All over the plain, for many miles around, the ruins of Celto-Roman villages have been found, and in many cases partially explored by Dr. Sarmento and others; the objects discovered, like those found in Citania, having been deposited in the museum at Guimarães belonging to the explorer, but in consequence of his death henceforward to be a public institution subsidised by the State. As I shall point out when I describe my visit to the museum, the objects unearthed at Sabroso, St. Iria, and other neighbouring places are immensely more numerous than those from Citania itself; great masses of coins, personal ornaments, arms, inscriptions, and utensils in the museum proving that these places existed far into Roman times, and perhaps much later. The chaotic condition of the Sarmento collection at present, and the apparent absence of any skilled and enthusiastic guardianship, have probably been a reason why certain investigators have attributed to Citania many objects discovered elsewhere, and have founded upon them theories which must necessarily be misleading. Dr. Hübner, who did not see the place personally, aroused the wrath of Dr. Sarmento in this way, and other archæologists have spoken somewhat loosely as to the nature of the finds in the Citania excavations. The great interest of the hill stronghold, indeed, consists in the fact that we have here practically an unspoilt Celtic or Celtiberian town, in which Roman civilisation had but little part. It will be seen by the objects actually unearthed that the place was inhabited after the Roman influence and language had dominated the district, as late, indeed, as the time of Hadrian; but of purely Roman remains, so plentiful elsewhere in the district there are in Citania hardly any; the construction and plan of the houses having much in common with the Irish and Scotch Celtic cashels, and the absence of all indications of Christianity being complete.
Following a well-paved causeway of some seven or eight feet wide, the flat stones of which have been worn smooth by countless generations of forgotten footsteps, we can perceive perfectly the ground plan of the houses on each side. In most cases Dr. Sarmento has excavated down to the stone-laid flooring of the houses inside, and to the base of the masonry outside; and it is possible to wander through the main lanes or streets of the town, crossing each other at right angles here and there, and interspersed by little circular paved open spaces, and to reconstruct in the mind’s eye the primitive life of this city of long ago. Here, for instance, just inside the wall by which we entered is a little square house, some twelve feet wide, containing two rough millstones, of which many have been found. The walls are of huge, rough stones, evidently taken as they came and fitted together with small stones where necessary to fill in interstices, the whole cemented together by some hard rubbly compost. Running past this building and through the town (in one or two cases, indeed, through the houses themselves) is one of the several stone water channels protected by low walls on each side, and supplied in ancient times by the springs that still gush out plentifully on the hillside.
Some of the houses are much larger, and must have contained two or more apartments. But what strikes the eye of the observer most is the relatively large number of purely circular edifices, and this it is that has mainly attracted the speculations of archæologists. Mr. Oswald Crawford, who went over the place whilst the excavations were in their earlier stages many years ago, was mistaken in his estimate that the round buildings were eight or nine times more numerous than the square, and he founded upon this and other data the opinion that the whole place was a great granary, where the food of the tribes might be stored in safety. So far from the round houses being eight times as numerous as the square, found at least four square houses to every one round; but that which struck me as most curious, and so far as I could learn, it had not specially attracted the attention of previous visitors, was that in a large number of cases the round houses were enclosed in a square or angular walled space, not very much larger than the circle, but leaving a passage of some two feet wide, in most cases on the right-hand side, between the two walls, leading to a space at the back between the circular wall and the wall of the square enclosure, the left-hand side of the circular wall being mostly built to touch the square wall on that side. Dr. Sarmento was of opinion that the space thus formed was for the purpose of sheltering cattle and domestic animals, and says that he had found some rough stone excavations like troughs in them, with, in one or two cases, a ring in the wall as if to tether beasts. The width of the entrance passage and the extent of the enclosed space in the rear of the circle would be too small to admit any large animal; but probably goats would be housed in them easily. In one or two cases I noted that the stone post forming a jamb to the entrance to the passage between the round house and the square enclosure was grooved on the inner surface. This in Dr. Sarmento’s opinion proved that the entrance to the passage was closed by a lifting hatch of wood, which to some extent confirms the idea that the back space was intended to shelter animals such as goats, as a lifting door set in a groove would be much less likely to be forced by them than a swing door turning, as the house door did, on wooden pegs.
There are very few instances of party walls being utilised for two adjoining houses, though the buildings are often only a few inches apart. Even in the case of the round houses enclosed in square spaces and touching the square wall, the circular structure is quite complete at the point of contact. In one instance I measured a large walled parallelogram fronting on the principal causeway, seventeen yards in length, enclosing within it one square house of nine yards wide, and two circular houses, one on each side, the structures in each case being complete, but the circular walls in this instance merged for a few inches only at the point of contact with the square outer wall at the side. Whether these square or outer enclosures were tiled or were merely enclosed yards it is difficult to say, but that the houses themselves were so covered is evident from the immense number of well-made red shards scattered everywhere, and particularly inside the houses, the tiles being turned up at each end, so that a concave tile to cover over the joint between them would make a roof covered with them quite watertight. A door jamb and lintel in one house showed a well-carved rope moulding, but in most cases they were plain, the lintels and doorsteps containing, however, at the side a square-cut hollow, in which a block of wood was apparently inserted to receive the wooden peg or pivot which formed a sort of hinge for the door, an arrangement still adopted for the doors of barns, &c., in the neighbourhood; though Dr. Sarmento was of opinion that no wood was employed in the construction of the houses themselves, the polished rounded stones fixed to the walls in some of the houses, which Dr. Hübner considered to be bases of pillars, being in the opinion of the Portuguese archæologist seats for the inhabitants.
The round houses are usually about fourteen feet in diameter, and the walls remaining rarely rise above four or five feet from the surface. The doorstep is usually raised a foot or so above the level of the ground. One round house has been tentatively rebuilt by Dr. Sarmento on the level space on the top of the hill, an unattractive beehive-looking structure without windows, but later investigation convinced him that he had built it too high; and that it should not be of so great an elevation as the measure of its diameter. The principal thoroughfares running transversely on the slope of the hill are carefully walled upon the scarped inner side, and in some cases the stone water channel runs alongside of it.
On reaching the bare space at the very summit of the hill, upon which the little modern Christian chapel stands, a good idea may be formed of the whole plan of the place. The town, covering perhaps five or six acres, all lies over the crest and down the south and south-west slopes. The wall by which we entered from the south is apparently the inner wall of three, and practically encloses the top of the hill and the centre of the town on the slope. The second wall, which shows signs of a moat, is of greater extent, following the irregular contour of the hill, whilst the third or outer defence extends far down almost to the plain on the west and south-west side; traces of buildings, although but little explored, being very abundant between the two inner walls on the south and south-west, and clearly defined paths leading down from the main city to the outer defences and the suburbs. In consequence of the formation of the ground, attack was to be looked for mainly from the most accessible point, namely, the north-east; for here the three lines of defences are almost close together, and each of the walls is here brought to a rough angle. From the apex of the outer wall on this side there are indications of another defence running straight out at right angles along the saddle which connects the hill with an outlying spur easy of approach, and at the end of this long projection there appears to have been two parallel horizontal outworks running across the end of the saddle, this being the vulnerable point of the fortress.
It is easy to imagine how almost impregnable such a place could be made. The hill at any other point than this could only be scaled, if at all, with the greatest difficulty, and the huge boulders on its side would enable even weak defenders under their cover to hurl down stones or spears upon an advancing foe. The south side of the hill is the least accessible of all for any considerable body, and there the defences are the most distant and the weakest.
In the midst of the ruined town I found a bright intelligent peasant lad, busy arranging fragments of pottery upon a stone for the later inspection of some one in authority; and from him I heard much quaint and simple local folklore. His own interest was greatest in what he called the cemetery, four or five small grave-like troughs, about three or four feet long and a foot deep, neatly made and lined with dressed stone slabs. The so-called graves lie close to the causeway and amongst the houses, in an irregular group, and can hardly have been sepulchral, considering their size and position; Dr. Sarmento inclining to the belief that they were troughs for feeding cattle. The cemetery, if there be any, would probably lie far down the slope outside the second, perhaps outside the outer, wall, but here no excavation of any importance has been executed. At some little distance down have been found three perfectly plain dolmens of the usual shape, which are usually sepulchral; and doubtless extensive exploration around them would reveal human remains. My peasant friend was also much concerned in a mysterious “mine,” as it is called, from which he assured me, in awe-stricken tones, that enchanted Moors came at night and carried evil over the plain. It is supposed that this cave, which is of no great extent, some two yards in diameter at the mouth, and a few yards deep, was adjoining or under the place where the great slab which the country-people call Pedra Formosa, the handsome-stone, to which I shall revert presently, was found.
I have mentioned that Mr. Crawford was of opinion that the round houses were granaries, but seeing that the Celts of Ireland and Scotland frequently built and lived in round houses within their cashels, and bearing in mind the existence of the spaces for animals, which I have described as attached to those of Citania, I am strongly of opinion that, comfortless as they appear, these were the veritable dwellings of many of the neolithic folk who for centuries held their foes at bay upon this headland jutting out upon the rich plain of Guimarães. Still another solution of the round-house problem is, as I understand his words, suggested by my friend Professor Altamira in his Historia de Espana y de la Civilization espanõla. The earlier generations of this people, he says, buried their dead under dolmens which when covered were circular; and later generations retained the tradition of circular sepulchres. “They were built round,” he says, “with a sort of domed roof, the middle of which was supported by a pillar of wood or stone. Some of such tombs had passages (or galleries) to enter by—which was frequently the case also with the dolmens—and some had lateral chambers.... Of this class are those discovered at Citania, on the hill of San Roman in Portugal.” Apart from the fact that no human remains have been found in these round houses at Citania, there is no sepulchral suggestion about them. They are, it is true, if Dr. Sarmento be right, windowless and rough, but the comparison must not be made with the dwellings of to-day, but with the haunts of cave men, who had been the progenitors of the early settlers of Citania; and judged by that standard, these stout, weather-proof, stone houses, with doors and an enclosed separate space behind for cattle, were almost luxurious. In any case, a close examination of them left in my mind no doubt at all that they had been the dwellings of human creatures in the earlier stages of civilisation.
It required no great effort of the imagination to people the narrow paved paths on the hillside and the little round central spaces with the dwellers in these rough abodes: wild-looking, shaggy men, with long hair, and clad in skin or rough woollen garments, going about their daily toil as hunters, husbandmen, potters, or smiths, to paint to oneself the alarm of an approaching foe, the savage warfare to repel attack, and finally the victorious host of Roman legionaries of Augustus levelling the poor homes, slaughtering, ravishing, destroying, until the poor remnant of the vanquished knelt in the dust and bowed their necks evermore to the yoke of discipline and civilisation.