THE AID OF THE ACHROMATIC FRINGES. 59 



with fixed objective and with small resistance (200 ohms) in circuit. After 

 this a large resistance (10,000 ohms and more) may be inserted and the work 

 continued with the interference fringes when further adjustment is possible. 

 The vibrator cc' is more liable to become unstable and stick to the magnets 

 of the telephone. But the advantages gained are manifest, as the sensitiveness 

 will at once have increased 50 times or more. 



In a later design of apparatus, the bar v, rigidly fixed at one end, was con- 

 trolled at the other by the screw / (fine-threaded) pushing against the rigid 

 socket s. This makes an excellent fine adjustment for resonance and does 

 not disturb the interferences when used. The fine adjustment may therefore 

 be made with the interference fringes at once. 



43. Observations with the new apparatus. The first trials were made with 

 the white slit-image. Using the coil with a virtual electromotive force of 

 about 0.7 volt and a small resistance (200 ohms), the resonance conditions 

 are easily reached in a well-braced apparatus. The resistance was then in- 

 creased to 10,000 ohms and the corresponding band-width found. This is 

 more difficult, for the maximum deflection is painfully sensitive to slight 

 differences of tension or to slight changes in the position of the telephones. 

 When obtained, however, the band-width 5 persists. The sensitiveness ob- 

 tained is thus, the observed s being 18 scale-parts, o.y/iSX io 4 = 4Xio~ 6 

 ampere per scale-part, or 4Xio~ 4 per ocular centimeter. Hence, per AAT = 

 io~ 4 cm. of the micrometer of the interferometer, where b6 = AN cos i and = 

 s / 2 f (/= 23 cm. being the focal length of the telescope and s = io~ 2 cm. the value 

 of i scale-part, b= 10 cm. the breadth of the ray parallelogram, 2 = 45), the 

 current would be estimated as 



4Xio- 4 Xs=(4Xio- 4 2/A7Vcos*)/6 = I -3Xio- 7 ampere. 



A fringe would usually represent much less than this. 



In the interferometer experiments now begun with this adjustment the 

 fringes were found much more easily than was anticipated. The slit-images 

 soon became adequately quiet. There will be considerable difficulty, however, 

 in testing the degree of resonance by means of the fringes alone, unless the 

 slide z of the tension-screw y is very true. Any rotation of y around its axis 

 will send the fringes out of the field. Adjustment may sometimes be made at 

 the telephones as specified, or the wire may be stretched by lever mechanism. 

 In the present experiments I made the resonance adjustment with the slit- 

 image before using the fringes. Later the device v, t, s, figure 61 was adopted. 



The 5-coil of mean voltage 0.7 volt was tried first. But this was rather 

 too strong for these measurements, so that coil No. 10, with about 90 turns 

 giving o.oo i volt per turn was substituted. With the slit-image alone the 

 sensitiveness was 4.5 X io~ 6 ampere per scale-part; with the sliding micrometer 

 of the interferometer, AAf = io~ 4 cm. was equivalent to 8Xio~ 7 ampere; with 

 the ocular micrometer I obtained io~ 7 ampere per scale-part. This is some- 

 what less than what was estimated; but a shortage is here inevitable. The 

 fringes used were fairly large, say 2 scale-parts, but completely under control. 



