glass, and for neutralizing the green colour of inferior kinds of glass, and the crystals so coloured are usually called amethyst. The faint violet tint so often seen on ordinary banded agate is due to the same cause, of weak solutions of manganese, which, along with iron, is always to be found in lava.

ONYX.

In a normal agate the layers of chalcedony are deposited in thin bands, one upon the other right up to the centre, and are more or less conformable in pattern to the shape of the cavity; but there is one very remarkable exception to this general rule, in the bands of what is generally known as onyx or opal. In these the dissolved material subsides in horizontal bands to the lowest part of the cavity, and never rises along the walls; that is to say, the layers of an onyx are laid down solely under the law of gravitation, and are always parallel, while those of an ordinary agate may vary in thickness. Conversely, all layers that do not climb the sides are onyx or opal, and all that do so are chalcedony. As no Scottish agates show a second opal layer inclined to the first, it follows that the opals of Scottish agates were deposited before any disturbance of the rocks took place. It may be further stated that it is from these onyx agates that cameos are cut.

MOSS AGATES.

Beautiful specimens of moss agates have been found at Kinnoull Hill and Ballindean. It is still a widely popular, though foolish, belief that the moss-like material enclosed within the agate is real moss, and no argument seems convincing enough to dispel this preposterous idea. When we consider that the rock enclosing the agate was originally a molten mass at white heat, how absurd is it to believe that any vegetable organism could have come through the fiery ordeal without being annihilated. The moss-like structure is, however, so real and life-like that the popular belief still lives, and even good observers have been deceived, and regard the enclosed moss-like material as of vegetable origin, much in the same way as flies are to be found enclosed in fossil amber. It has already been observed that the lining of the cavity is first made up of the most easily dissolved constituent of the rock, to which the name celadonite has been given, but which is popularly known as one of the green earths. It is quite soft and easily detached. Torn and shredded portions hang in irregular moss-like masses from the dome of the cavity, or became detached, and, having afterwards been enveloped in the waterclear chalcedony, show all kinds of an open network of interlacing tortuous strings of celadonite. Moss agates, when cut in

H

thin slices, show the moss-like arrangement to perfection if held up and viewed in the bright light of a gas jet.

As has already been observed, chalcedony, when pure, is waterclear like ice, and of silver grey colour, which deepens into a slate colour in thick masses; if admixed with impurities, these give rise to cream colour, pale lavender, dove and dark slate colour. If the celadonite in solution goes on concurrently with the silica, then a deep green ivy leaf colour is the result, such agates being known as heliotrope; if speckled with red, as bloodstone.

CARNELIAN AND JASPER.

When the clear chalcedony is mixed with some weak solution of iron, the agate is known as carnelian, and is translucent when held to the light. If the iron solution has developed to a greater extent than in the last, then jasper is produced, which, in cut sections, is quite opaque. If the iron solution, instead of ordinary ferric oxide, is ferric hydrate, then the jasper is yellow and opaque. Ferric hydrate gives the yellow tint to bog iron ore, and to the yellow sandstones of the younger geological formations. It is precipitated as a yellow-brown deposit in bog pools and stagnant ditches. This ferric hydrate never occurs in milk-white chalcedony layers, as these do not stain readily. It must be observed, however, that we have conclusive proof that many agates have been coloured, subsequently to their formation, by staining from without, and at a late period, from colouring material percolating from the lake bottoms of the Old Red Sandstone, the process being analagous to that employed in the preparation of biological subjects for the microscope.

XI.-Notes on the Discovery of the Remains of an Earth-house at Barnhill, Perth.

BY ALEXANDER HUTCHESON, F.S.A. Scot., Broughty Ferry.

(Read 13th April, 1905).

In the month of April, 1904, in the course of the construction of a new road at Barnhill, near Perth, on the property of Sir Alexander Moncrieff of Culfargie, K.C.B., some lines of stonework suggestive of a structural formation were uncovered.

The discovery was communicated to Sir Alexander Moncrieff, who immediately stopped the works and made intimation of the discovery to the Society of Antiquaries of Scotland. I was honoured

by the Council of the Society with a request that I should visit the site and report, which I did, and my report appears in their volume of Proceedings for the year. By favour of the Council, I am permitted to give an adaptation of that report for the information of the Perthshire Society of Natural Science, and in doing so I take the liberty of amplifying the introductory part, with the view of recording briefly the general features of the particular class of structures of which that recently discovered at Barnhill is an example.

These structures, so far as known, are peculiar to Scotland, but are fairly common over the whole eastern area from Berwickshire in the south to the Shetland Isles. Many examples have been explored and recorded in volumes dealing with Scottish antiquities, but mainly in the Proceedings of the Society of Antiquaries of Scotland, to which latter volumes I beg to refer those who may desire to investigate the subject more closely. It may suffice to state here a few of the special features of these structures. As the name earth-house usually given to them implies, they are found wholly below the surface, concealment having been apparently one of the purposes they subserved. The usual type presents a long narrow curving gallery, entering by a low and confined aperture nearly on a level with the surface, thence widening and deepening from the entrance inwards, turning usually first to the left in a curving direction until about half the length has been traversed, and then to the right for about an equal distance, it terminates abruptly in a closed end. The letter S-like form is not, however, invariably followed. Some few curve to the left without any turn to the right, others curve at once to the right, while still fewer although curving, deviate very little from a straight line.

The typical mode of construction is side walls built of rough boulders or slabs without any dressing or mortar in the joints, gradually converging somewhat towards the roof, which at a height of 5 or 6 feet from the floor is formed with large and weighty slabs of stone overlapping on the rude walls. In some cases it would seem as if timber must have been used as a roof covering. This may have been the case in the Barnhill example.

Some of the earth-houses possess an inner doorway, probably as an additional protection, while one or two have two external entrances. It is probable that the earth-houses were always associated with overground habitations, whence in the case of danger from an approaching foe, or in winter for greater warmth the people could descend by what was doubtless the well-concealed and narrow entrance into the laboriously constructed secret chamber. As these chambers have not been found elsewhere than in Scotland, underground structures which have been found associated with fortified places in Ireland and elsewhere not being of the Scottish type, it is possible they would be unsuspected

by a foreign foe, but against marauding neighbours retreat into these dark labyrinthine recesses would only be at best a mode of passive resistance to foes who probably would grudge the time and labour, not to speak of the personal danger which breaking in would involve, especially when it is considered that all the plunder they were likely to get was probably already above ground and in their hands.

Dr. Joseph Anderson (Scotland in Pagan Times; The Iron Age) has summed up the evidences and the argument as to the age of these structures. He assigns them to the comparatively narrow period that "will lie between the time of the general establishment of Christianity and the departure of the Romans from Scotland," with much propriety judging that the relics found in the earth-houses and our knowledge concerning them are as yet all too scanty to warrant wider conclusions.

To come now to the Barnhill earth-house, despite of certain peculiarities of construction, which, however, are I think susceptible of explanation, I had no difficulty in recognising the remains as those of one of the underground structures referred to.

I prepared the annexed plan showing the form and dimensions of the remains. The structure has now unfortunately been removed, the completion of the new roadway not having permitted of its retention.

The entrance to the earth-house, which was 2 feet 3 inches in width, faced towards the south-west. The wall forming the left-hand side of the entrance was continued inward to form the western wall of the earth-house, but the right-hand side of the entrance was prolonged inwards only 6 feet 4 inches, and then formed a projection, behind which was a recess 3 feet in depth and about 4 feet in width. The entrance passage, this recess, and a portion of the structure extending backward 8 feet from the inner end of the passage were rudely paved with cobble stones laid on the rock. From this point the earth-house exemplified the usual characteristics of its class by sloping downwards and curving rapidly to the left with an average width of about 8 feet, for a distance in all of about 45 feet from the entrance, measured along the medial line, to where the two side walls abruptly terminate, having doubtless been cut off when the public road between Perth and Dundee was diverted and cut through it, presumably in the early years of last century.

As to the features of the situation, it is known that these structures have been commonly found occupying level or at least arable ground, in other words, sites suitable for, and in modern times given over to, agriculture; hence they have been generally discovered by the plough coming into contact with the roofing slabs, and so leading to an examination of the obstruction. But the Barnhill earth-house has differed from the usual type in occupying the summit of a rocky knoll,

where presumably, if covered or roofed over in the usual way with large slabs of stone, it must have been partially formed above ground, and afterwards covered from sight by earth being heaped above it to such depth as afforded that concealment which seems to have been the invariable rule, if not indeed the originating cause, of the typical form of these structures. No covering slabs now exist, nor has any evidence of them here been discovered. The walls were dry-built, and formed of superincumbent lines of stones. The stones of the first or lowermost line averaged 2 feet 6 inches long, 2 feet high, and I foot 6 inches thick. They were water-rolled and ice-scratched boulders of whin, diorite, granite, etc., nowhere exhibiting toolmarks or any evidences of artificial shaping. The stones had, however, been placed with some recognition of a principle of construction. They were set with their longer axes in the line of the wall, and had their smoothest and flattest surfaces facing inward to the earth-house. At the date of my visit practically only one tier of stones remained, with here and there portions of a second tier; but I learned from the workmen that before the artificial character of the remains was recognised, one and in some places two tiers of stones had been removed from the walls. The result of this removal has been to deprive us of the possibility of now deciding whether the walls in their complete state were erected with that inward convergence of the upper part which characterises these structures.

It was a singular place to select for an earth-house. The rock, which protruded through the surface at the apex of the knoll, must have shown the prospective builders what they had to expect in forming there an underground structure. Are we therefore to assume that they contemplated a certain amount of scarping of the rock to attain their ends? I was at first inclined to think they had done so, from certain appearances of the rock, which forms everywhere the floor, shelving downwards at the entrance, and also on the left side about half way towards the end; but on reflection I gave this up, since, even if necessary to scarp the rock at the entrance and further in, it was not necessary to make this supposed scarping extend underneath the stones forming the side walls, which, on examination, it was found to do. I therefore concluded that the supposed scarping was only the natural slope of the rock; and in this opinion I was glad to have the concurrence of Mr. Alex. M. Rodger, Curator of the Museum of Natural History, Perth, who is well acquainted with the geology of the district. It seems, therefore, that this structure, which conforms to so many of the features of an earth-house that it seems impossible to assign it to any other known class of early structure, yet differed from the type in having been only partially excavated, and consequently formed partly above ground, being afterwards covered over