580 Profs E. Taylor Jones on 



with correction for permeability, was 1*132 amperes. The 

 observed least current giving a 1-cm. spark was 1*13 amperes, 

 but generally a little more current was required. In 

 Case III. the values were *815 ampere calculated, and 

 •86 ampere observed. The differ ence is much greater in 

 this case, but the correction for permeability is also greater 

 and probably not so reliable. 



Of the three cases Case I. gives the most definite value 

 for the least sparking current, and is the most suitable to 

 adopt as a standard, while the numbers in Table VII. show 

 that the potentials in the other two cases are fairly closely 

 proportional to the calculated values. The above results 

 show therefore that, so far as the 1-cm. spark potential may 

 be taken as a guide, the theoretical expression (1) leads to 

 approximately correct values (within about 2*5 per cent.) of 

 the maximum secondary potential over the range of current 

 and primary capacity covered in the present experiments. 



Further support of the theory is afforded by the similarity 

 in the forms of the calculated and experimental curves 

 (figs. 1, 2, 3, and Plate VIII. figs. 7, 8, 9), and by the agree- 

 ment between the observed and calculated values of the 

 frequency of the longer wave in each case. Thus the 

 calculated frequencies are (from Table V.) 194*3, 252*8, 

 304*4 : the observed values, obtained by measurement o" the 

 photographs, 193*9, 232*3, 301*6. 



Possibly still closer agreement between the calculated and 

 observed results would have been found if second-*rder 

 quantities (e. g. 6i, *9 2 2 ) had been retained throughout but 

 the additional labour involved in the calculations vould 

 make the method too cumbersome to be of much vilue. 

 What is more required at present is a simplified expr^sion 

 which will give approximately the maximum secoidary 

 potential, or an upper limit to its value which is not very 

 greatly in excess of the actual value. 



Such an upper limit may be arrived at by neglecting the 

 resistances of the circuits so that k lf k 2 , &i, and S 2 all become 

 zero. If we further assume that the maxima of the two 

 potential waves in the secondary occur simultaneously the 

 expression for the maximum potential becomes 



T . n y n 2 2 ±n^n 2 

 2ttL 2] i n ,_ n , 



i.e. fcrL,' n ^ 



'21*0 



n 2 — n x 

 This may also be expressed in terms of the ratio LoCgL^Ci. 



