MR. thedgold's theory. 27 



olf heat over the descending pipe, as in the annexed figure ; cooling in this 

 manner will increase the velocity. 



6th. The aperture of the upper pipe should not be more than about 

 one inch below the surface of the water, or as much as prevents it draw- 

 ing air, in an open boiler, but the lower it is below that, the less effect 

 v^dll be obtained ; the lower pipe should enter the boiler w^here it has less 

 tendency to cool and check the fluid rising to the upper pipe from the 

 fire surface. 



7th. In a close boiler a pipe, at any distance from the boiler, may 

 rise to any height and descend again, but it must neither rise twice nor dip 

 after leaving the boiler ; when it is necessary to raise it, there should be an 

 open pipe inserted at each extremity of the height of the rise ; advantage 

 has been taken of this circumstance to avoid door ways, as in the annexed 

 diagram. 



" 8th. A certain quantity of motion woidd be obtained by a single 

 horizontal pipe between any points except the bottoms of the vessels ; but 



