tionally done to strengthen the parapet, we are met by some difficulties.

In the first place, it is almost certain that vitrification of the larger masses often met with, if intended as a structural method, must have been a troublesome business, and a process to which recourse would have been had only when ordinary building was impossible. But we find in various forts, notably at Tor Duin, near Fort Augustus, and at other places, that the loose stones below the vitrification are supported by ordinary masonry, which apparently might have been carried up the whole way had the builders so desired.

Another difficulty which presents itself is that the vitrification is not as a rule continuous all round the parapet, although for structural purposes, if that were the object, it would seem to be equally required at every point.

But not only is this the case, but we find that the greatest amount of the slaggy mass occurs often-I think I may say generally—where a strong parapet is least needed. At the top of an inaccessible cliff is often found the bulk of the vitrification. This is well seen at Shielfoot, Dunagoil and Ard Ghaunsgail (Arisaig). The last-named fort, which stands on a peninsula, has on the land side a defended entrance which would appear to be the weakest point; but the parapet is there devoid of vitrification, or nearly so.

It will also be admitted that, if the builders were determined to have solid walls, they were not very wise in setting them up on a foundation of loose stone, for they might have anticipated that the vitrified blocks would slip down the hill, as we find a great many of them have done, by the foundation sinking.

Lastly, I think it may be fairly assumed that, if the builders had designed a wall built with a mortar of semi-melted stone, they would have restricted its width to much less than a thickness of say 4 to 6 feet. In this connection, what are we to say to the mass of vitrified matter (described by Fraser-Tytler more than a hundred years ago)

extending along the east end of Craig Phadric, 40 feet wide and 70 feet long?

For these among other reasons, it seems unlikely that vitrification was undertaken as a structural method.

As the air of mystery still hangs over vitrified forts, it occurred to me that it could be to some extent dispelled if we could reconstruct a vitrified parapet; and my purpose to-night is to explain the experiments made with this object during the last five or six years, at long intervals and with insufficient leisure. The positive results have been poor, but a good many negative results have been obtained.

Beacon-fires seemed, at first sight, the most likely source of the necessary heat; so their results were first examined. Undoubtedly a large amount of slag can be obtained from burning grass or straw. This can be easily seen by inspecting the site of any large stack fire. I was fortunate enough (if I may use the expression) to see the results large stack-yard fire which occurred at Hay Mount Farm, near There fifty-seven stacks of grain and eight of hay were conIt was found, where the stacks had been recently erected and the straw was strong, producing an open texture in the heap, so as readily to admit the air, as is the case with a rick of wheat, that there slag at the bottom. The silica had gone off in vapour, which had partially condensed on the lee side of the rick in the form of pellets about the size of a pea. These could be found in considerable number on the ground. Where the material was closer in

Kelso. sumed.

was no

small

texture,

as in the older oat ricks, some slag was seen at the bottom among the débris; but most of the slag was found at the bottom of the closely packed haystacks, where it lay in a nearly continuous crust, in

places about 2 inches in depth.

I should have said that the wind was very high when the fire occurred, and that the combustion of the oldest stacks alone was at all slow.

The chemical analysis, by Professor Macadam, of the slag from this fire is given below, as it may be useful in further investigations:

As the stack fires, when nearly burned out, had been extinguished with water and otherwise, and as the stone bottoming might have thus escaped fire action, the following experiments, among others, were carried out to ascertain the action of open fires having a base of selected stones. The first experiments were made on the high moorland near Riccarton. To begin with, on a base of stones constructed like a saucer, having a diameter of 5 feet, 100 stones of old moorland hay were burned. The process took about eight hours. The result on the stones was nil, but some very small streams of slag were found outside

the lip of the saucer on the lee side of the fire. The stones in the bottom were covered with ash. They were chiefly whinstone.

The next fire was arranged with the stone heap raised in the centre and having a diameter of 10 feet. On this, some hay and an unlimited amount of bracken were burned for six hours. The heat was so great that with difficulty could the men pile on the fuel. The result on the stones in the heap at the bottom was that many of them were fractured and burned, but there was no vitrification.

forts.

The third experiment was carried out on the sea-shore at Arisaig— that district being selected because it is the site of a group of vitrified On the stone heap in this case a mixture of four loads peats, four boat-loads brushwood, twelve cart-loads of sea-weed, and a boat-load of burned, and the blaze was kept up for thirty-nine hours. The result on the stones below-they were principally moine schist— was, as before, fracture and roasting, but no melting or approach to fusion.

grass was

With

this experience I had become satisfied that a beacon-fireunderstood in the sense of an open blazing mass— -had not produced vitrification of stones lying in a heap below it. Another solution had therefore to be found.

In a

good many papers on the subject of vitrification, it is explained

how easily it can be produced; but it does not seem such an easy matter in the field, when you tie yourself down to use only the materials found on the ground. It may be convenient here to refer to a remark made by Mr Ramsay of the Geological Survey, writing in 1859, and repeated later by others.

The observation was to the effect that he thought the vitrification at Knock Farrell had been produced by burning with wood, and explained how the rocks near Barnsley were more or less vitrified by being burned in heaps with coal and brushwood. The stones were described as sandstones. I thought this so interesting that I communicated with the Borough Surveyor, Barnsley, Mr J. H. Taylor, who informed me that

never had any stone been burned, but that some clay shale used to be burned in clamps about sixty years ago, so as to harden it for use on roads with light traffic, or for blinding—as we now see the same thing done on railways for ballast. That is a very different thing from fusing pieces of mica schist such as are found in vitrified forts, and I think some mistake seems to have been made by Mr Ramsay as to the sufficiency of the explanation.

To return to the experiments at Arisaig: What had already been done and observed led to the conclusion that the burning was too rapid, and that the supply of air had to be reduced.

A rough stone square enclosure was therefore set up, the walls being about 3 feet high and the sides about 4 feet long. A layer of loose stones was put in the bottom of the enclosure, and over them peat and brushwood were filled in up to the top of the wall, and a fire set agoing. When the fuel was about half burned, more wood was added, and a layer of stone about 1 foot thick was placed on the top and covered over with peat. As this sank down, hay was added to check the draught.

The result of this was that incipient fusion, producing a sort of glaze, was noted on some of the stones which had been placed on the top and had sunk down into a part of the fire where there was little or no draught. The remainder were only burned. The burning lasted six hours, but the enclosure remained hot for twenty hours.

Another experiment was made under similar conditions, except that the walls were less open and damp moss was placed on the top; but no vitrification resulted.

A third trial was made in the same built enclosure, with 2 feet of small branches below, with some peat and then 1 foot of stone above, and the top was kept covered with rough hay. The stones which had most heat were burned; where they were out of the draught they showed signs of commencing fusion on the surface. This took about nine hours.

Some other similar experiments were made.

It will be seen that the attempt to vitrify by heat above the stone