828 



Prof. Carl Barns on the 



is never lost, but may always be brought back again to the 

 fiducial and coincident sodium lines by moving the micro- 

 meter, however remote the centre may be, a crystal (or 

 indeed a column of any thickness) may be placed in one of 

 the component beams, and the micrometer equivalent of the 

 difference resulting from presence and absence of the crystal 

 ascertained. This makes the method quite sensitive; for 

 the two micrometer readings for the presence of air and 

 crystal, respectively, may differ by many centimetres, if de- 

 sirable, and each reading may be to '00005 centimetre. 



The method, however, requires an accurately plane parallel 

 plate, however thick. If the faces are at even a slight angle, 

 the beam passing through them will be deflected by the 

 introduction of the crystal and will not be returned in its 

 own path. This requires a readjustment, and therefore a 

 correction for the displacement of the mirror difficult to 

 estimate. It is to be accentuated, however, that the method 

 for normal incidence admits of a plate of any thickness, 

 i. e., of a long column of glass with plane parallel end faces, 

 so that this method of measuring the index of refraction 

 must ultimately exceed in sensitiveness the method of the 

 spectrometer, while even the importance of the readjustment 

 in question becomes of less consequence. It will thus be 

 available for answering refined questions relative to this 

 index even in glass, under conditions where the spectrometer 

 fails of application. It is in this direction that the results 

 are to be prosecuted as soon as the preliminaries discussed in 

 the present paper have been clarified. 



It is also probable, that if the deviation of the ray mm due 

 to non-parallel faces in the crystal is corrected not at the 

 micrometer mirror M, but at the other opaque mirror N 

 beyond the grating, the discrepancy in question will vanish. 

 For in this case, on restoring the interferences by rotating N, 

 the air path of both component beams mm and nn is equally 

 incremented. 



Fig. 1. 



The plate crystal X may be mounted on a horizontal axle 

 ah (fig. 1) normal to the interfering beam mm in question. 



