The late Sir George Mackenzie of Coul adopted yet a fourth view. He held that the vitrification is simply an effect of the ancient beacon-fires kindled to warn the country of an invading enemy. But how account, on this hypothesis, for ramparts continuous, as in the case of Knock Farril, all round the hill? A powerful fire long kept up might well fuse a heap of loose stones into a solid mass; the bonfire lighted on the summit of Arthur Seat in 1842, to welcome the Queen on her first visit to Scotland, particularly fused numerous detached fragments of basalt, and imparted, in some spots to the depth of about half an inch, a vesicular structure to the solid rock beneath. But no fire, however powerful, could have constructed a rampart running without break for several hundred feet round an insulated hill-top. "To be satisfied," said Sir George, "of the reason why the signal-fires should be kindled on or beside a heap of stones, we have only to imagine a gale of wind to have arisen when a fire was kindled on the bare ground. The fuel would be blown about and dispersed, to the great annoyance of those who attended. The plan for obviating the inconvenience thus occasioned which would occur most naturally and readily would be to raise a heap of stones, on either side of which the fire might be placed to windward; and to account for the vitrification appearing all round the area, it is only necessary to allow the inhabitants of the country to have had a system of signals. A fire at one end might denote something different from a fire at the other, or in some intermediate part. On some occasions two or more fires might be necessary, and sometimes a fire along the whole line. It cannot be doubted," he adds, "that the rampart was originally formed with as much regularity as the nature of the materials would allow, both in order to render it more durable, and to make it serve the purposes of defence." This, I am afraid, is still very unsatisfactory. A fire lighted along the entire line of a wall inclosing nearly an acre of area could not be other than a very attenuated, wire-drawn line of fire indeed, and could never possess strength enough to melt the ponderous mass of rampart beneath, as if it had been formed of wax or resin. A thousand loads of wood piled in a ring round the summit of Knock Farril, and set at once into a blaze, would wholly fail to affect the broad rampart below; and long ere even a thousand, or half a thousand, loads could have been cut down, collected, and fired, an invading enemy would have found time enough to moor his fleet and land his forces, and possess himself of the lower country. Again, the unbroken continuity of the vitrified line militates against the signal-system theory. Fire trod so closely upon the heels of fire, that the vitrescency induced by the one fire impinged on and mingled with the vitrescency induced by the others beside it. There is no other mode of accounting for the continuity of the fusion; and how could definite meanings possibly be attached to the various parts of a line so minutely graduated, that the centre of the fire kindled on any one graduation could be scarce ten feet apart from the centre of the fire kindled on any of its two neighboring graduations? Even by day, the exact compartment which a fire occupied could not be distinguished, at the distance of half a mile, from its neighboring compartments, and not at all by night, at any distance, from even the compartments farthest removed from it. Who, for instance, at the distance of a dozen miles or so, could tell whether the flame that shone out in the darkness, when all other objects around it were invisible, was kindled on the east or west end of an eminence little more than a hundred yards in length? Nay, who could determine,--for such is the requirement of the hypothesis,--whether it rose from a compartment of the summit a hundred feet distant from its west or east end, or from a compartment merely ninety or a hundred and ten feet distant from it? The supposed signal system, added to the mere beacon hypothesis, is palpably untenable.