I 



83 



maghitudes however have 3 dimensions, while those which we 

 wish to compare, namely the horizontal section of the furnace and 

 the portion of it occupied by fuel have only two, and therefore are 

 as the squares of the cube roots of the above numbers, that is as 

 1,502 : 1, or nearly as 3 : 2, the apertures in the horizontal section 

 are then about one third of the area of the furnace. It appears 

 fi'om what has been previously stated that the chimney should be 

 thrice as great or equal to the total area ; and this gives the con- 

 struction wliich is recommended by Baume and Saussure, a ver- 

 tical tube of uniform diameter with a grate at the bottom. The 

 intervals between the bars of the grate must be greater than ~ of 

 the area, because they are partly obstructed by the coals, but as 

 the fragments of fuel are in this case in contact with flat surfaces, 

 about zds of the intervals are pervious ; on this supposition the 

 breadth of the bars should be equal to the intervals, and no evil 

 can arise from making them narrower. Furnaces are in general very 

 badly constructed in both these respects, the grate is often made 

 of square bars whose thickness is three times their interval, and 

 it seems to have been feared that the heat might escape up 

 the chimney, to judge from the smallness of its aperture. Pott 

 allowed a flue of 4 inches for a fireplace of 18, and Boerhaave 

 one of 3 to a fireplace of 12, even now we sometimes see them 

 in the proportion of one to four, but we may be certain that all 

 such perform badly. We are not sufficiently acquainted with 

 pneumatics to pronounce on Venturi's chimney, whicii depends 

 on the fact that a diverging ajutage increases the discharge of 

 water from an orifice ; if the same is true of air, the tube should 

 be a frustum of a cone with the wide end upward, and De 



N 2 



