OP ARTS AND SCIENCES. 137 



of carbonic oxide and hydrochloric acid results, which shows that 

 the forrnylchloride, 



/H 

 OCT 



X C1 



first formed decomposes spontaneously. 



On treating aniline potassium, C 6 H 5 NK 2 , (which is obtained as 

 a dark brown hygroscopic powder, on adding potassium to an excess 

 of aniline, and then heating the mixture in an atmosphere of 

 hj'drogen, first in order to dissolve the potassium, and finally to 

 get rid of the aniline, to a temperature of 310°,) with 1 molecule 

 of chloroform, a violent reaction takes place in the cold, but the 

 resulting product is not pheuylirnidoforinylchloride, 



C 6 H 5 NK 2 + CHC1 3 = C 6 H 6 N=C**, 



but its decomposition products, phenylisocyanide and hydro- 

 chloric acid. The last named substance then unites with part of 

 the isonitrile set free to form the salt, 



2C 6 H 6 N=c]? 1 , HC1. 



The above observations are sufficient to clear up fully the re- 

 action which takes place on treating a primary amine with chloro- 

 form and caustic potash.* The amine P£NH 2 forms at first with 

 the caustic potash present a mono- and a di-potassium salt, RNHK 

 and EJSTKo. These salts then react with the chloroform, giving 

 rise to the amide and imidechlorides, 



^Cl 2 H 



RXH-CT and RN=C ( 



X H X C1 



which are then converted by the caustic potash present into potassic 

 chloride and alkylisocyanide. 



This noteworthy tendency to split off hydrogen and another 

 univalent group (here chlorine) from one and the same carbon 

 atom is not limited to the cases observed above. It is also, as is 

 known, a property of thioformanilide, f 



.SH 



c 6 h 5 n=c; 



* A. W. Hofmann, Ann. Chem. (Liebig), CXLIV. 116. 

 t A. W. Hofmann, Ber. d. chem. Ges., X. 1097. 



