332 Prof. E. Taylor Jones on the Secondary 



Thus when Rj is small the parallel connexion gives a longer 

 make-spark than the series-parallel. The reverse is the case 

 when R x is large. With the coil used in these experiments 

 the effects are equal at the value R t = 25 ohms. 



It may be remarked here that for a given P.D. at the 

 primary terminals the spark-length at break is also greatest 

 when the primary layers are in parallel. This results, of 

 course, from the large value of the primary current. Con- 

 sidering, for example, the change from series to parallel, L^ 

 and R x are both reduced to one sixteenth, L 2 i to one-quarter, 

 of their " series " values. Thus the primary current at break 



E 



— given approximately by -^-(1 — e 5T ), where T = time of 



contact — is increased to sixteen times, and L 21 i to four times, 

 its original value. The coefficient of coupling being un- 

 changed, the primary capacity being supposed in each case 

 to be the optimum, and the effect of damping being regarded 

 as unaltered, the secondary potential is proportional to L 21 z , 

 and is therefore quadrupled. On the other hand, if the coil 

 is driven through a considerable series regulating resistance, 

 the spark-length at break is greatest when the layers are in 

 series. In this case the lowering of resistance due to the 

 change from series to parallel is of little consequence, and 

 the diminution of L21 reduces the secondary potential. 



The effect on the potential at make of varying the primary 

 self-inductance is further illustrated by the three oscillo- 

 graph photographs shown in figs. 1, 2, and 3 (Plate Y1L). 

 In obtaining these the coil was worked by a motor inter- 

 rupter, the oscillograph being connected directly to the 

 secondary terminals. Each photograph shows a " make " 

 followed by a " break," the movement being from right to 

 left. In each case the 98-volt battery was used, the total 

 primary resistance Ri being 46*7 ohms. This rather large 

 value of Rf was employed in order to reduce the secondary 

 potential at break. If Ri is increased the break-potential, 

 being proportional to the primary current immediately before 

 interruption, falls off more rapidly than does the make- 

 potential. The break-potential was further reduced by em- 

 ploying a large primary capacity — 10 microfarads, a value 

 far greater than the optimum. The interval between make 

 and break is in each photograph 0*00936 sec. The deflexion 

 from the zero line is proportional to the square of the 

 secondary potential. 



Fig. 1 shows the curve obtained when the primary coil 

 was connected in the series-parallel arrangement (cf. case III. 

 above). The potential at make is greater than that at 



