620 TRANSACTIONS Of SECTION A. 



I must add a few words as to the screens employed. They have been mndi? 

 ill my laboratory with great care, and the light transmitted by them follows very 

 closely the colour sensation curves after white has been abstracted from them, 



I have with me a few negatives, which show the focus for parallel rays and 

 also for portraits. 



I do not know whether it is common knowledge that the best way of testing 

 equal sizes of image is to take a transparency from one image and superpose it 

 over the one to be tested. A want of equality of size is at once detected. 



2. On the Measitretnent of Larye Indiictanceit containing Iron. ^ 

 By Sir Oliver Lodge, F.B.S., and Benjamin Davies. 



AVe have bad occasion lately to measure large inductances, up to and above 

 loo henries, with core consisting of subdivided iron in a nearly closed circuit, as 

 used for certain telegraphic purposes ^\T[th very weak currents, .Since the in- 

 ductance may vary rapidly with strength of current, it is necessary in any measure- 

 ment to imitate the conditions of practice, and to determine the inductance as a 

 function of current under those conditions. To this end we have designed a 

 maximum-amplitude galvanometer, consisting of a well-damped coil moving- 

 (/end beat in a strong magnetic field, and attached to a mirror so that the ampli- 

 lude of its excursion can be observed. It can subsequently be calibrated by 

 means of a steady current giving the same deflexion. The inductance to be 

 measured is connected up in series with this galvanometer, aod with a specially 

 designed alternator of small power and known frequency /j/Stt, giving a sinuous or 

 simply harmonic current. A switch allows the inductance to be suddenly re- 

 placed by a non-inductive adjustable resistance R'; and when under these con- 

 ditions the same oscillation is produced in both cases, then the self-induction is 

 equal to that equivalent resistance divided by the frequency constant ; or h = Ft,' jp. 

 The strength of tbe current involved in this measurement is known by imitating 

 the deflexion with a known steady current ; and the main measurement consists 

 simply in observing the deflexion caused by the sine-alternalor, at .a measured 

 frequency, either with the inductance, or with the adjustable non-inductive 

 resistance, indiscriminately. It is to be understood that the ohmic resistance of 

 the wire wound on the self-induction is low. If not, a correction must be applied 

 for that, which is easily done, since \/(R- +^)'-L-') = R'. 



The following empirical expression is found able to give the self-induction of a 

 nearly closed magnetic circuit excited only by very weak currents, since for such 

 currents it is found to be practically constant. 



L = ^"" 



where n is the number of turns, G the width of the air-gap in millimetres, and 

 the other quantities are constants to be determined by experiment : though a will 

 naturally be nearly 2. In an actual case of a neai'ly closed magnetic circuit the 

 following values were found : — 



a = 1-99 ; _ i = 0-47; 



A- = 18 henries; and 5' = 0"1 millimetre. 



Hesults. 



The result of this method of measurement applied to Inductance colls of this 

 type shows (1) that measurements based on a determination of the square root of 

 mean square of current would serve fairly well for low magnetising forces ; (2) that 

 the self-inductance of such coils is for weak currents nearly independent of fre- 

 quency, or say for all frequencies up to about 20 per second, when the magnetising 

 force does not exceed 0'04 c, g.s. ; (3) that the self-inductance of a nearly closed 

 magnetic circuit is a definite and dependable function of the width of the air-gap 

 for moderate currents and frequencies. 



