82 



height of Ihe cliiniiiey, but have not considered tlis influence 

 of its diameter, as it does not directly affect the velocity of 

 the current ; but it is not to be neglected, for as we have al- 

 ready stated, a column of air of large diameter, will be less 

 cool'd than a small one, and in addition to this, as the current 

 moves through the fuel with a determinate velocity, the wider 

 the chimney is, it will move the slower there, and experience 

 less friction. But still more necessary is it to have it of large 

 dimension, on account of the dilatations wliich the air must ex- 

 perience in passing through the fire-place, for it is obvious, 

 that a free passage should be afforded to all the air which can 

 enter between the fragments of fuel. Now the air in the flue 

 or narrowest part of the funnel may be heated 1000°, and 

 this temperature will triple its volume, the area of the flue there- 

 fore should be three times that of the apertures in the fuel at 

 least. The magnitude of these depends on the quantity of space 

 occupied by the fuel, compared with the vacuities between them, 

 and I endeavoured to ascertain the proportion by the following 

 method : A vessel, whose capacity was known by weighing the 

 quantity of fine sand which filled it, was filled with fragments 

 of coal, of such a size as I found produced the best effect in 

 my furnace (about the size of a walnut) ; sand was then 

 poured in till all the interstices were filled up, and the ratio of 

 this to the entire content of the vessel was the ratio of the whole 

 space to that which was unoccupied by coal. In three 

 sets of experiments made with vessels of different shapes and 

 sizes, I found that the numbers differed much less than I ex- 

 pected, and the mean gave 219 to 100, therefore the solid content 

 of the furnace is to the bulk of the fuel as 219 : 119. These 



