44 DISPLACEMENT INTERFEROMETRY BY 



on t and the width of the band read off at once on the micrometer scale, though 

 it is probably as easy to read off both ends. In my earlier improvised appa- 

 ratus, single scale-parts (o.i mm.) only could be guaranteed; but with a per- 

 fected optical system there is no reason why tenths of scale-parts should not 

 be equally trustworthy. This makes a scale of 1,000 parts in the ocular. 



28. Observations. There is sometimes difficulty in finding the coarse 

 adjustment for resonance, unless the design of figure 42 is adopted. After 

 this the adjustment at m is easy. As a source of alternating current, I selected 

 a small induction coil with a mercury break (Kohlrausch's design) to facilitate 

 the initial tests. This was put in series with a rheostat (to 30,000 ohms), a 

 Siemens precision dynamometer reading to just within milliamperes, an 

 ordinary telephone to indicate the continuous action of the coil, and the 

 vibrator above described. 



The Siemens dynamometer was first standardized with a Clarke cell. 

 Accepting the effective current i as being 



where tp is the deflection on a scale at about i meter of distance, the constant 

 C= i.i2X iQ- 3 relative to amperes was found and the mean resistance of the 

 coils included about 250 ohms. Virtual currents of io~ 4 ampere would escape 

 detection. 



The coil was now started and measurements made simultaneously both 

 at the Siemens dynamometer and at the vibrating telephone (slit distance 

 a = 35 cm., ocular distance 6 = 75 cm.), with results an example of which is 

 given in the diagram (fig. 44), the virtual currents of the dynamometer (abscis- 

 sas, milliamperes) being compared with the width of the image band (in scale- 

 parts, o.o i cm.) seen in the vibration telescope. The relation of the two is as 

 closely linear as the observations warrant, seeing that the optical system 

 (spectacle lens) was imperfect and the currents of the induction coil essentially 

 not quite uniform. In fact, it is an encouraging feature that the relation is 

 so nearly linear, even within the limits of observation. To obtain these 

 different virtual currents, resistances from 10,000 to 2,000 ohms were put into 

 the circuit. The dynamometer showed deflections from 2 to 20 cm. on the 

 scale. For larger deflections the coil current would have been too irregular 

 for use. 



Figure 44 shows that the vibrator indicates about 10 scale-parts per milli- 

 ampere, that readings beyond 5 milliamperes would be possible, that the deflec- 

 tions are fairly proportional to the current, and that with a good optical system 

 virtual currents as small as io~ 5 ampere should have been perceptible. Judg- 

 ing from the inclosed telephone, I should estimate the currents of an ordinary 

 voice at about i scale-part. The apparatus is thus very well worth further 

 development and would be particularly useful where alternators with a definite 

 period are in question. 



Throughout these measurements a variety of interesting observations were 

 made. It is obvious that current must not be increased until, with the ap- 



