ON OUR KKOWLEDGE OF SPECTRUM ANALYSIS. 



311 



These groups are developed when examined either by photography > 

 or by fluorescence, with the aid of a spectroscope having lenses and 

 prisms entirely of quartz or of fluor-spa, and focussing at a metre ; the 

 order of brilliancy being 30, 32, 31. 



The following experiments demonstrate that the intensity of the 

 extreme rays is modified considerably by atmospheric absorption in the 

 manner indicated by the formula. 



"With a spectroscope (comprised of one prism and an object-glass) 

 focussing at six metres, the ray 32 was invisible, the feeblest ray, 31, being 

 still quite distinct. On adding a collimator, so as to reduce the distance 

 to l-50m., the ray 32 again became visible, notwithstanding the absorption 

 due to the additional object-glass. 



To complete the demonstration, M. Cornu had a spectroscope con- 

 structed focussing at 0-25m. With this apparatus the ray 32 had a 

 superior intensity to ray 30 ; so that the order became 32, 30, 31. Still 

 not satisfied, he arranged, between the collimator and the prism of 

 the spectroscope, a tube four metres long, closed at its two extremities 

 by flnor spa. When the tube is full of air no trace of 32 is visible, but 

 if a vacuum be gradually made, 31 gains in intensity, 32 soon appears, 

 and finally surpasses in intensity 31. As the air is readmitted the phe- 

 nomena repeat themselves in inverse order.^ 



In order to put this property of the atmosphere to the test of further 

 direct experiment, M. Cornu carried out a series of experiments of great 

 interest in the Alps. The following are the results of these experiments, 

 which were made under particularly favourable conditions : — 



The most remarkable series is that of July 25 at Riffelberg. The 

 curve traced by taking for ordinates the logarithm of the sine of the true 

 height of the sun, and as abscissee the wave-lengths for the limits ob- 

 served, is nearly an absolutely straight line. The deviation in the curve 

 for the Rigi corresponds to the time at which there was mist. Expressed 

 in wave-lengths, the extreme limits of the ultra-violet solar spectrum were 

 as follows : — 



' When wet-plates are used they should be washed with pure water after sen- 

 sitising, as nitrate of silver is very opaque to the extreme rays. 

 - Cmiiptes Rendus, t. Ixxxviii. p. 1289. 



