223 



K('i,';ii(lii!.i,' this matter of calibration, existing hand bdoks, hii)orator.v 

 text books, and maker's catalogs are not clear, and the reader, whe'.her he 

 be a student, an inexperienced instructor, or a practical man can be, and to 

 to the writers knowledge often is, misled. In discussing the use of the 

 l)a]listic galvanometer in iron testing for example, the statement in va- 

 rious sources which should be authoritative is not infrequently seen, that 

 the galvanometer may be calibrated with a standard cell and condenser, 

 and students have been known to follow these directions, without couuse! 

 fi-om the instructor, although the condenser was introducing perhaps 4,000 

 megohms in series with the galvanometer, while the resistance of the sec- 

 oiidary circuit otherwise used was leSvS than 100 ohms. This procedure may 

 give rise to errors of several hundred per cent, with corresponding influence 

 on the values for the B-H curve. 



IJecognizing that this problem is satisfactorily treated in much of the 

 existing literature, it must also be admitted that many of the current helps, 

 to which one first turns for reference, are quite inadequate and mislead- 

 ing, and it is the purpose of this article to offer a wider discussion of the 

 facts. A single example with calibration curves of a Leeds and Northrup 

 type H galvanometer will serve to illustrate the principle. In figure 1, 

 curve D gives the relation of charge to deflection for the case of calibra- 

 tion with a standard condenser. For the same galvanometer. A, B and C 

 •Ave the corresponding calibration curves when the total circuit resistances 

 are i-es|iectively 486, SSO, and 1,486 ohms. These curves show clearly tht* 

 influence of diminishing total circuit resistance upon the value of the gal- 

 vanometer constant. Curve D shows 8.2 scale divisions for 1 micro-coulomb, 

 while curve A, for a circuit resistance of 486 ohms, shows 1.4 scale di- 

 visions for the same charge. 



Curves A, B, C, and D were taken with the small rectangular damp- 

 ing coil removed. A similar set of curves, Ai. Bj. Ci. and Dj give the call 

 bration values after the damping coil has been removed. 



Various suggestions have been made for calculating the true value ol 

 the ballistic constant for any given condition from the known constants of 

 the galvanometer such as period, moment of inertia, moment of torsion, 

 strength of field, etc. These methods, entirely adequate theoretically, are 

 nevertheless ditticulc to apply practically, because the values of the con- 

 stants are seldom known with sufficient precision, and are themselves liable 

 io change when the galvanometer is readjusted. 



