The question of a fire resisting roof is one which has long exercised, and still continues to exercise the minds of architects and builders. An ordinary slate roof has no pretensions that way; tin is an acknowledged failure; gravel is by no means infallible; and where the accustomed red tile is used success is by no means a certainty. Into these considerations, of course, the new style of fireproof roof as practised today does not enter. Such buildings are not in common use for other than business, manufacturing, hotel, large apartment, or assembly room and theatre purposes. What is wanted is a fire-resisting roof for the ordinary dwelling or the smaller stores.


Towards attaining the requisite knowledge on this subject the British Fire Prevention committee has been experimenting as to what style of roof will best withstand a fierce fire of one hour, beginning at a temperature of 500 degrees Fahr., gradually increasing to 1,500 degrees, followed by the application of a stream of water for three minutes. The superficial area covered by each roof experimented upon was about 100 square feet; the time allowed for construction (the work being done in spring) was about four weeks. The slated roof was constructed of deal. The wall-plates were four and onehalf inches by three inches; rafters, five inches by two inches; ridge, nine inches by one and one-half inches; ceiling joists, four inches by two inches; slate battens, two inches by three-quarters of an inch; gutter boards, one inch thick; and bearers, one inch by three inches. The roof was covered with American green slates, twenty inches by ten inches, and three-sixteenths of an inch thick, with a two and a half-inch lap, and each slate was nailed by two zinc nails. The ridge was covered with blue Staffordshire ridging. The laths were sawn sprucelaths one and one-quarter-inch by one-quarter-inch The first coat of plastering consisted of coarse stuff composed of lime and sand in the proportion of 1 to 3, mixed with a proportion of hair, and a small admixture of plaster of Paris (1 in 16). Th.e second coat was of the same material, finished fair; the total thickness of plastering was about one inch. The gutters were lined with No. 14 gauge (Vielle Mnntagne) zinc. After the ceiling was completed a salamander was lighted in the chamber, for one day. to assist in drying the plaster. The Vulcanite roof was also of deal. Its wall-plates were four and onhalf inches by three inches; joists, seven inches bv two and one-half inches; boarding to flat, one and one-quarter inch thick in four-inch widths, grooved for iron tongues. The hoop-iron tongue was seven eights-inch wide; the angle fillet, four and one-half, inches by three inches. The Vulcanite roof covering consisted of asphalted felt in three thicknesses, laid to break joint, cemented together with a composition termed “Vulcanite,” which was brought to a plastic state by being heated in a cauldron and applied to the asphalted felt with a brush. The edge^ were turned up against the angle fillet and covered with a zinc flashing of No. 14 gauge (Vieille Montagne) zinc. The top of the vulcanite roofing was covered with a layer of gravel and sand two and one-half inches thick. The laths were sawn sprucelaths one and one-quarter-inch by one-quarter-inch. The first coat of plastering consisted of coarse stuff composed of lime and sand in the proportion of 1 to 3. mixed with a proportion of hair and a small admixture of plaster of Paris (1 in 16). The second coat was of the same material, finished fair; the total thickness of plastering was about one inch. After the ceiling was completed, a salamander was lighted in the chamber, for one day, to assist in drying the plaster. The roofs were then covered with tarpaulins for twenty-five days.

The gas having been lit (in the inside chamber in each case) for the tests, in thirteen minutes the plaster began to fall in patches from the slated roof, half falling in sixteen minutes. Three minutes afterwards the ceiling joists were well alight, and in two minutes more all the interior of the roof, with melting zinc dropping through fourteen minutes afterwards. In forty-two minutes after the gas was lit the first slate fell, followed by others in rapid succession. In forty-seven minutes the burning rafters began to fall, and in forty-eight the roof collapsed. In one hour from the beginning i f the test (in each case) the gas was extinguished, and 1 be test closed. The gas being lit at the same moment under the Vulcanite roof, it was forty minutes before the plastering began to fall and fiftyfour before the underside of the flat surface was a sheet of flame.

Observations on the outside of each roof were as follows: In one minute after the gas was lit. smoke was seen to issue from the undercover flashing of the gutter on the north side (slated). In five minutes it was issuing through the joints of the slating, increasing in two minutes, and in nineteen minutes be ing so dense that the slates could be seen only at intervals. mam being cracked and broken. In twentv-seven minutes from the beginning of the test flames, not continuous, were seen on the north slope: seven minutes afterwards on the south slope, near the gutter; in another minute showing all along the flashing of the north side gutter, and in another minute long and continuous on the south slope, near the gutter, with the north side woodwork well alight. In forty-one minutes they showed along the edges of the ridge tiles and slates, the latter beginning to fall at the southeast corner. In five minutes more the rafters began to fall. In sixtyone minutes after the test began (two minutes after it closed), the water was applied and turned off in three minutes. In the case of the Vulcanite roof, it was thirty-seven minutes before any smoke was seen at all. and then it was very slight and issuing from under the cover of the flashing about twelve inches from the west wall. Seven minutes afterwards slight smoke issued from under the flashing at northwest, southwest, and southeast angles: in three minutes more from under the flashing all round. In twelve minutes—one minute before the test was closed—the smoke was very dense and continuous, especially at the southeast angle. The water at the close was applied and turned off as in the case of the slated roof.


On the same day observations were taken as follows: The slated roof was entirely consumed, only the charred remains of the wall-plate remaining in position. The Vulcanite roof was sound enough to be walked and jumped upon, and the fire had not passed through it. The gravel and sand were cleared off for about a square yard, and the Vulcanite covering removed. It was soft and pliable, but otherwise unaffected by the fire. The top of the boarding was also clean and unaffected by the fire. A portion of the boarding ws cut out. and the underside was found burned through to the iron tongue, except where it passed over the wooden joists. The next day no plastering remained in position, and the joists to the flat were reduced by charring to six inches bv one and five-eights inches, the boarding to three-cigfiths-inch.

No posts to display