ticular inclination for the lens panels, but when that inclination was settled, the lantern should be constructed to correspond. The uniformity of the light from the apparatus used by Mr. Douglass would be improved by having the racks of the upper and the lower reflectors placed so as not to come vertically over each other. He strongly disapproved of vertical framing for the reflectors, while inclined framing was used in the lantern; and of inclined framing for the reflectors, when a different inclination was used in the lantern; as under such circumstances the framing of both the apparatus and the lantern stopped light. The Trinity House light at Gibraltar was marred by the vertical lantern bars not coinciding with the vertical framing of the apparatus, by which the obscuration of light was more than doubled.

As lanterns of a sufficient strength could be constructed on any of the known plans, it was unnecessary to discuss which form of framing was theoretically the most perfect, in regard to a maximum strength with a minimum weight of material; but it was important to construct a lantern of ample strength with a minimum loss of light, and at the least expense. He claimed for his lantern superiority in these three points. He considered a shallow capping inch thick more suited to prevent the accumulation of snow on the lantern framing than one upwards of 1 inch deep, as used by Mr. Douglass. During a snow-storm one of the light-keepers should be outside the lantern to sweep the snow off the capping and glass. He was of opinion that Mr. Douglass' lantern would carry a greater bulk of snow than any other. No doubt, when Professor Faraday reported in reference to Mr. Douglass' lantern that it was nearly a matter of indifference as regards shadow how the uprights of the optical apparatus were placed in relation to such a lantern," he was aware that the whole of the framing would stop light, no matter where placed.

The composition of the glass varied considerably; and as a natural result the index of refraction, colour, hardness, and nonliability to change colour, all varied. The allowance of inch for grinding was necessary on account of the rough surface of the cast glass, the inaccuracy of shape resulting from the imperfection of the moulds, and the contraction in cooling and handling which sometimes took place before the glass was set.

The lantern which he had described was as good as any of the old lanterns for illuminating the sea from the horizon to the tower: it was also suited for any required dip, as that was accomplished in the apparatus, and no obstruction resulted from the lantern framing, unless it was incorrectly placed. The brightest light was generally sent to the horizon, but with a properly adjusted apparatus there would be a good light in-shore.

CHARLES HUTTON GREGORY, President,
in the Chair.

No. 1,185.-"The Roman Rock Lighthouse, Simon's Bay, Cape
of Good Hope." By JOHN FREDERICK BOURNE, M. Inst. C.E.
SIMON'S BAY, which is the north-western indentation of False Bay,
contains the dockyard and depôt of the Cape Naval Station; and is
the port chiefly sought by merchantmen of all nations coming in
distress from the eastward, or damaged at sea immediately off the
'Cape of Storms.'

The Roman Rocks consist of a large mass of granite boulders, lying in the bight of Simon's Bay, about 2 miles from the shore, and immediately in the track of shipping entering or leaving the harbour, which is inside the point off which these rocks are situated. There are ship channels inside and outside the rocks. Until within the last seven or eight years, a lightship had been moored to leeward of them. But the lightship showing signs of decay, and occasionally getting adrift during heavy gales of wind, the Board of Trade, at the instance of the Admiralty, decided on erecting, instead of it, a lighthouse upon one of the rocks. A plan and contour of one of the rocks were sent home about eleven years ago; and the late Mr. Alexander Gordon (M. Inst. C.E.) was intrusted with the design, and with the superintendence of the construction of the iron-work and lantern in England.

The design was for a circular tower, 15 feet in diameter and 48 feet in height, of cast-iron plates, with a central column, 16 inches in diameter, in which the weight of the revolving machinery descends. The tower is surmounted by a lantern and gallery as usual. There are eight plates in the circumference of the tower, and six plates, each 8 feet long, in the height, the plates being 14 inch in thickness. But to admit of the horizontal joints of each vertical set of plates breaking joint with those of the contiguous vertical sets, there are four plates of 4 feet high each, and four of 8 feet each in the first and last horizontal sets of plates. The door cill and level of the first floor are 24 feet from the foundation; the whole interior up to that level being intended to be filled in with concrete. Below that level the flanges of the plates are outside; above it they are inside. If the building, as designed, had been skilfully and carefully erected, and filled in [1868-69. N.S.]

E

1

1

with good material, there is no reason to doubt that it would have answered the required purpose, and have stood well. But it failed, and was condemned as being dangerous.

The object of this Paper is to point out the causes of failure, and to give an account of the mode of securing the tower against further injury.

Of the Roman Rocks the principal is about 80 feet in length at low water, by about 40 feet in width at the widest part, which is about one-fourth of the distance from the leeward, or shore end. The rock tapers irregularly from this spot to a sharp point at the windward or southern end, and it is somewhat rounded off towards the shore end. The extreme point, to windward, stands about 6 feet above the level of dead low water at spring tides; the centre of the rock is about 3 feet lower than the point, and the contour rises again, to about 5 feet above low water, at the highest spot near the rear. This rock, therefore, may be compared to the back of a horse, with the shoulder to windward. The best position for the lighthouse would certainly have been about 10 feet further to the rear than that actually selected, in fact, the highest available position; but, unfortunately, the hollow of the back was chosen, on account of its being more level. Spring tides rise about 5 feet; neap tides 1 foot or 2 feet, and sometimes scarcely at all. There is almost always a heavy surf breaking upon the rocks; and only on one day, in the course of three years (1865-7), was it so calm that no sea broke over at low water. On that occasion, for an hour or more, water had to be dipped up for the use of the drillers.

The highest position should have been selected, to prevent the workmen being impeded when cutting the foundation pits and fixing the first row of plates, but the lowest spot was chosen; and this was the first cause of trouble, and led to immense additional expenditure of time and money. For even at dead low water, and when it is considered quite calm, there is invariably sufficient swell to strike the point of the shoulder; and it was always found that the 'run' of the sea swept over the backfrequently a 'run' from each side, first from one side and then from the other: only one now and then, a 'master sea,' being of sufficient bulk and force to wash completely over the rear.

The lowest spot having been chosen, the next error was cutting the foundation pits too deep into the rock, for the purpose of getting as much solid core as possible for the inside of the tower. In order to obtain a core of 6 inches at the lowest spot it was necessary to leave it 2 feet 9 inches high at the highest point; and as the groove was formed by blasting, for the sake of saving labour and time, the rock was much injured. Every sea of course filled the annular foundation pit, rendering it difficult to work.

Two channels were therefore made, by blasting, one on each side, to allow the water to run off; and these channels were very annoying at a later period. It was found impossible to cut the foundation pit true and level, or so difficult that the attempt was abandoned; and the holding-down bolts were so imperfectly secured, that some of them drew when screwed up. Nor was the circle true in plan. Not only had the bottom flanges of the plates to rest upon uneven bearings, being wedged up in some places with blocks of teak, but they were forced, when screwed together, to take a form to which they were not cast.

It was necessary to wait for fine weather to set the first tier of plates and fill it in with concrete to the level of the half plates, or 4 feet from the bottom flange. It was feared, it is said, that the concrete would swell and burst the plates; and therefore a system of sheet-iron bulk-heads, not originally designed, was introduced; and wooden shores were also inserted and wedged up, which would, no doubt, have a worse tendency than the concrete, especially if the wedges were dry when driven. At all events, it so happened, that, when the plates were all tightly bolted together, and the concrete was filled in to its full height of 24 feet, the plates began to crack vertically in six different places, one crack extending 28 feet high; so that it became necessary to hoop the tower with wrought-iron hoops, of double 1 inch round iron, the hoops being about 6 inches apart.

In this condition the lighthouse was completed and was used for some time. The erection occupied five years, and the cost is stated to have been about £17,000. It had been previously estimated that one year would be sufficient for the erection; but the difficulties arising from the constant surf, the tides, and the weather were altogether unknown.

The

An arrangement had been made, between the Imperial and the Colonial Governments, that the lighthouse was to be built by the Imperial Government, and, when completed to the satisfaction of the Colonial Government, to be thereafter maintained and lighted by the colony. But, on inspection, the state of the plates was found to be very bad, and the wrought-iron hoops were already rapidly corroding; as, on account of the surf, it was not possible to give the lower ones any efficient protection. structure was reported as altogether insecure; and the Colonial Government refused to take it over for maintenance in its patched up and unsafe state. It was, however, lighted by the colony. In 1862 and 1863 a long correspondence took place between the two Governments on the subject; and soon after the Author landed at the Cape in 1863 he was desired by Sir Philip Wodehouse, the Governor, to inspect the lighthouse and report to him. This led to a proposal being made by the

Author, which was eventually adopted by the Board of Trade, after it had been approved, in preference to others, by Sir Baldwin W. Walker, the Admiral in command of the station, that the tower, as it stood, should be surrounded to the level of the first floor, a height of 24 feet, by a concentric ring-wall of granite, 4 feet thick, with a backing between the wall and the iron plates of about 8 inches of cement concrete. The proposal was to have, in the first six courses, sixteen stones only in each course, fastened together with dowels; each stone to be accurately dressed, to be about 2 feet thick, 4 feet wide on the bed, and to be carefully laid in Portland cement, which the Author had successfully used for about eight years in marine works in British Guiana. The backing was designed to consist of the débris of the quarry, mixed with the same cement. The upper courses were to be of rather smaller stones, dressed only on the face and bed; to be of irregular width on the bed, as the quarry might allow; and to be backed by rubble masonry in cement up to the plates. The coping was to be of dressed stones extending to the flanges of the iron tower. Iron stanchions and a hand-rail were to form a gallery round the tower at the level of the first floor.

As it is seldom possible to land on the rock, even at low water, for the purpose of visiting the lighthouse, a light stage had been constructed, on a level with the floor of the house, and extending from the door towards the shore, over the rear of the rock, above which it was supported by two iron stanchions. Gallows posts, with block and tackle, and a Jacob's ladder,' to be let down when required, formed the means of communication with boats, which could only approach from the leeward, or shore side. There a mooring had been laid down, with a buoy and surf line to the iron stanchion of the stage.

The same system was adopted for the landing of the stones, &c.; but the whole had to be renewed with heavier material, as the stage was required to carry a traveller, with a stone weighing upwards of 3 tons, not hanging steadily, but jerking and swinging with the excessive motion of the barge. To carry the stones round the house to their proper sites, it was proposed to fix a circle of fished railway bars to the tower, on brackets projecting about 2 feet from the plates; and to use a traveller with only two double-flanged trucks, the hook to which the stone was to be suspended being bent round and below the rail, so that the point of suspension should be directly under the tread of the truck wheels. This was found to answer exceedingly well. The rail was below the level of the landing-stage, so that a stone might run completely round the tower, and it was afterwards built into the wall, and the upper stones were run round on iron rollers coned to the proper circle.