I-it'cersaLs in Spectrum Pliotoiiraphs. 581 



Tro\vbriclo;e's photographs of spectra of single sparks, the 

 shock was the ahiiost instantaneous exposure to a bright-line 

 spectrum of exceeding brief duration, followed by an exposure 

 to a superimposed continuous spectrum of longer duration, 

 which may have been due to incandescence of the inner wall 

 of the capillary tube, or to phosphorescence of the gas. By 

 employing a tube with a bore of about '25 mm. I have obtained 

 reversed lines in the blue with single discharges of a medium 

 sized induction-coil and condenser. 



The fact that the faint continuous spectrum of the spark 

 is uniformly reversed shows that there is no selective reversi- 

 bility so far as the initial light-shock is concerned. 



I next endeavoured to determine whether the wave-length 

 of the foggiug-light had anything to do with the matter (in 

 which case we should expect reversals in some parts of the 

 spectrum and not in others), in the particular case where the 

 fogging illumination was spread out into a spectrum as in 

 Professor Trowbridge's photographs. Having already found 

 that foo-aino- the plate with X-ravs never oave reversals of 

 spark-images, it occurred to me that possibly ultra-violet 

 light might act in a similar manner. Having impressed a 

 number of spark-images on a plate^ it was illuminated with 

 lio'ht of wave-leno-th in the neio-libourhood of X = 23, from a 

 discharge between cadmium electrodes, a screen provided with 

 a slit being placed in the focal plane of the quartz spectro- 

 graph, with the plate a short distance behind it. The spark- 

 images were not reversed, and it appeared at first sight as if 

 ultra-\dolet light of this wave-length acted like the X-rays. 

 I was, however, not willing to accept this conclusion without 

 further studv, since the foo'D'ing; illumination in this case con- 

 sists in reality of a number of feeble light-shocks, that is, it 

 is of much briefer duration than in the case of candle-light. 

 In the first experiment the fogging illumination by ultra- 

 violet light was produced by the passage of perhaps a score 

 of sparks before the slit of the spectrograph. To get a 

 feebler illumination of longer duration I moved the next 

 plate to a distance of about two metres from the screen and 

 let the coil run for about a minute. The room was absolutely 

 dark, the spark-terminals and the front of the spectrograph 

 being covered with heavy black cloth^ so that the only light 

 that reached the plate was of the wave-length above men- 

 tioned. On this plate the spark-images which had been 

 previously impressed were strongly reversed, showing that 

 the time factor comes in the foo^oinor-lio-ht as well as in the 

 light-shock, and that ultra-violet light is as efficient as any 

 other, if it is not of too brief duration. This appears to 



FhiL Mag. S. 6. Vol. 6. No. 35. Noi\ 1903. 2 Q 



