188 Prof. Mitscherlich on the Spectra of Compounds 



logous deportment with the first at high temperatures, the 

 opinion might be expressed that ordinary iodine is a compound 

 body. 



From this it would follow that iodine at ordinary temperatures, 

 and iodine at the temperature of hydrogen flame, must be con- 

 ceived as two different compounds, because the spectrum of 

 iodine formed at ordinary temperatures is different from that pro- 

 duced by the hydrogen ilame. A compound of hydrogen with 

 iodine in this flame cannot be the cause of it, because the same 

 spectrum is obtained when iodine is in a carbonic-oxide flame ; 

 from the ready decomposability of oxygen compounds of iodine, 

 the presence of one of these cannot well be the cause. 



Spectra resembling those of iodine I could not detect by the 

 flame in the case of chlorine and bromine. If the preceding- 

 supposition is correct, bromine must be regarded as a compound 

 body ; as it has two spectra, one formed by absorption, and the 

 other by the electrical spark. 



If we compare with the flame spectrum of iodine the spectra 

 of the metalloids as formed by the flame of selenium, tellurium, 

 phosphorus, and those of sulphur and nitrogen, resulting from 

 feeble electrical charges, it is found that all these metalloids have, 

 in their spectra, the character of this iodine spectrum, and 

 would thus, if the above-expressed supposition be confirmed, be 

 compound bodies. 



In comparing these spectra, several peculiarities are observed : 

 thus in the distinctly marked part of the sulphur and selenium 

 spectra, the number of luminosities appears to be inversely as the 

 atomic weights ; a similar relation appears to obtain to that which 

 exists between the spectra of the haloid compounds of barium. 

 I must, however, add here that the spectrum of sulphur, like 

 that of tellurium, could not by the methods which I have hitherto 

 used be obtained with the distinctiveness which the spectra of 

 other bodies exhibited. 



These communications on the connexion of the spectra with 

 one another and with the atomic weights can only be regarded 

 as precursory. I shall continue the investigations with more 

 accurate apparatus, and in due time make further communications. 



Appendix. 



In conclusion I will adduce some observations which refer to 

 the flames giving spectra. 



The flames giving spectra are in most cases produced by the 

 luminous gases of the volatilized bodies, — thus, for instance, in 

 burning cyanogen, and in the metals and metallic compounds 

 which are brought into the flame. 



