Lemon. — On Duplex Telegraphy. 399 



place only 200 there, what would be the result ? Through our lowering the 

 potential at D, the current arriving at C, instead of passing to line, a portion 

 of it would rush through D. The effect of this would be similar to the effect 

 I described in figure 3, when two water-pipes have a connection made between 

 them from a higher to a lower level ; so, in like manner, a greater portion of 

 the current arriving at C would find its shortest road to earth to be through 

 C D, as the potential at the points C and T> would have been disturbed by- 

 altering the artificial resistance j thus the first step towards accomplishing 

 duplex would be defeated. 



It will be seen, on turning to the diagram, that between the back contact 

 of the key and the earth, 100 units resistance is inserted. This is placed there 

 to prevent the balance being disturbed when the key is at rest j for as the 

 battery in itself represents an internal resistance of 100 units, by inserting the 

 100 units resistance on the back connection of the key, it follows that, whether 

 the key is up or down, the same resistance is always in circuit. I may 

 remark, however, that there is an instant of time when the key is on the 

 hang — that is to say, when neither back nor front contact is made — when 

 both these resistances are cut out, but, as far as my experience has led me, 

 I have found no inconvenience from it. I have, however, in order to reduce 

 the interval between the two contacts, devised a key which practically removes 

 the difficulty, as near as it is possible to do so. This resistance in the back 

 connection of the key will vary as the resistance of the battery varies j for 

 instance, on a duplex circuit of 200 miles, by experiment, I find it requires 

 thirty cells to work it efficiently j and as thirty cells are eq^ial in their 

 internal resistance to about 300 units, it takes 300 units in the back connec- 

 tion of the key to preserve the required conditions. 



It will be observed that the line measures 776 units resistance, and that 

 the artificial resistance measures 520, which is equal to one-half the resistance 

 of the line, and one-half the joint resistance (268 units) at the B side. The 

 same applies to the A side when telegraphing from B. This joint resistance, 

 268 units, is the actual resistance that the current measures when arriving at 

 E or C, as the case may be, and is arrived at in the first instance by testing 

 v/ith a galvanometer at E or C, the circuit being through E H, E F direct to 

 earth, and from HF direct to earth, and from H to earth through the 

 100 units in back contact of key. It may appear strange that whilst the 

 resistances that the current arriving at E has in its path to travel should in 

 their sum represent 800 4-400 4-400-1-100 = 1,700 units, that the actual 

 resistance in circuit is only 268 units ; but it must be borne in mind that 

 instead of one path for the current there are no less than four for it to dis- 

 tribute itself over, consequently the resistance in the aggregate is lessened 

 thereby. 



