Another (Fig. 256) excavated a few years ago at Clova, near Kildrummy, also in Aberdeenshire,^[111] differs from these in being so slightly curved to the left as to be almost straight. It measures 57 feet in length, 2½ feet wide at the entrance, suddenly widening to about 8 feet at about 20 feet within the entrance. At a short distance from the entrance there were checks for two doors about 8 feet apart. The covering stones had been removed from the first 15 feet of the narrow part, but the roof remained entire over the whole of the wider part of the structure, at an average height of about 6 feet from the floor. The earth with which the chamber was filled was largely mixed with charcoal and bones of animals, among which those of the horse and dog were recognised. No manufactured relics were found, but two of the stones in the walls, one being a large boulder, were covered with the small hemispherical pits known as cup-markings.

An Earth-house at Eriboll, in Sutherlandshire^[112] (Fig. 257), resembles that at Clova in presenting so little curvature as to be almost straight. The curvature which it has is to the left, and only extends for a few feet within the entrance. It is said, however, to have been 10 or 12 feet longer than it was when examined in 1865. It was then 33 feet in length. It is peculiar for the smallness of its size, being nowhere more than 4½ feet in height, and for the greater part of its length only 2 feet wide, expanding to 3½ for about 3 feet only from the inner end. In view of this feature of its character, Dr Mitchell remarks that it is exceedingly difficult to see what purpose such a structure could have served; but he adds that it is worthy of note that in this district similar underground constructions are not rare, and that they are called by a Gaelic name which signifies Hiding-beds. The use of such underground places of concealment is referred to in the Saga of Gisli the Soursop, which relates to events occurring between the years 930 and 980, and was written in Iceland about the beginning of the twelfth century. It states that when Gisli was outlawed and every man’s hand was against him, he went to Thorgerda in Vadil “She was often wont to harbour outlaws, and she had an underground room. One end of it opened on the river-bank and the other below her hall.” Again it states that “Gisli was always in his earth-house when strangers came to the isle.”^[113] The form of Earth-house thus described as then in use for concealment in Iceland is not the form of the Earth-houses found in Scotland, which have rarely two openings, but the passage is interesting because it shows that the traditional use ascribed to the Scottish examples is a use which was practised among a people who had close relations with the district in which the tradition still remains attached to these structures.

But whatever may have been the actual purpose or purposes to which they were applied, the fact which is of importance in our investigation is that these Earth-houses, though ranging in area from Berwickshire to the north coast of Sutherland, are all of one special character, long, low, narrow galleries, always possessing a certain amount of curvature, sometimes greatly, and at other times doubly curved, always widening and increasing in height from the low and narrow entrance inwards, usually built with convergent walls and roofed with heavy lintels, which are always lower than the surrounding level of the ground, so that the whole structure is subterranean. Occasionally they present variations in structure as in the case of one at Murroes, in Forfarshire, which, instead of being built, has its walls constructed entirely of flagstones set on edge. Similarly, the example at Kinord, in Aberdeenshire (Fig. 258), has its walls constructed of single boulders set on edge or on end, and it presents the further peculiarity of the chamber being divided into two branches at the farther end. One at Pirnie, in the parish of Wemyss, in Fife, and another at Elie, had steps leading down to the entrance.

Occasionally they occur in considerable groups, as at Airlie, in Forfarshire, where there is a group of five. One of these is of great size, its length being 67 feet, and its average breadth, from the farther end to within about 12 feet of the entrance, 7½ feet. The height at the entrance is only about 22 inches, and the floor slopes down for about 20 feet till a height of about 6 feet is obtained. The walls are built of rough undressed boulders laid in pretty regular courses, and they converge from a width on the floor of a little over 7 feet to about 4 feet at the roof. The covering stones are of great size, many of them 7 or 8 feet in length and 4 feet wide. It contained the usual traces of cookery in the accumulation of ashes and bones of animals upon the floor. The only other relics found in it were a brass pin, a stone mortar-like vessel, and fragments of querns. The other four examples in the same neighbourhood are known to have existed, but have neither been measured nor described.

A still more remarkable group was brought under the notice of the Society in 1816 by Professor Stuart of Aberdeen. They are spread over a space of a mile or two in diameter on what was then a dry moor in the parishes of Auchindoir and Kildrummy, in Aberdeenshire. These excavated houses, he says, are most frequently discovered by the plough striking against some of the large stones which form the roof. The only opening to them appears to have been between two large stones placed in a sloping direction at one end, and about 18 inches asunder. Through this narrow opening one must slide down to the depth of 5 or 6 feet, when he comes to a vault generally about 6 feet high, upwards of 30 feet long, and 8 or 9 feet wide. The floor is smooth, as if of clay, and the sides are built of rude undressed stones without cement. The walls bend inwards to form a rude arch, and the roof is covered with large stones 5 or 6 feet long, some of them being over a ton in weight. The whole structure is beneath the level of the ground and quite invisible, but many of them were detected by the existence close to them of a square space about 10 to 15 paces each way dug a foot or two deep with the earth thrown outwards. These he conjectures to have been the sites of the summer huts of the people, who retreated to these underground places in winter, and stored their provisions and concealed their valuables in them all the year round. But he adds that no article of furniture, and no utensils or instruments either of stone or metal have been found in them so far as can be learned, but only a quantity of wood-ashes and charcoal, chiefly at the farther end, where there sometimes appears a small aperture at the top as an outlet for the smoke. The whole number discovered in this locality he estimates at between forty and fifty. They are found, he says, in other localities, but so great a number collected in one place has probably never before occurred. The number is certainly very large, and may probably be over estimated, but it would not be difficult to find in other parts of Scotland, and specially in Aberdeenshire, a series of groups of similar structures which, though not so numerous or so closely aggregated, are so distributed over wide districts as to show that the custom of constructing these underground edifices was general and prevalent. Wherever they occur they present the same individuality of character and the same strongly marked typical features. Their range in area extends from Berwickshire to Shetland. They occur in greater or less abundance in most of the counties bordering on the east coast. A few doubtful examples only are recorded in those bordering on the west coast. But it is only of late years that the importance of securing a permanent and exhaustive record of such casual discoveries has begun to be recognised, and in this direction of defining the areas of the respective types of structural antiquities, we are still groping in darkness on the threshold of a great investigation.

I now proceed to notice a few examples which, by their associations or their contents, disclose indications of the period of the type.

In the spring of 1859 an underground structure of this type (Fig. 259) was discovered on the farm of West Grange of Conan, near Arbroath, Forfarshire. It occupies an elevated situation on the south-east slope of an eminence commanding an extensive view. The structure differs from all those that have been described, inasmuch as in addition to the long, low, narrow, and curved gallery widening and increasing in height from the entrance inwards, which is the typical form, it presents the additional feature of a circular chamber (A) attached to the long curved chamber near the narrow end, and also communicating with the surface by a passage (C C), thus giving to the composite structure a second entrance. The main chamber or gallery is 65 feet in length along the curvature of the central line of the floor. Its entrance is 2½ feet wide, and apparently little more than 18 inches high. It widens but slightly, till at a distance of 20 feet from the entrance there is an offset formed by a large stone set at right angles to the passage, beyond which it widens more rapidly to about 8½ feet across at the farther end. The walls are built of undressed stones, but in some places they are partially cut out of the soft rock, which, for a considerable portion of its length, also forms the floor of the chamber. The circular chamber (A) is about 10 feet in diameter and 7½ feet high. The floor is partly excavated in the underlying rock. The walls are rudely built of undressed boulders. They converge almost from the floor, and the covering stone was a large boulder resting on the circular apex of the vaulted roof which impeded the plough and thus led to the discovery of the structure.

About 20 feet to the north of the underground chamber there was a circular space from which the soil had been removed. It was rudely laid with a pavement of undressed flags forming a circular floor a few inches below the level of the surrounding soil, and about 20 feet in diameter. Among the flags of this paved space there were found a portion of a plain bronze ring about 3 inches diameter, the upper stone of a quern or hand-millstone, two whorls of lead, a number of rudely-hollowed stone vessels of various sizes, and fragments of implements in iron so greatly corroded as to be unrecognisable except as fragments of implements with cutting edges.

The articles found in the underground chambers were few in number. They consisted of some fragments of pottery, coarse, but wheel-made, pale yellow in colour, and differing in texture and manufacture from the usual handmade pottery of native origin found in many of the other structures of the same class. It closely resembles some varieties of pottery that are constantly found in the vicinity of Roman stations in Scotland. A bronze needle and a portion of a quern were the only other objects found. But that the place had been long occupied was sufficiently apparent from the quantity of ashes mixed with calcined and broken bones of the common domestic animals which it contained.

In this case we have distinct evidence of an underground chamber associated with an overground habitation of less permanent structure, of which time and cultivation had removed all traces except the circular paved floor and the casual relics which it contained. There can be no doubt that the people who occupied this overground habitation also possessed the underground structure, and used it for purposes connected with their daily life. There is little now left to disclose what the manner of that life was, but that little is highly significant. It discloses that they were a people cultivating grain and rearing cattle and sheep. They had utensils of stone it is true, and these of the very rudest form and fabrication, but they also possessed wheel-made pottery and weapons or implements of bronze, iron, and lead.

A singular interest attaches to this little settlement, inasmuch as it not only shows us the association of the two forms of underground and overground structure which united to make one habitation, but also gives the associated grave-ground of the family. A few yards distant from the dwelling there was a group of six graves. They were full-length, stone-lined graves, rudely constructed, with three or four flattish slabs forming the sides, and one stone placed for each end. They lay so near the surface that the covering stones had mostly been removed by the plough, and the remains in them were greatly decayed. The only manufactured object found in them was a single ring or child’s bracelet of cannel coal. This is the only instance on record of the discovery of a cemetery associated with the double dwelling of the people who constructed these subterranean galleries.

Among the rubbish thrown out in the course of the excavation there was found a beautiful spiral bronze bracelet of the form of a double serpent, decorated in that peculiar style of art which has been described in the third Lecture of this course as the precursor or earlier development of the art of the Celtic Christian time.^[114] Here we find the earlier art associated with this peculiar type of structure, and with a manner of sepulture which is destitute of all indications of Christianity. It is associated also with wheel-made pottery of a type that is only found in situations suggestive of Roman intercourse, and therefore indicates a period when Christianity had not yet supplanted the Paganism of the country. It was also in a precisely similar association with one of these underground structures that the massive bronze armlets (Figs. 115, 116), described in the same Lecture, were discovered at Castle Newe. They also are decorated in this peculiar style of art and enriched with enamels. Their workmanship evinces skill and taste of a very high order, and the occurrence of these works of art in such associations may serve to remind us how greatly we should have erred if we had estimated the capacity and culture of the inhabitants of these structures by their architectural character alone, or if we had measured their condition and acquirements merely by the fact that they burrowed under ground.

Another structure of the same type (Fig. 260), but of larger dimensions, was discovered in 1871 in a field at Tealing. It was 80 feet in length measured along the curve, 3 feet wide at the entrance and widening gradually to 8½ feet at the inner end, where it is a little more than 6 feet high. It has checks for a door at a little distance within the entrance, and a second pair about 16 feet from the farther end. The usual evidences of occupation were found in the presence of ashes, charcoal, and animal bones throughout the excavation. The manufactured relics unfortunately have neither been described nor figured, although they constitute the largest and most varied collection of objects ever obtained from such a structure. They are enumerated by Mr. Jervise as follows:—A piece of the red lustrous ware commonly called Samian, a bracelet, bronze rings, and coarse pottery, no fewer than ten querns, a number of whorls and stone cups, and an article made of iron slightly mixed with brass. The occurrence of the red lustrous ware in these Earth-houses, as well as in the Brochs and Crannogs, is an indication of the period of the occupation of these structures which is of great significance. The large size of the gallery, in the present instance, and the occurrence in it of ten querns, indicate that it was frequented by a considerable number of people. It has another feature of interest in the presence, on one of the rude boulders which form the walls, of a number of cup-markings, one of which is surrounded by five concentric circles. Another stone with forty-six cup-markings on it lay on the margin of a circular paved space close to the entrance of the structure. These cup-markings form one of the enigmas of archæology. They are shallow pits, roughly hemispherical in form, hollowed by pointed tools in the surfaces of rocks, boulders, and standing stones. Sometimes they are on vertical surfaces, sometimes on horizontal surfaces, occasionally on the under surfaces of stones placed as the covers of cists. Most frequently the cups are simple rounded hollows, but very frequently they are surrounded by a series of concentric circles of varying number, and often a straight gutter proceeds from the central cup through the circles. They are sometimes hewn in groups upon the solid rock of a hillside, sometimes on earth-fast boulders, occasionally on the stones of stone circles, and often on stones in sepulchral cairns or in connection with cists. They are not confined to Scotland, or even to Britain. They are found in Scandinavia, in France, in Germany, and Switzerland. They appear on the Continent in associations which refer them to the Bronze Age at least, but they also occur in associations which show that the custom survived to the late Iron Age, and even in a modified form to Christian times. Their occurrence here, in connection with this underground structure, has therefore no special significance with respect to the age of the structure, and there is nothing in the association or the circumstances in which they occur in this particular instance which contributes to our knowledge of the purpose or significance of the markings themselves. They may or may not have been sculptured on the stone before it was taken to form part of this underground gallery, and the only thing they tell us for certain is that here, at some time or other, there was a custom of which traces are found scattered over a wide area of Western Europe.

But other indications have been found in connection with the structure and contents of these singular buildings, which carry the period of their construction close up to the time of the Roman occupation of the southern portion of Scotland. An underground structure of this special type (Fig. 261) was discovered near the village of Newstead, in Roxburghshire, in 1845.^[115] It fortunately came under the observation of Dr. John Alexander Smith, who has given a carefully prepared notice of its peculiarities in the Proceedings of the Society. It was of the usual form, a long, low, and narrow gallery turning sharply to the right and widening and gradually increasing in height from the entrance. It measured 54 feet in length along the curve of the central line of the floor, and widened gradually from 4 feet at the narrow end to 7 feet at the farther end. The height was not ascertainable, as the roofing stones were gone, and scarcely more than 3 feet of the height of the side walls remained. But the walls presented the peculiarity of being built with hewn stones, laid in pretty regular courses, though not jointed with mortar or any other cement. Among the fallen stones in the interior of the structure there were many flat slabs bevelled on one edge, and two measuring about 4 feet in length which presented a rope-moulding (Fig. 261) of distinctively Roman character. No relics were obtained from the excavation of the building, but the character of the squared and bevelled stones and the presence of the Roman moulding indicate that the construction of the underground structure was subsequent to the period of the Roman occupation of that part of the country. Another structure of similar character was found in an adjoining field, but not built with squared stones. In all probability the squared stones of the one structure were due to the presence in its immediate neighbourhood of some Roman construction, the stones of which were utilised by the underground builders.

Another structure of the same type (Fig. 262) was found in 1869 at Crichton Mains, in Midlothian.^[116] It was of the usual long, narrow, curved form, 51½ feet in length, and gradually widening from 5 feet 10 inches just within the entrance to 9 feet at the farther end. A number of the roof-stones remained in position, and the floor throughout being formed of the natural rock it was seen that the average height was about 6 feet. The walls converge from the floor for about half their height and rise somewhat perpendicularly above that, thus giving to the cross-section the form of an ogee vault. The door (A, in Figs. 262, 263), formed of two upright stones crossed by a lintel, is 3 feet high by 33 inches wide, and the top is about 5 feet under the present surface of Fig. 264.—Ambry in Earth-house at Crichton Mains.
Fig. 264.—Ambry in Earth-house at Crichton Mains. the ground. Fourteen feet inwards is a passage at right angles to the gallery, the entrance to which is shown at F in Fig. 262. It is 13 feet long, 3 feet wide, and 3 feet 6 inches high, rising by low sloping steps in the rock to what seems another entrance nearer the surface, also shown in Fig. 262 at E. At G is a small ambry (Fig. 264). No relics were found in the excavation; but here and there in the interior faces of the walls there were a number of squared stones faced with the diagonal broaching and stugging which is so common in late Roman work.^[117] About thirty of these are visible. Two are shown in Fig. 265. There were also other hewn stones, some of which had apparently formed portions of water conduits (H, Fig. 265), and others adapted to different Fig. 265.—Stones in Earth-house at Crichton Mains.
Fig. 265.—Stones in Earth-house at Crichton Mains. purposes. The lintel-stone of the door of the side passage is moulded and bevelled on the edge in a similar way to the bevelled stones found in the structure at Newstead. It seems evident that this underground building, like that at Newstead, has been partially constructed with stones taken from a ruined building of late Roman workmanship, and that both are consequently later than the commencement of the Roman occupation of the country. Similar indications are given by the discovery of wheel-made pottery of Roman type in the Earth-house at Cairn Conan, and of fragments of the red lustrous ware commonly called Samian in the Earth-houses at Tealing, Pitcur, and Fithie. The presence in most of them of querns and implements of iron, and the entire absence of such implements as are characteristic of the ages of stone and bronze, are indications pointing to the same conclusion. On the other hand there is a complete absence of indications of Christianity, and the characteristics of the ornamentation of the bronze armlets found in association with them are those which belong only to the earlier and partial development of Celtic art, which preceded its subsequent and complete development under the new impulses and opportunities afforded by Christianity. It seems therefore that, so far as our present knowledge will carry us towards a definite conclusion, the period of this peculiar type of structure will lie between the time of the general establishment of Christianity and the departure of the Romans from Scotland.

The range of the type includes the whole eastern area of Scotland, stretching from Berwickshire on the south to the Shetland Isles. Its special form is so peculiar that it must be held to constitute a distinction sufficiently characteristic to separate the Scottish group from all other varieties of underground structures, and sufficiently constant to warrant us in assigning to it a specific value. There is an Irish group and a Cornish group of underground structures, but they do not generally present the special features of form which characterise the Scottish group. The Irish examples are usually associated with raths, thus resembling the specimen in the rath of Dunsinnane, which is the only one known of that special variety in Scotland. They are excavated in the area enclosed by the interior rampart of the rath, and consist of one or more chambers, sometimes circular or oval in plan but often rectangular, and connected together by low narrow passages. Sometimes the chambers are lined with masonry, and roofed by overlapping courses forming a rude dome-shaped vault; at other times they are simply excavated in the hard earth, while the passages and doorways are lined with stones. They thus differ considerably in form from the Scottish variety, and they differ also in being usually associated with raths or earthworks, while the Scottish structures are usually contiguous to the sites of overground habitations.

The general features of the Cornish group, on the other hand, are more allied to those of the Scottish area, inasmuch as they are often associated with clusters of overground habitations. One at Chapel Euny, in the parish of Saucreed, near Penzance, contiguous to the sites of four circular huts, is an underground gallery presenting features of remarkable similarity to that at Cairn Conan, in Forfarshire. The gallery, which is slightly curved, is about 60 feet in length, 6 feet wide, and from 6 to 7 feet high. A circular chamber, 16 feet in diameter, constructed of large granite blocks, each overlapping the one below it and thus forming a domed roof which must have been 10 or 12 feet high, was connected with the wider end of the gallery by a passage 10 feet long, opening off one side. Another small offset near the narrower end of the gallery, also about 10 feet long, slopes up to the surface, presenting an entrance doorway 2 feet 8 inches in height, with recesses on either side as if to retain a slab to close the doorway. The floors of the gallery and chamber were paved with flat stones, and provided with drains underneath the pavement. The relics found in the structure were whetstones; hammer-stones; spindle-whorls; several varieties of domestic pottery, red and black, mostly plain, but occasionally ornamented with markings made by a pointed instrument; an iron spear-head; and a fragment of the red lustrous ware commonly called Samian.^[118] Another at Halligey, near Trelowarren, consists of a slightly-curved gallery 90 feet in length, from 3 to 5 feet in width, and about 6 feet high in the middle, becoming lower towards the extremities. It has a small rectangular chamber off one side at the farther end of the main gallery. The main gallery opens off the middle of the side of a shorter and wider gallery 28 feet in length, 5½ feet wide, and 6 feet high. At one end of this shorter gallery a narrow passage rises to the surface. The entrance passage is provided with checks for two doors, and the whole structure is strongly and substantially built and lintelled with large flags. On the surface there are traces of two embankments with an intervening ditch surrounding a large area within which there may have been a cluster of overground structures.^[119]

Like the Scottish examples the Earth-houses of Cornwall are long narrow galleries of dry-built masonry, but they are not so strongly marked by the peculiar feature of single or double curvature which distinguishes the Scottish group. They are comparatively few in number, and any indications of the period of their occupation that have been observed point also to a time not far distant from the close of the Roman occupation of the country. No other group of such underground structures is known in any other part of Europe, or indeed anywhere else in the world. These excavated chambers, possessing the characteristics which have been described, are peculiar to the Celtic area, and the specially typical form with the strongly marked curvature is found only in Scotland.

Of the culture and civilisation of the people who constructed these strange subterranean cells, it may be impossible in the present condition of our knowledge to form an adequate estimate. But we can say this of them with certainty, that whatever may have been the special motives and circumstances that induced them to give this peculiar expression to their architectural efforts, they exhibit in other respects evidences of culture which, though it may be held to be inferior in range and quality to the culture of the Christian time, compares not unfavourably (so far as it goes) with that which is exhibited in connection with the superior architecture of the Brochs.

And while on all these lines of investigation we have traced the manifestations of these early forms of culture and civilisation up to points at which they seem to touch the culture and civilisation of the Roman Empire, it is to be observed that they do no more than touch it—they are not merged in it. In all their distinctive features they are still Celtic, and Celtic exclusively. There is nothing Roman in the forms of the prevailing types; there is nothing Roman in the art that decorates these forms; there is nothing Roman in the typical character of the structures in which they are found. The forms, the art, and the architecture are those of Scotland’s Iron Age—the Pagan Period of her Celtic people.