406 Scientific Intelligence. 



meters of gas. The cost of the oil would be therefore in Paris 

 about a hundred times that of the electric light, or that of gas 

 fifty times, to produce the same light. The comparison with the 

 smaller machine woiild be less favorable. The carbons for the 

 larger light had a cross section of 81 mms., and the ordinary con- 

 sumption was a little over a centimeter in length per hour,— 

 Comptes Eendu9, Ixxxii, 299. e. c. p. 



_ "-" - oy increase of temperature on the Index of Refraction ; 

 m TT. T.r ^ Letter to the editors, dated 



-Dear Sirs : I have in progress 

 an investigation of the effect of increase of temperature on the 

 index of refraction, which has at this time yielded some results 

 of considerable importance to spectroscopists. In 1858 Messrs. 

 Gladstone and Dale announced as the conclusion of a research 

 upon this question, that in every substance the refractive in- 

 dex diminishes as the temperature increases. I am satisfied that 

 glass at least does not obey this law ; that, on the contrary, with 

 it the index increases with the temperature. In my experiments 

 I have used equilateral glass prisms with indices of refraction of 

 about 1-63. The change in the position of the D line has been 

 observed with a parallel wire micrometer. That the effect was 

 not due to a change in the angle of the prism during the process 

 of cooling I satisfied myself — both by measurement of the angle 

 when hot and when cold — and by receiving the image of the slit 

 of the collimator reflected from both faces upon the cross hairs 

 of two telescopes properly adjusted upon the instrument. No 

 appreciable change in angle could be discovered. Numerous ex- 

 periments agree well in fixing the " index of sensitiveness," but 

 the quantitative results I have not fully worked out. I only wish 

 at present to direct attention to the fact, that, in the use of a 

 train of several glass prisms in a spectroscope, ordinary changes 

 of temperature to which the instrument may be subjected will 

 produce a very noticeable change in the position of the spectrum 

 lines. In my own, of five large prisms, of an angle of 64°, the 

 change in the position of the D line on removing the prisms from 

 an open window, the temperature outside being about 32° F., 

 to the room, at ordinary temperature was as much as 95 divisions 

 of the micrometer screw head. With a smaller number of j)ri8m8 



the change was closely proportioned to that number. I wish 

 suggest that in this way may be found the cause of many discrep- 

 ancies which occur in tables of wave-lengths, as furnished by dif- 

 ferent workers, in such cases as those in which the dispersion spec- 

 trum has been made use of, and the wave-length computed by in- 

 terpolation. A great many such cases occur in Watt's Index oi 

 spectra. In an instrument of many prisms the observation oi 

 temperature will be a matter of vital importance in fixing the 

 exact position of a line. I propose to pursue the investigation, 

 especially in respect to the observation of lines more or less re- 

 frangible than the D, and also as to the effect of change of temper- 

 ature upon other than glass prisms. 



