The site of this ancient city is a very fortunate one. It commands the western entrance to the basin of Ambracia, and is defended on three sides by a navigable river. The ruins themselves are not of a very early date; the plan of the Acropolis is regular, and proves that, at the time of its erection, the science of military architecture was far advanced. In the surrounding wall, rectangular towers occur at regular intervals, and built in horizontal courses of masonry. The structure of the wall itself is beautifully symmetrical, and partakes almost of the precision and compactness of mosaic.

This city is rendered very interesting by the excellent preservation of its remains. In a few weeks it might be made ready for the reception of a colony from Corinth or Ambracia, and put in a condition to stand a siege. The restorations of its walls, both of Roman and Gothic times, prove that it was considered of importance in both. In the centre of the citadel are the ruins of a monastic church, containing some frescoes which might have furnished materials to Dante for sketches in his Inferno. At the southern angle of the Acropolis is a pleasing view of the river winding at

the foot of the city, and of the broad expanse of the Ambracian Gulf in the distance.

Proceeding among the woods on the right or western bank of the river, we bear to the north-west, among the low hills which lead into the plain of Lelovo. At a few miles to the north-east of that village is the woody, steep, and conical hill of KASTRÍ, on which are the ruins of an ancient Greek citadel. Beneath the hill, on the east, is seen a beautiful valley, with a lake. The remains of Kastrí appear to be of the same date as those before described of Rogús; but there is this difference in the character of the two places, that the former seems to have been built almost exclusively for the purposes of attack and defence, and not for habitation; while the latter was well adapted for both. The one was a citadel, the other a city.

But, in the mean time, who were the inhabitants of these two places, which we have just visited?-by what achievements were they distinguished? -to what nations did they belong?-what were the fortunes, what were even the names of the cities themselves? The evidence which is afforded for replying to these questions is very scanty, and can only conduct to a conjectural result. Philip, the son of Demetrius King of Macedon, was induced by the urgent entreaties of the Epirots to besiege AMBRACUS, which was favourably situated for furthering the designs of an enemy desirous, as the Epirots were, of hovering over the territory of Ambracia. Ambracus is described by the historian Polybius, in his narrative of Philip's campaign, as situated among marshes, and having only one approach from the Ambracian country by a narrow artificial causeway through the morass. Supposing, as seems most natural, that Ambracus was on the frontier line between Epirus and Ambracia, this marsh must have lain between Ambracus, and the plain of Ambracia, that is, on the south-east of the former. It was of no great extent, but was well fortified by a wall and towers. After a siege Philip took it, and delivered it to the Epirots. He then marched

242

CHARADRA IDENTIFIED WITH ROGUS,

rapidly by a city called CHARADRA, aiming at the narrowest part, that is, the mouth of the Ambracian Gulf, which he was eager to cross in his route into Acarnania. It appears from this circumstance, that Charadra was in a direct line between Ambracus and Actium. These details are, we think, sufficient to warrant the surmise, that Ambracus coincided in site with Kastrí. The Epirots, from their situation and from their want of strength, would neither have desired, nor have been able, to maintain a fortress hanging over the Ambracian frontier, had it not been in their own neighbourhood, and therefore on the north-west side of the enemy's country. Again, we hear of no opposition made by the city of Ambracia to the besiegers of Ambracus; which, had Ambracus been on the Gulf of Ambracia, and thus the door of the commerce of that city, would surely have been the case. The site of the lake and morass on the south-east of the hill of Kastrí, confirms the above conclusion.

Allowing its correctness, we have little difficulty in ascertaining the ancient name of the city above described, which stood upon the site of Rogús. Philip, after the siege of Ambracus, hastened towards Actium, and passed Charadra in his way, which was therefore in a direct line between these two points. The only site which satisfies this condition, and where there are any remains

AND AMBRACUS WITH KASTRI.

243

of an Hellenic age, is Rogús. The name, too, of Charadra, supposes a river in its vicinity, which is there the case. The term Charadrus, by which the river of Charadra was known, indicates a soil broken into ravines and gullies; and, in a similar manner, the word Rogús seems to be derived from the appellation given to abrupt chasms and gorges in the ancient language of Greece.

The lake of Xero-Limne bears some resemblance to that of Ulleswater in Westmoreland. It lies among high limestone rocks, which are covered by many varieties of dark-green shrubs and trees hanging over the water, and deepening the shade cast on it by its steep banks. Here and there some water-flowers bloom upon the surface, and throw a little light upon its gloomy colour. The only sounds near it are those of the wild-fowl startled from its weeds by the footsteps of the traveller. The road lies on its eastern brink; it is skirted by a forest of oak, beech and maple, which thickens on both sides as we proceed. The river is heard dashing along its rocky bed, at no great distance to our right, but is not seen from the route till we cross a path which passes over it by the bridge of the "Pasha's Lady." The views here are magnificent, such as Salvator Rosa would have exulted in. The river tosses itself in cascades; shattered plane-trees torn up by its violence are lying over the stream; along their trunks some speckled goats may be seen climbing, while on the other side of the water the goatherd appears with his scarlet cap shining through the trees. We continue our track on the right bank of the stream. The traveller who diverges from his course here will probably be driven back to it by the wolves, which are not uncommonly met with in this solitary wood. Ascending to the right over some grand castellated rocks of grey limestone, we enter a more open country, but there is no appearance in it of living creature or human habitation. After proceeding a little further to the north, we again approach the river, and enter a small village by its side.

In the summer and autumnal months, CHARADZO,-for that is its name,—is deserted; the inhabitants close their windows, lock up their doors, and quit their houses, which they surrender at this season to the mosquitoes that infest the place, and drive them from their homes. The luxuriant fields of rice, as well as an abundance of marshy plants in the neighbourhood, present infallible indications of the prevalence of malaria, which operates as another cause for their emigration.

The road from Charadzo lies along the wide and pebbly bed of the ancient Charadrus, a name which seems to be connected with that of the village above mentioned it leads through long thickets of thorny paliurus, and

occasionally deviates into swampy fields of Indian corn. The valley contracts itself into a defile, on both sides of which are very lofty rocks: those on the left are clothed to their summit with trees; the shrubs which wave in the wind at the top are scarcely visible, on account of their height; the cultivated patches of the valley are filled with granone, and the soil is everywhere irrigated with limpid streams. An ancient fortress rises on the rocks to the right, to guard the entrance of the gorge; it is called TERAVO: a second, named MESOTITZA, is at two miles' distance from it on the left; in the valley beneath is a water-mill, pleasantly sheltered by trees. The ridge of mountains to the north is called TOMARITZA, that to the northwest OLITZA. Following the valley, we leave on our right the narrow pass of ZAGATORO, where, it is asserted by the peasantry, are ruins of an ancient citadel on the north side of the defile. Similar remains are said to

exist between Mules and Kopáni, at THERAKÈS.

The present route offers a prospect of many geographical and antiquarian discoveries to the traveller; and, even should his success in these respects be below his anticipations, it will more than repay him for his labour by the singular beauty of its natural scenery. After a ride of twelve hours from Charadzo we arrive at the gates of JANINA.