34 



DISPLACEMENT INTERFEROMETRY BY 



thermal discrepancy largely vanishes. Figure 36 shows the relations of the 

 fringe displacement A<? in the ocular to the micrometer displacement AN". The 

 two curves were obtained successively and show slightly different rates io- 6 X 

 3.45 and lo^Xs-ss- The reason of this is to be sought in the change of fringe- 

 breadth if the rod appreciably expands. It also accounts for the difference of 

 contractions in figure 35, found later from those for the same field found earlier 

 in the series; for as the coil and rod inevitably increase in length during the 

 measurements, the fringe-constant of the ocular micrometer changes, and this 

 constant can be found only at the beginning or after the end of the measure- 

 ment when all temperatures are again constant. The work with Ae should 

 always be confined to the range of the ocular micrometer, which is here much 

 exceeded. 



100 



300 400 500 600 700 800 



20. Coefficient of expansion. To arrive at a definite reason for the occur- 

 rence of the anomalous contraction mentioned above, it seemed desirable to 

 modify the magnetic apparatus for the purpose of measuring the coefficient 

 of expansion of a given metal tube. This could easily be done by using the 

 coil merely as a heater. 



Accordingly a thin-walled brass tube was prepared, effectively 40 cm. long, 

 7 mm. in diameter, with walls but 0.3 mm. thick. One end of this is shown at 

 T, figure 37, inclosed by the water-jacket (now empty) w, of the coil CC. The 

 tube T is closed at this end by a brass plug ending in the stylus e (to be made 

 of invar), which actuates the plate of the contact lever, as above explained. 

 The other end of the tube T was strengthened by a cylindrical sleeve soldered 

 to it, the latter projecting beyond the coil CC and rigidly held by a clutch, 

 the attachment being quite independent of the supports of the coil CC, for 

 the reasons already indicated. To measure the temperature of the brass 

 tube 7, a long (60 cm.) normal thermometer (Baudin) t, graduated in tenth 

 degrees, capable of sliding easily from end to end of the tube, was used. The 

 stem projected considerably beyond the left end of T, figure 37, so that the 

 whole thread could be left within the tube except during the reading. 



The method of observation consisted in passing a current of a few amperes 

 through the coil for a short time, breaking it, and making the observations 



