926 ) 
Indicating the self-inductance of the latter by 4, and of the former 
by Ly, Rosa ealeulates the correction 4, L in order to obtain /, 
from £,, so that: 
bypass Ee 
_ This correction 4,7 consists of times the difference of the self- 
inductance of one winding with square section from that of a winding 
on the cylinder, added to the sum of the differences of the mutual 
inductances of all the windings. The correction term A,/ is brought 
in the following form : 
AL = Aman (A + B) 
wherein 7 in the said number of windings —, « = the mean radius. 
t 
A is the part of the correction due to the difference in the self- 
inductance, and B the part due to the differences in the mutual 
inductances. 
t 
Rosa gives two tables, wherein A is given as a function of — 
a 
and / as a function of 7. 
The error in Rosa's method is concealed in this correction term J, 
Which, as I shall show in a subsequent communication, is not 
only a funetion of # but also sensibly depends on the value of 
so that for this term a table with double entrance would be 
> 
a 
necessary. 
: . . . . . t . 
I have computed for a few different values of — a table for the 
a 
term B, by means of which I am able — for these special values of 
f . . 
—~— to get an idea about the degree of accuracy that can be 
a 
obtained in calculating self-inductances by the formulae (14) and (15) 
Example 1. 
[= 50-cM. Ho SOM: (= 04 eM. m4 n= 
calculated: 
by formula (14) L, = 70.5976 millihenry 
5 fe (15) i= 10.5988 A 
, by accurate correction method /,, == 70.5992 - 
… by Rosa’s method 10.54 5 
» by formula (1) of ConeN Dod en 
For this example the correction term used by Rosa is: 5 == 0.3440 
whereas the above mentioned table gives: 5 = 0.3247. 
