NOTES ON STATIONARY ENGINES. 35 



much more, whioli is chiefly due to the bettor draught avail- 

 able ashore. Chimneys ashore are usually much taller than 

 the funnels of steamboats, and of a substance which does 

 not so readily allow the cooling of the contents ; consc- 

 ()ucntly the heated column on which the di'aught depends is 

 both hotter and longer, so as to be in a double sense more 

 offootivo. 



Some years ago it was comparatively seldom that a feed 

 pump was provided in a boiler house, the boiler feed de- 

 pending on the engine feed pump, or at most a rough sort 

 of donkey pump was provided for use when the main cnn-ine 

 was not at work. More recently, however, there are so 

 many donkey pumps in the market of good design, that it 

 appears to be an open question whether or not engine feed 

 pumps should be altogether abandoned. By this means one 

 great lisk is avoided, wliich is often a serious one in connec- 

 tion with an engine driven feed pump, this is that the feed 

 valve is sometimes screwed down when the pump is working, 

 when, if there is no overflow valve, or if this valve is out of 

 order, a burst pipe is one of the least serious things which 

 can hiippen. If a separate donkey-pump is adopted, this 

 risk is entirely avoided, as any appreciable obstruction causes 

 the complete stoppage of the pump. The stoker ought to be 

 able to !i,tfcond to a donkey-pump as well as to a feed valve, 

 so that no objection need be raised on this score. 



Thoro is one feature which does not receive anything like 

 the attention which it ought to receive in connection with 

 almost all pumps and water-pipes. It is in the fact that 

 unless the pipes, etc., are throughout exposed to a moderate 

 and fairly uniform pressure, all possibility of lodgment of 

 air in any part must be avoided. Almost all the mysterious 

 troubles encountered in the working of feed pumps may bo 

 traced to this one point ; there is some space where air may 



