This fortress is a mountain peak somewhat resembling a castle, and to which they are in the habit of driving their flocks for safety in times of danger. The monument has certainly no architectural merits, but it is constructed in a solid and careful manner of finely-cut stone.

It is rectangular in plan; on each side are six Corinthian pilasters, and four on each end, surmounted by a bold and massive entablature. Within, it is divided into two chambers, each 9 ft. 9 in. broad. The inner or mortuary one is 6 ft. 9 in. long. Several columbaria exist in the walls. It had two doors; the outer one has disappeared, but it must have been of a single slab of stone turning on pivots, the holes for which still exist. Another door in the partition wall led into an outer chamber 10 ft. long, with seats in the wall, and lighted by a window. This door is still lying on the floor; it was decorated with a bas-relief representing a winged figure holding some large object in his right hand. Above the aperture of the door half of a wreath of laurel is sculptured on the wall. The roof was of immense blocks of stone laid across, one of which still exists in each chamber. This is evidently the building erroneously laid down in Ste. Marie’s map as ‘Mausolée de Verrius,’ but the tomb of C. Verrius Rogatus, described by Sir Grenville Temple,[173] is situated on the opposite side of the city, near the aqueduct.

The ruins of Roman constructions which we had observed all along to-day’s route became more frequent as we approached Mukther. Instead of encamping within the circle of the ruins there, we proceeded a little further on, and pitched our tent in a charming dell, full of clear springs and rich grass, and shaded by a grove of fig-trees, nearly south-west of the koubba of Sidi Ali ben Omar.

Here, again, the old scene of wrangling took place before we could get any supplies. The Khalifa of the district told us that the people were completely poverty-stricken, they had not wherewithal to nourish a single horse or mule. Sheep were as much a matter of ancient history as the Roman cities (the hills were covered with both!), and that if we insisted on camping there, we must supply our own provisions, and our animals must be satisfied with the grass which grew on the ground.

I took the head man apart and putting a sum of money in his hand, told him that we had no desire to be a charge on any one; all we asked was permission to pay a fair price for what provisions we might require, and that I would give him whatever more might be due before leaving. His objections vanished in a moment, and we were in hopes that at last we had found a means of purchasing honestly what we required. In an hour, however, he returned with the money in his hand, saying that it would be a disgrace to our Lord the Bey, if a Consul were permitted to travel through his country save as a guest, that everything we required should be supplied, but that payment was out of the question. Our spahis had got hold of him in the interim, and, as usual, they managed to have their own way, which was that we should rather expend our money in presents than in payments, and for very obvious reasons.

The grove where we were encamped was full of pigeons and other birds. My companion, who could never see a pigeon or game-bird without an irresistible longing to bag it, was on the point of loading his gun, when the Arab implored us not to shoot them, as they were in the habit of frequenting the Saint’s tomb, and were considered as partaking of his sanctity. I need hardly add that we were delighted to respect this pleasant superstition.

CHAPTER XXIV.

This city has been identified by nearly all the older travellers as Tucca Terebinthina. Bruce, amongst the number, appears to have entertained no doubt on the subject. It was M. Guérin[174] who first pointed out that the similarity of the modern and ancient names, coupled with their recorded distances from known places, left no doubt that Mukther was the Mactar or Oppidum Mactaritanus in the list of African bishoprics,[175] and that Tucca Terebinthina was to be sought for at the modern Dougga, about eleven miles to the south-west, where considerable ruins exist, but which must not be confounded with the more important city of the same name, near Teboursouk.

The position of Mukther is admirably chosen; it is built on a wide and elevated plateau between two water-courses, the Oued Sabon, or river of soap, on the north, and the Oued Mihran on the south. The banks of the former are high and precipitous, and serve as a natural defence on the north-west side of the town. This has been further fortified by a wall constructed of immense blocks of stone, placed loosely and irregularly together, resembling more the retaining wall of a terrace than a regular line of defence.

We commenced our explorations from this side of the ravine, opposite to the lower triumphal arch. On the north-west face of it was evidently the necropolis of the city. The hill-side is covered with tombs. Many contain simply records of the names and ages of the deceased, others have rudely sculptured figures. One tomb was a carefully constructed vault of cut stone, near to which was a handsome cippus, which probably was erected over it before its violation. One only struck me as sufficiently curious to deserve copying. It was evidently a family tomb, containing four inscriptions, the first of which was the rudest and the most injured by time; but what rendered it curious was the existence of two hieroglyphical figures sculptured on the stone in relief before the inscription was engraved.

On the opposite side of the ravine, and rising directly from its right bank, is a large triumphal arch, which forms the first of Bruce’s illustrations. Of this he has left us eight sheets.

1. A rough pencil outline, which has not pleased him, and which he has subsequently cancelled by a waved pencil line of obliteration; on the back are numerous architectural details and memoranda of measurement.

2. Another sheet of similar details and measurements.

3. A beautifully executed perspective view, done on the spot, showing the actual condition of the monument, without any foreground or other accessories. (Plate XVII.)

4. A highly finished Indian-ink drawing, from the same point of view, decorated (!) with an impossible landscape, probably by Balugani; instead of the gentle slope of the hill on the other side of the ravine, bare of trees and destitute of water, there is a foaming cataract on one side, and an extensive vista of river, wood, plain and mountain on the other.

5. A highly finished plan and elevation.

6. A highly finished drawing in Indian ink of the Corinthian order of arch.

7. A highly finished drawing, representing architectural details of soffits, cornice, and architrave. (Plate XIX.)

8. A highly finished drawing of soffits of architrave of order.

This arch is of a richer architectural composition than that of Trajan, which will presently be described. It has but a single opening, without archivolt, flanked by two square niches, with rounded heads, underneath the level of the impost. Four Corinthian columns corresponding to pilasters enrich both façades. These existed nearly entire in Bruce’s time, and even M. Guérin mentions ‘un petit vestibule soutenu sur deux colonnes corinthiennes,’[176] but these have now entirely disappeared, although the stylobates are still in place. The whole entablature seems as though it had been thrown down since Bruce’s drawing was made, and the stones all piled up again without much regard to order. It is more likely, however, that this may be only a general dislocation of the building, caused by an earthquake. Fragments of cornice highly decorated may be seen here and there in the crumbling mass, and the very rich treatment of the impost, which is in a tolerably good state of preservation, would attest the magnificence of the building, even if Bruce’s exhaustive designs did not exist. To show the present condition of this monument, a photograph taken by my companion, the Earl of Kingston, is also given (Plate XVIII.)

The two façades are identical; on the ravine side the foundations are entirely exposed to view, and consist of a mass of rubble masonry, twelve or thirteen feet in height, between which and the floor of the arch are three courses of substantial cut stones, but there is no appearance of this having been connected with any other structure.

It is difficult to understand that a building of this nature could have been constructed at the edge of a ravine without there being an arch of some kind to span the latter. Mr. Davis says that he observed part of a paved road on the opposite side of the ravine, from which he infers that a bridge must have existed;[177] but after the most careful search, with his work in my hands, I was unable to trace any vestige of this. Probably therefore the ravine, if it existed at all in ancient times, has been much deepened by the action of water during the last sixteen hundred winters.

M. Guérin mentions that he discovered amongst the débris, at the base of the monument, ‘a sculptured Greek cross, which would assign to this monument a date posterior to the Christian era.’[178]

I recognised this stone without difficulty, and though it is much obliterated by time, I feel confident that it is actually the detail of soffit, figured by Bruce in Sheet 7, above mentioned. This is shown on Plate XIX.; the lowest figure, with all the details of the acanthus leaves worn off, might easily be taken to represent a Greek cross.

Unfortunately, no trace of inscription remains, although a search amongst the débris, which encumbers the base of the inner façade, might be rewarded by success.

The dimensions of this arch given on the plan and elevation are as follows:—

Close to the arch is a group of olive-trees, and a delicious spring rises from the ground a few yards further off. During the whole of our ride yesterday we noticed at intervals rocks consisting of a mass of petrified oyster-shells; here also they recur, and in the spring I have just mentioned are two frusta of columns made of this material, which has an excellent effect.

In the ravine below the arch we observed a Lybian tombstone, containing the representation of a figure with the face mutilated, holding objects like ears of corn in his hand, and three lines of inscription below.

From this arch a paved road led towards the south-west, where it conducted to a large building, on the site of which the tomb of one of the members of the family of Sidi Ali bin Amer has been built. Numbers of monolithic columns of limestone lie scattered about, and one still remains in position. Bruce, in his notes, thus alludes to this building:—

Here I identified a small fragment of the inscription recorded by Sir Grenville Temple,[179] from which he concludes that either this temple or the neighbouring amphitheatre was built in the reign of Trajan. The inscription as it now exists is . . . AIANI PAR . . . Sir Grenville supplies two other letters, TH. As the name of Trajan here appears in the genitive case, it is more probable that it was built in the time of Hadrian, and that the inscription would have been, TRAIANI PARTHICI FILIVS.

Continuing along this road, on the left hand, is a small amphitheatre constructed of rubble masonry, faced, no doubt, at one time with cut stones, to judge from the number lying about. It is of elliptical form, the major axis being about 160 feet, and the minor one about 114.

A little further off, and to the right, are the remains of a building which M. Guérin believes to have been a temple of Diana,[180] he having found the name of that goddess on a fragment of marble, which had probably formed the base of a statue. The building is too much ruined to permit any conjecture as to its destination, and though I searched diligently, I failed to find the inscription in question.

The road now leads to the triumphal arch of Trajan, which appears to have stood in the centre of the town facing the south, and which forms the second of Bruce’s illustrations. There are six sheets of designs of this monument.

1. A sketch taken upon the ground apparently, in Indian ink without any landscape. (Plate XX.)

2. A highly finished Indian-ink perspective view with back and foreground of the usual character, a camp of soldiers, sheep, cattle, Arabs, &c.

3. A finished Indian-ink plan and elevation to scale.

4. Finished drawing in Indian-ink, to scale, of details of minor order and its pediment.

5. Finished Indian ink drawing, to scale, of details of major order.

6. Pencil memoranda of details and measurements.

This building in its proportions and treatment is very grand and simple. The solid mass of the front is much higher than it is wide. The principal order is Corinthian, with a single three-quarters attached column near the angle, supported on a lofty pedestal or stylobate; it has the usual base and capital, and the shaft is seven frusta high. The architrave and frieze over the columns consist of one plain block without any mouldings, which is not carried along the face. The cornice is the usual one, handsomely but not elaborately enriched. Three courses of the attic remained in the time of Bruce; portions of two courses only are now in place. In the centre of this block is an archway, having its own peculiar treatment. On each side of the opening is a smaller semi-engaged Corinthian column, raised on a stylobate equal in height to that of the major order; the impost of the arch runs all round the monument, intercepted only by the columns. There is no impost to the arch-head, which consists only of the radiating cunei; and the soffit is quite plain. The capitals of the columns flanking the archway rise somewhat higher than the archway itself, and are surmounted by an architrave, a lofty plain frieze, on which is the inscription, and a simple cornice with a pediment above. In the middle of this there is a square opening, giving access probably to a chamber over the arch. The whole treatment is dignified and reserved.

The inscription on the frieze of the minor order has been given more fully by Bruce than by succeeding travellers, and much more so than can be deciphered at the present day. It runs thus:—

The mention of the twentieth year of the tribunate of Hadrian fixes the date as the last year of his reign, A.D. 117.

The arch has been built up with loose stones, and has evidently been surrounded by a wall, to convert it into a fortified position.

The building has not suffered greatly since Bruce’s time, but it is buried almost to the level of the impost in débris. As Bruce’s sketch shows no sign of the loose masonry wall, it is possible that this may have been constructed by the Arabs; at the same time it is not improbable that he might purposely have omitted it, as having no connection with the history of the building.

The dimensions given on the plan and elevation are:—

To the south-south-west of this is a very large building, which Sir Grenville Temple and M. Guérin believe to have been a palace or a church, but which I am inclined to consider public thermæ. There is a large central hall, twenty paces long by twelve broad, the north and south ends of which were almost entirely occupied by immense arches. On each of the other sides were three arched openings, the central one of which was the largest, giving access to lateral chambers, now entirely destroyed. Near the north-west angle is a large well, more than twelve feet in diameter, whence probably the water supply of the establishment was derived. The construction was very massive, partly of rubble and partly of cut-stone masonry. This building, like Trajan’s arch, had been converted into a fortress by a bastioned wall constructed of old materials, amongst which I observed two Christian tombstones; one had a cross inscribed within a circle, the other the monogram of Christ, with the Alpha and Omega.

About midway between Trajan’s arch and the double-storied mausoleum was a large and important pile of buildings, which might well have been a governor’s palace. The construction is of rubble masonry, but the voussoirs of the arches and some of the facings are of huge blocks of cut stone. It is much encumbered with débris, in which it is buried up to the imposts of the arches.

Beyond this is the mausoleum just alluded to; it consists of a sepulchral chamber containing seventeen carefully cut columbaria. Above this was an open chamber with a portico in front for statuary, the whole being surmounted by an elevated pyramid. Exteriorly the sides are ornamented by Corinthian pilasters. Above the door was an inscription; this has quite disappeared with the exception of the dedicatory letters D.M.S., which are of unusual size. Below this was a sculptured representation, now almost effaced, but from what remains there can be no doubt that the subject was exactly the same as that on the mausoleum of Julius Proculus, the procession of a bull to the sacrificial altar, one probably in common use like the cross on Christian tombs.

The second mausoleum, to which I have just alluded, is on the right bank of the Oued Sabon, near the lower triumphal arch; it was of similar construction to the other, but has only the lower storey remaining. Above the door the sculpture is in a very perfect condition, except that the faces of the persons assisting at the procession have been obliterated by chisels. Above this, occupying the whole length of the front side is a long inscription, part of which has been thrown down and is lying amongst the débris. Both Sir Grenville Temple[181] and M. Pelissier[182] have given portions of this inscription. M. Guérin,[183] however, has given as much of it as it is possible to decipher. It shows that the tomb was erected to commemorate Julius Proculus, who, ever since his assumption of the toga virilis, had devoted himself to the study of elocution, and was much admired by his fellow-citizens.

To the west of Trajan’s arch is an aqueduct of large cut stones. The piers of the arches are five feet six inches in thickness, and the intervals between them eleven feet eight inches.

Although the ground between the ruins is so thickly strewn with loose stones as to render walking a difficult and painful operation, it was during our visit covered with crops of wheat and barley. It is difficult to understand how any plough could be got to work under such circumstances. There was, of course, no attempt to clear away any of the stones; that would be too much to expect from an Arab husbandman. Perhaps from their point of view they are right; land is so abundant and population so scanty, that it pays them better to sow two acres in a superficial manner than to cultivate one with greater care and thus produce heavier crops.

CHAPTER XXV.

About six and a half miles from Mukther we reached an elevated rocky pass, under a peak called Djebel Deir Subah. All along our route we had observed, as usual, the remains of Roman buildings even in places where nothing but diss grass now grows. From this point there was a most extensive view. To the south Djebel Trozza bounds the horizon, and the ruins of Mukther are seen in the middle distance; to the north we could plainly see both the city of Kef and the ruins of Zanfour. Descending from this we passed Magherawa, the first stone village we had met since leaving Susa: it is composed of very poor hovels, built on the site of an ancient town, but its position is delightful. Above it is a fine spring, which soon widens out into a little stream; the ground is richly cultivated, and there are numerous fine groves of olive trees. Here it was that Sir Grenville Temple[184] found the Punic inscription, No. 142, which he has given in his work, and fragments of rudely-sculptured bas reliefs of men and animals, but no Roman inscriptions. The mere fact of these people living in permanent stone villages is sufficient to prove them descendants rather of the original Berber races than of their Arab conquerors. Soon the valley widens out into the plain of Sers, by far the richest and most highly cultivated that we have hitherto seen; olive groves are numerous, and stone villages commence to replace the woollen tents of the nomad Arab.

We stopped to rest for a short time at El-Lahs, a small village near a magnificent wood of olive trees. A little before entering it we passed a fine spring of water, which issues from a cavity of the rock; a number of Arab girls were washing their clothes in it, and did not appear particularly averse from seeing or being seen by us; but as soon as our escort came in sight, veils were brought into use, and the youngest of them scampered away and hid their faces till we had gone out of sight. The appearance of Europeans must be rather a startling event, to be talked of for years afterwards, and to serve, no doubt, as an epoch in their simple chronology. I only trust that they did not jump at the conclusion that all Europeans resembled us; they might well style us, as the Chinese did, ‘red devils.’ Our faces and hands were scorched and excoriated by constant exposure to the sun; and our costume, though very convenient for travelling in Tunis, was rather travel-stained and would have been thought peculiar in more civilised lands. Modesty prevents me from making any allusion to my own personal appearance, but I am bound to admit, that my companion looked about as disreputable a character as it would be possible to meet even in the wilds of his native Ireland.

We did not enter the village, but sat down under the shade of the olive trees on a grassy mound outside, and regaled ourselves with a bowl of fresh warm milk. Generally it is difficult to obtain this, save early in the morning, or late in the evening; the milk is kept in vessels, which are never thoroughly cleansed and which turn it almost immediately; but the Arabs like it so, and even prefer it the second day when it has quite curdled.

In the vicinity are some curious ruins, not at all Roman in appearance, but we could not decide whether they were anterior or posterior to the Latin epoch. Sir Grenville Temple also noticed them in these terms: ‘The most remarkable curiosities, however, are several very ancient constructions scattered about in the fields round the villages, formed of large unwrought stones or slabs, some measuring 17 ft. 10 in. in length by 6 ft. 2 in. in breadth, and 1 ft. 8 in. in thickness. With these a number of little chambers are formed, generally in two rows, divided by a passage resembling in their plan some of the tombs of the Pharaohs at Biban el-Melook, near Thebes; that is to say, those which have little chambers on each side of the entrance gallery. The whole edifice is also roofed with similar slabs laid flat. I should feel inclined to attribute their construction to the aboriginal inhabitants before the arrival of the Phœnicians; probably, they are not tombs but magalia or houses. Many of them are in such good preservation as still to be inhabited and used as stables.’[185]

After passing this place, and crossing the Oued Ez-Zerga, and the Oued Zanfour, we came to the Henchir, which now bears the latter name.

Bruce was the first of modern travellers to recognise that Zanfour, and not Kisser, as Shaw supposed, was the ancient Assuras; he read the name on the triumphal arch, the inscription of which he has recorded. Pliny mentions the Oppidum Azuritanum as inhabited by Roman citizens;[186] it is also cited by Ptolemy,[187] in the Itinerary of Antoninus,[188] and in the tables of Peutinger as Assuræ. Several bishops of this place are recorded, and one of the epistles of St. Cyprian was addressed ad episcopum et plebem Assuritanorum.[189]

The most important of the ruins is the triumphal arch above mentioned. Bruce has left two drawings of this monument, one a very elegant sketch in Indian ink, evidently his own work, showing its actual condition, and exhibiting in the background the accurate position of the temple on the left, and of the theatre on the right. The other is a highly-finished Indian-ink drawing of the same subject, in which the hand of Luigi Balugani is very apparent. The entablature and one of the columns have been restored, and a camp of Arab soldiers has been pitched among trees behind it; numerous figures in fanciful attitudes are scattered about, and a cart, which must have had some difficulty in crossing the stream, contains a family party in the foreground. This I have chosen for reproduction, as, in spite of these embellishments, the drawing is more accurate, and the Corinthian temple to the left is given in greater detail. (Plate XXI.)

There are also five other sheets to illustrate this monument. One contains a slight pencil sketch, on which the dimensions are given; another is a finished Indian-ink drawing to scale of the plan and elevation: the other three are beautifully executed plates of architectural details of various parts of the monument.

This arch has but one opening, decorated with a Corinthian order, composed of single isolated fluted columns near each angle of the mass, with unfluted pilasters behind, corresponding to them. It is in a very ruinous condition, but enough remains to show, even had these drawings not existed, that the entablature was rich and complete, and crowned by an attic with cornice and basis. As is usual in such buildings, the impost turns all round the monument, without however interfering with the pilasters and columns, and, which is less common, the arch is decorated with an archivolt.

The dimensions given in the plan and elevation are as follows:—

The columns have all fallen down, and lie in fragments around. Bruce’s sketches show that during his visit the right one on the north-east side was complete, as was half of the corresponding one on the opposite side of the opening, but the condition of the south-west side of the arch is not indicated.

On that portion of the attic above the north-east face was a long inscription, part of which still remains in situ. Some of the stones, however, on which it was engraved have been thrown down and broken; fragments of these still exist amongst the débris, but the most part have disappeared.

The following is the inscription, as recorded by Bruce:—

It was also copied by Sir Grenville Temple,[190] on whose authority, supplemented by his own observation, it has been produced by Guérin.[191]

The most important words of this inscription, COL. IVL. ASSVRAS, were rendered by Sir Grenville Temple as OLIVI . AVSVRA.[192] And Pelissier, accepting this rendering as accurate, imagined the letters OLIVI to be an abbreviation of olivifera.[193] M. Guérin has settled, beyond the possibility of doubt, that the words really are