70 Professor TT. E. Aijrton [March 24, 



Consequently Professor Perry and myself have for some time past 

 been working out practical means for overcoming these difficulties, 

 and we have arrived at what we hope is an extremely satisfactory 

 solution. Instead of supplying electricity to one very long, not very 

 well insulated rail, we lay by the side of our railway line a well- 

 insulated c^ble, which conveys the main current. The rail, which is 

 rubbed by the moving train, and which supplies it with electric 

 energy, we subdivide into a number of sections, each fairly well in- 

 sulated from its neighbour and from the ground ; and we arrange that 

 at anv moment onlv that section or sections, which is in the immediate 

 neighbourhood of the train, is connected with the main cable ; the 

 connection being of course made automatically with the moving train. 

 As then leakage to the earth of the strong propelling electric current 

 can only take place from that section or sections of the rail, which is 

 in the immediate neighbourhood of the train, the loss of power 

 by leakage is very much less than in the case of a single imper- 

 fectly insulated rail such as has been hitherto employed, and which 

 being of great length, with its correspondingly large number of points 

 of support, would offer endless points of escape to the motive current. 

 Dr. Siemens has experimentally demonstrated that an electric 

 railway can be used for a mile or two ; Professor Perry and myself, 

 by keeping in mind the two essentials of success, viz. attention to 

 both the mechanical and electrical details, have, we venture to think, 

 devised means for reducing the leakage on the longest railway to less 

 than what it would be on the shortest. 



For the purpose of automatically making connection between the 

 main well-insulated cable and the rubbed rail in the neighbourhood 

 of the moving train we have devised various means, one of which is 

 seen from the followinoj ficjure. 



A B is a copper or other metallic rod resting on the top of and 

 fastened to a corrugated tempered steel disc D D (of the nature of, but 

 of course immensely stronger than the corrugated toj) of the vacuum 

 box of an aneroid barometer), and which is carried by and fastened to a 

 thick rine^ E E made of ebonite or other insulatinfj material. The 

 ebonite ring is itself screwed to the circular cast-iron box, which latter 

 is fastened to the ordinary railway sleepers. The auxiliary rail A B 

 and the corrugated steel discs D D have sufficient flexibility that two 

 or more of the latter are simultaneously depressed by an insulated 

 collecting brush or roller carried by one or by all of the carriages. 

 Depressing any of the corrugated steel discs brings the stud F, which 

 is electrically connected with the rod A B, into contact with the 

 stud G electrically connected with the well-insulated cable. 



As only a short piece of the auxiliary rail A B is at any moment 

 in connection with tbe main cable, the insulation of the ebonite ring 

 E E will be sufficient even in wet weather, and the cast-iron box is 

 sufficiently high that the flooding of the line or the deposit of 

 snow does not affect the insulation. The insulation, however, of G, 

 which is permanently in connection with the main cable, must be far 



