bounded by those walls, by a marsh, and by the Trojan army. In like manner the pursuit of Hector by Achilles occurred in sight of the Trojans, collected on the ramparts on one side, and of the Grecian army drawn out in the plain on the other. And the poet, in describing the action, mentions no objects passed by Hector and Achilles, except the Scæan or Dardanian gate, the carriage-way under the walls, the Erineus, and the source of the Scamander[488]; all places which we know to have been on the side of the city towards the plain. Can it be supposed that Homer intended to describe the heroes as following such a track as must have concealed them entirely from the view of both armies, except in a small portion of the circle?

It has justly been observed by Lechevalier and Choiseul Gouffier that the word περὶ, which has given rise to the erroneous interpretation of this passage, means, in other passages[489] perfectly similar, _near_ or _before_ the city, and not _around_ it. To this I may add, that no supposed situation of the city, which is not entirely in the plain, will suit the idea of a course round the entire circuit of the walls; and that such a situation would be totally unadapted to the description which Homer has given of Troy, as windy[490], lofty[491], and as surmounted with a citadel bordered by precipices[492]. Strabo in fact, following Demetrius, makes use of this very argument to prove that the ancient city did not stand at New Ilium; round which, he remarks, it would have been impossible for Achilles to have pursued Hector[493]. It would seem, therefore, that the poet, as a keen observer of nature, intended to describe that circular course, which a person invariably takes when he runs from another, and finds no shelter or advantageous position for defending himself. The track of the two heroes was from the Scæan gate, along the road under the walls, by the Erineus, and by the fountains of the Scamander back again to the Scæan gate. Ὣς οἳ τρὶς Πριάμοιο πόλιν περὶ δινηθήτην[494].

* * * * *

It remains to offer a few remarks in justification of the north-eastern portion of the map which accompanies the present volume. This part of Asia Minor was called Pontus by the Romans, from its bordering on the Euxine _sea_: though it still retained the divisions of its ancient inhabitants, the Bithyni, Maryandini, Caucones, and Paphlagones. Here, as in many other parts of the peninsula, modern travellers have not yet afforded us sufficient information to enable us to make the best use of the evidence of ancient history. The astronomical observations of M. Beauchamp and Capt. Gauttier have been of great importance in giving the correct length of the coast, its general outline, and the exact position of the principal places: but it requires such a careful survey as that of the southern coast by Capt. Beaufort, to illustrate fully the three ancient Periplus of the Pontic coast[495], and to correct the numerical errors which their disagreement with one another proves to exist in them.

On the sea-coast all the most important sites of antiquity are determined by the actual names.—These sites are _Rhebas_, now Ríva; _Calpe_, Kerpe; the river _Sangarius_, Sakaría; _Heraclia_, Erégri; the river _Parthenius_, Bartan, in Greek Parthéni; _Amastris_, Amásera; _Cytorus_, Kídros; _Thymena_, Temena; _Carambis_, Kerempe; _Abonuteichus_, afterward _Ionopolis_, Aináboli; _Cinolis_, Kinóli; _Stephane_, Istefán, in Greek Stéfanos; _Sinope_, Sinub, in Greek Sinópi; _Carusa_, Kerze; _Amisus_, Samsun. With these data it will not be difficult for the future traveller to fix the intermediate names of the three Periplus: especially as existing vestiges of antiquity, and the rivers which form a large proportion of the ancient names, will greatly facilitate the task.

Although the route along this coast, in the Peutinger Table, is unworthy of much notice, and conveys very little information, it is right to point out the obvious correction of one remarkable error which it contains. The author, misled by the similarity of the name of Amastris (written Mastrum in the Table) with that of Amasia, has substituted the coast-road from Amastris to Sinope for that leading from Amasia to Sinope. Of this the names along the latter route in the Table, although disfigured, leave no doubt.—Cromen, Cythero, Egilan, Carambas, Stefano, Syrtas, are obviously intended for Cromna, Cytorum, Ægiali, Carambis, Stephane, Syrias; and the sum of the distances, 149 M. P., is tolerably correct. It is probable, therefore, that the two roads should change places in the Table; although it must be confessed that no proof of this inference is to be found in the road of the Table from Mastrum to Sinope; for the sum of the distances of the three places on that route is not above half the real road-distance, and I can find no traces of their names (Tycæ, Cereæ, and Miletus) in any other ancient author.

Another and a more important defect in the routes of the Table through Paphlagonia, is the omission of the name of the place which by its _two towers_ is shown to have been the most remarkable on the road leading from Nicomedia to Gangra, with a branch to Amasia. As this route of the Table lies between the coast road and that leading from Nicæa to Amasia by Juliopolis, Ancyra, and Tavium, it seems evidently to have been the same as the modern road from Nicomedia to Amasia by Boli; for the structure of the country, and the direction of its mountains, passes, valleys, and rivers, must naturally have led the main ancient road in the same direction as the modern. The position in the Table of the place with two towers without a name, relatively to the two ends of the route, shows that it stood on or about the site now occupied by Boli. Now Boli is evidently an abbreviation of some name ending in Polis, which in process of time was vulgarly used in that form, like ἡ πόλις for Constantinople. In Honorias, which under Constantine formed a district separate from Paphlagonia proper, lying between it and Bithynia, there were three places with the termination of polis—Claudiopolis, or Bithynium; Flaviopolis, or Cratia; and Hadrianopolis[496]. The other towns of Honorias were Tium, Heraclia Pontica, and Prusias on the Hypius; so that the district seems to have chiefly comprehended the country lying between the Sangarius and the Billæus. Bithynium or Claudiopolis was on the Sangarius[497]; and having been originally a colony from Greece[498], was probably not far from the mouth of that river, Greek colonies having generally settled in maritime situations, as we see instanced in several cities on this coast. Flaviopolis was twenty or thirty miles from Claudiopolis, on the road leading from that place to Ancyra[499]; consequently to the westward of Boli. Boli, therefore, seems to have been the ancient Hadrianopolis. It is singular that among the numerous inscriptions which so many travellers agree in having observed near Boli, not one should yet have been copied, containing the name of the ancient city.

The other places on this road in the Table have been inserted in the Map, in the situations which I have thought the most probable, trusting less to the distances in the Table, (which are probably not more correct in detail than they are in the general result,) than to the situation of the valleys and fertile districts. Potamia, a place which Strabo has noticed as being in this part of the country[500], seems to have stood in the valley of Beinder, where the branches of the Parthenius first unite into a considerable stream.

On another route in the Table, which crosses the preceding nearly at right angles, the only place named between Gangra and Sinope is Pompeiopolis. This place seems to have occupied the site of Tash Kiupri, as well from the position of that modern town, as from the considerable remains of antiquity found there, and which are apparently of the date when Pompeiopolis may be supposed to have flourished.

Of Germanicopolis, or Germanopolis, we know only that it was one of the principal places of the interior of Paphlagonia, and that it continued to be so in the sixth century[501]. It has probably left some remains similar to those of Pompeiopolis, though they have not yet been discovered by modern travellers. D’Anville supposed Germanicopolis to have occupied the site of Kastamúni; but the words in the Novellæ of Justinian seem to place it near Gangra[502].—Kastamúni is the modern corruption of Castamon, which we find mentioned in the Byzantine history[503], and which may have been a more ancient name, although it is not found in Ptolemy, nor in any authority earlier than the 12th century.

The subordinate districts of Paphlagonia and Cappadocia Pontica; namely, Timonitis, Bogdomanis, Zygiani, Marmolitis, Blaene, Domanitis, Cimiatene, Gazelonitis, Saramene, Phamezonitis, Diacopene, Babamonitis,—have been inserted in the map, from the information, as well as it could be understood, of Strabo and Ptolemy; and some of the Turkish names from the still obscurer description of Abubekr Ben Behrem.

It is much to be regretted that no modern traveller has visited Tshorúm, which there is the strongest reason to believe occupies the site of Tavium, the chief fortress of the Trocmi, and a very important point in the ancient itineraries.

Upon comparing the road from Tavium to Cæsareia (Mazaca) in the Table with that in the Antonine itinerary, we find that none of the names agree—that the distance in the Table is nearly double that in the Antonine—and that both of them give an incorrect rate to the Roman mile. It might be supposed, in explanation of this difficulty, that there were two roads from Tavium to Cæsareia; but I am inclined to think there is some error here in the Antonine, as it places Soanda on this road, which we have good authority for believing to have been in a very different situation, namely, on the great western road from Cæsareia, between that city and Garsabora[504].

ADDITIONAL NOTES.

I have reserved to this place all observations on the geographical information contained in the Latin historians of the 12th century, who have described the first crusade[505]; because, upon a careful examination of it, I have not found any thing either to invalidate or materially to confirm that which is deducible from the ancients or from the Byzantines. At the same time there are several passages in the Latin historians which may receive some illustration from the cotemporary Greeks, or from the ancient geographical authorities.

NOTE TO PAGE 9.

The following is the substance, of a short account, by Anna Comnena, of the military operations in Bithynia in the autumn of the year 1096, which proved fatal to so many of the followers of Peter the Hermit. Peter having passed over into Asia, contrary to the advice which the Emperor Alexius gave him to wait for the other crusaders who were then on the way, encamped at Helenopolis, from whence the Normans proceeded to ravage the country around Nicæa; and having successfully defended themselves against a body of Turks, which advanced against them, they carried back their spoil in safety to Helenopolis. In a second expedition they occupied the fort of Xerigordus, but the sultan Kilidj Arslan, having sent one of his officers against them, retook that place, slew many of the Normans, and made many of them prisoners. He then sent two men to raise a report in the camp at Helenopolis, that the Normans had taken possession of Nicæa, and were plundering it; when the other troops, desirous of sharing in the spoil, proceeded in a disorderly manner towards Nicæ: and thus they fell into an ambuscade which the Sultan had stationed in a place called Draco, and were cut to pieces. The number that fell was so great that their bones formed a mountain. Peter then retired to Helenopolis, where he was invested by the Turks: but the Emperor, unwilling that he should be taken, sent his officer Catacalon with some ships to his succour, upon whose arrival the Turks retired, and Peter returned with his surviving followers into Europe.

From the Latins there is great difficulty in extracting any clear account of these events, which may partly be ascribed to the want of a good map, partly to the ignorance of the authors in ancient geography, but chiefly to the circumstance of none of those writers having been personally engaged in Peter’ s imprudent expedition. They agree tolerably well with the Greek Princess in regard to the principal events, but are at variance both with her and with one another as to many of the particulars. They relate that the crusaders, having crossed the Bosphorus, marched to Nicomedia, and from thence to a place on the sea-side called Civitot or Civito, where they were amply supplied with provisions by sea. The French troops, separating from the others, spread themselves over the country and took possession of an abandoned fortress called Exerogorgo (the Xerigordus of Anna Comnena), the situation of which is variously described as four days beyond Nicomedia, as four days beyond Nicæa, and as three or four miles from the latter. Here they were soon surrounded by the Turks, who cut off their supply of water, slew many of them, and at length, by the treachery of one of the French chieftains named Reynald, captured many more. Soon after this event there was a general action in the field, which was fatal to the gallant military commander of Peter’s army, Gauthier Sansavoir, (Walter the moneyless,) as well as to several other distinguished leaders. The exact scene of action it is very difficult to understand, though it rather appears from a comparison of Anna Comnena with Albert of Aix-la-Chapelle, and William of Tyre, the two Latin authors who have given the fullest account of these transactions, to have been at the northern extremity of the plain of Nicæa, and on the adjoining hills. The chief slaughter of the Franks seems to have occurred in the passes leading from thence to the sea, of which passes the Turks had made themselves masters during the action, unknown to the enemy. According to the Latin historians, a part of their army found its way back to Civitot, where they were speedily surrounded by the Turks, and where they would have been in great danger of being all slain or taken, had not the Turks been induced, by the mediation of Alexis, to retire, and to leave the crusaders at liberty to return to Constantinople.