Manchester Metnoirs, Vol. xliii. ( 1 899), No. 5. 3 



gravity through the oil-hole, and as this was inconvenient, 

 tending to empty the bath, he drove a plug of wood into 

 the hole and tried again, when to his still greater surprise 

 he found that the oil forced out the wooden plug. This 

 led him to fit a pressure gauge to the hole ; this imme- 

 diately rose to the top of its scale, 200 lbs. per square inch. 

 Then, realising that he had before him evidence of an 

 action in lubrication until then unsuspected, Mr. Tower 

 turned his attention to its experimental investigation, 

 finding that when the journal was run at 400 revolutions 

 a minute, the pressure on the square inch indicated on the 

 gauge was somewhere about 3/2 of the pressure necessary, 

 if distributed over the whole horizontal area of the section 

 of the bearing, to sustain the load. The pressure in the 

 oil-hole would be 600 lbs. per square inch when the total 

 load was 9,600 lbs., whence, as the area of the horizontal 

 section was 24 square inches, the mean intensity of 

 pressure v/ould be 400 lbs. This, however, was only when 

 the speed of the journal was greater than a certain limit 

 depending on the load ; when the speed diminished below 

 this limit, the pressure on the gauge fell to any degree 

 below that necessary to sustain the load. But this was 

 not all. When the speed was such as to sustain the load, 

 the friction was i in 400, but when running slow the 

 friction reached i in 3, or the journal seized the brass. 



Taking these two things together^ it made clear the 

 fact which had never been surmised before, that the action 

 of lubrication consisted in the actual separation of the solid 

 surfaces by a film of fluid of finite thickness. 



These discoveries of Mr. Tower excited great interest 

 at the time, and, being myself occupied in the study of 

 fluid motion, I was induced to undertake the theoretical 

 analysis of Mr. Tower's experimental results, from which, 

 after two years' work, I was able to publish a complete 



