98 Prof. A. F. Sundell's Researches 



units of temperature ; therefore the absolute unit of tempera- 

 ture is equivalent to 



tjtk x 10 16 Centigrade degrees, 



= about 10 18 Centigrade degrees. 



These figures are only intended to convey a very rough 

 notion of the relation. They cannot of course be considered 

 as even approximately correct, owing to the great uncertainty 

 as to the mass of molecules. 



Having seen that temperature is a quantity of the same 

 dimensions as energy, and knowing that the same is also true 

 of heat, it follows that entropy, whose dimensions are heat -s- 

 temperature, is a purely numerical quantity ; and the unit of 

 entropy is therefore independent of all other physical units. 

 In fact, the entropy of a perfect gas increases by unity, when 

 (without altering in temperature) it receives by conduction a 

 quantity of heat equal to the mean energy of one of its molecules. 

 This is seen by putting 



ir- 1 ? •••ah=@=e, 



where © is the absolute (C.G.S.) temperature. 



XI. Researches on Spectrum Analysis. 

 By Prof. A. F. Sundell*. 



CERTAIN natural phenomena,, such as the aurora borealis, 

 zodiacal light, and solar corona, have occasioned nu- 

 merous attempts to obtain the spectra of gases in a highly 

 rarefied condition and at a low temperature. Under these 

 conditions spectra generally become very feeble, and therefore 

 difficult to observe. The following experiments show that 

 tolerably bright spectra may be obtained by an advantageous 

 employment of means already known. 



Since the brightness of the spectrum depends, in the first 

 place, upon the thickness of the radiating layer, " end-on " 

 tubesj are employed by preference. I have employed tubes 

 as long as possible (up to 1^ metre long). The end towards 

 the spectroscope was simply melted together, and rounded 

 as well as might be in the process. The other end was drawn 

 out and melted on to the tubes leading to the mercury-pump. 



* Translated from a separate impression from the Acta Societatis 

 Scientiarum Fennicce, vol. xv., communicated by the Author. 



f Such tubes have been employed by Prof. P. Smyth for observations 

 on gaseous spectra in vacuo (fieibl. vii. 1883, p. 286). 



