Dr. Blyth ow the Composition ofNarcotine. 365 



These widely-differing formulae, calculated from analyses 

 agreeing so nearly, made it particularly necessary to deter- 

 mine anew the atomic weight. I undertook the subject at the 

 request of Professor Liebig, to whose valuable assistance in 

 the further prosecution of my investigation I consider myself 

 much indebted. 



A. Composition of Narcotine. 



The determination of the platinum in the double salt of that 

 metal was the method adopted in my investigations. The first 

 attempts gave constantly varying results ; the per-centage of 

 platinum found being 15-441, 16-7, 14-303, 14-84, IG'l, 16-6. 

 The following per-centages of platinum obtained by my friend 

 Dr. A. W. Hofmann were kindly communicated to me, 

 16"63, 16'49, 16-77. From these numbers it appeared that 

 the atomic weight of narcotine could not be deduced accu- 

 rately from its platinum double salt. Upon carefully repeat- 

 ing the experiments, it was found that this difference arose 

 from two sources of error. The first is, that the platinum 

 double salt is decomposed if submitted to washing for too 

 long a time. A portion of the double salt, already freed from 

 the excess of the bichloride of platinum employed to precipi- 

 tate it, was thrown on a filter and submitted to further wash- 

 ing. The fluid passed through of a pale greenish colour, 

 from which, on adding a solution of hydrochlorate of narco- 

 tine, a precipitate was obtained. 



I. 0-571 grm. of this precipitate, carefully dried, gave on 

 combustion 0-09 metallic platinum = 15-76 per cent. 



II. 0*996 of the salt still remaining on the filter gave pla- 

 tinum 0-1538 = 15-441 per cent. 



The second source of error lay in the action of the bichlo- 

 ride of platinum under the influence of a temperature much 

 below 1 00° C. For the purpose of causing the precipitate of the 

 double salt to settle more quickly in an agglomerated form, it 

 was placed in a water- bath and heated much below the boiling 

 point. The salt speedily collected at the bottom of the vessel 

 in a half melted state, and the supernatant fluid became of an 

 orange colour, from which, on filtering and cooling, large dark 

 red crystals mingled with long white needles were deposited. 

 The same fact was observed on boiling a portion ofplatinchlo- 

 ride of narcotine in a larger quantity of water, with the view 

 of getting it for analysis in a more distinctly crystalline form 

 from its watery solution. On cooling the greater portion again 

 precipitated. The mother-liquor, also of an orange colour, 

 upon evaporation gave the same red crystals mingled with the 

 white. From a portion of the former collected and burned, 



