experiments were determined by the shape of the corner bricks which were in the market. These determined the inclination of the sides to be such that, if prolonged, they would meet at a right angle. This is the angle laid down by Rumford as the angle of selection, but as the largest angle admissible in a good fireplace. This angle, however, brought me into difficulties with my "lean-over" back. The openness of the angle made the back, as it ascended, spread out so rapidly that what was gained in width was lost in height. Moreover, my critics objected to its appearance as ugly. What, then, should determine the inclination of the sides? The point was thus determined: Seeing that a heated brick throws off the greatest amount of radiant heat at a right angle with its surface, the "covings" should be at such an inclination to each other that the perpendicular line from the inner margin of one "coving" should just miss the outer margin of the opposite "coving." Where the "covings," as in my earlier attempts and in Count Rumford's fireplaces, are at a right angle to each other, this perpendicular line misses the opposite margin by several inches. It was clear, therefore, that the inclination might be made more acute. Guided by this idea, and having determined the principle on which the shape of the grate should depend, an inclination was arrived at which turned out to be an angle of 60°, i. e., the inclination of the sides of an equilateral triangle.
Count Rumford came very nearly to the same conclusions: "I have said, in my essay on chimney fireplaces, that where chimneys are well constructed and well situated, and have never been apt to smoke, in altering them the 'covings' may be placed at an angle of 135° with the back; but I have expressly said that they should never exceed that angle, and have stated at large the bad consequences that must follow from making the opening of a fireplace very wide, when its depth is very shallow."
Rule VII. "The 'lean-over' at the back should be at an angle of 70°" (Fig. 1).—Commencing at a level (A) corresponding with the top of the front bars, and leaning forward at an angle of 70° with the horizontal line of the hearth, the back should rise to such a point that the angle where it returns toward the chimney (B) should be vertically over the insertion (C) of the cheeks of the fire-grate. This angle (B) will be about twenty-eight inches from the hearth, or sixteen inches from the top of the fire, and about three and a half to four and a half inches from the front line of the fireplace, according to the size of the grate. These points will be obvious from the vertical section of the fireplace here shown, and from C, Fig. 2.
So far, in the fireplaces built after my rules, the height of the grid from the hearth has been taken at two bricks, or six inches, and the height of the bars from the grid also at two bricks, or six inches. It follows, therefore, that the lean-over commences at twelve inches from the hearth. It is possible that a better angle than 70° may eventually
