CHAPTER XI. 



INTERFEROMETER OBSERVATIONS AND ACHROMATIC FRINGES 

 IN CONNECTION WITH THE HORIZONTAL PENDULUM. 



86. Apparatus and data. The observations begun in my last report * were 

 continued during the remaining summer months, and the graphs obtained 

 may conveniently be considered here, particularly as they bring out more 

 clearly than has hitherto been the case the effect of the temperature of the 

 laboratory on the horizontal pendulum. Figures 57, 58, 59 of the apparatus 

 of the last report should be consulted. 



FIG. 113. 



In the graph, figure 113, the initially lower curve shows the change of incli- 

 nation of the pier in seconds of arc, the upper curve the corresponding tem- 

 perature in degrees centigrade of the basement laboratory room. Observa- 

 tions were usually made at 10 a. m. and 6 p. m. Between August i and 21 

 the resemblance of the two curves is marked. After that the gross resemblance 

 is no longer so striking, but it nevertheless remains in the details. There is a 

 lack of quantitative equivalence only. 



Toward September 8 the two curves cross each other, i.e., the relatively 

 enormous fall in temperature due to the cold season has had no correspond- 

 ing effect on the pendulum. This is very curious, for thereafter the detailed 

 resemblance of the two curves (the temperature graph being below the incli- 

 nation graph) is in fact astonishing. The large fall of temperature toward 

 September 21, however, again fails of similarly marked expression in the 

 inclination graph. The alternations of temperature are thus sharply indicated 

 in both curves, while large changes in one direction produce less pronounced 

 resemblances. This behavior is very much like the backlash of a screw, but 

 it is exceedingly difficult to surmise how such a discrepancy can occur when an 

 ocular plate scale, only, was used for measurement. The effect of rain, apart 

 from temperature, appeared so inconsistent that it is hardly worth considering. 



The question is, therefore, where this temperature discrepancy has its seat. 

 It can hardly be in the interferometer, where the parts are of the same metal, 



Carnegie Inst. Wash. Pub. No. 249, part in, 47-51* 1919. 



114 



