10 Profs. J. Dewar and J. A. Fleming. On the Dielectric 



change of frequency of the electromotive force than is the case with 

 water. Ifc appears certain that as far as water at 0° C. is concerned, 

 the dielectric constant and the square of the electric refractive index 

 is a number not far from 80, for waves having wave-lengths between 

 8 mm. and infinity, or for electromotive force reversals having 

 -frequencies varying from 37*5 x 10 9 to zero. On the other hand, 

 the values found for ice at or a little below 0° C. seem to indicate a 

 dielectric constant of 78, when using very slow oscillations ; and a 

 value of about 2*0 when using oscillations having a frequency of 

 some millions per second. 



It is clear that in this matter there is still room for further investi- 

 gation. It is evident, since the optical refractive index of water is a 

 number lying between 1*3 and 1*4 for Avaves having a wave-length of 

 '0*00005 cm. or reversals having a frequency of —400 X 10 12 to 700 X 10 12 , 

 that water may be regarded as presenting the phenomenon of 

 anomalous dispersion beyond the range of the visible spectrum, 

 because the refractive index for waves of a length of 0*8 cm. and 

 upwards is a number not far removed from 8*9, and this number 

 is very much greater than that for wave-lengths of the order of 

 visible light. 



Within the octave of wave-lengths comprising visible light the 

 refractive index of water lies between 1*3 and 1*4. We know very 

 little about the refractive index of water for the fourteen octaves of 

 radiation lying beyond the extreme red end of the spectrum, but we 

 know that water has very considerable absorptive power for a large 

 range of this radiation. The next ten octaves beyond the last, include 

 the range of the Hertz radiation or of wave-lengths from ■?? to 500 cm. 

 in length, and for all this the refractive index of water is approxi- 

 mately 8*9. It remains to be seen how the high value is connected 

 with the low one, and whether this variation may be properly 

 regarded as a case of anomalous dispersion analogous to that found 

 in the case of an alcoholic solution of fuchsine within the range of 

 the visible spectrum. It is evident that since the dielectric constant 

 of any one substance, such as ice- water, is a function both of tem- 

 perature and time, it can best be represented geometrically by a 

 surface, which may be called the dielectric surface, and which is 

 defined by the co-ordinates representing dielectric constant, tempera- 

 ture, and frequency of electromotive force reversals. 



The details of two determinations of the dielectric constant of ice 

 at —185° C. are given in Table IY. 



The same readings were obtained both with 90,000 ohms and 

 1,000 ohms in the galvanometer circuit. 



The above figures of observation require two corrections to be 

 applied. In the first place, the pins which support the inner con- 

 denser plate, and which, pass through glass beads, have a total area 



