On the Hexatomicity ofFerricum and Aluminium, 115 



lizes out; the crystals are washed with water and purified by 

 recrystallization. 



The new body can also be produced by the action of sodium- 

 amalgam on berberine ; but this method is not advantageous. 



The new body is a base ; it crystallizes in forms belonging to 

 the monoklinoedric system. Originally colourless, its solution 

 acquires by continuous boiling, or by standing in the air, an 

 orange-yellow tint. It forms with acids a series of well-crystal- 

 lizing salts, the analyses and detailed descriptions of which are 

 given by the authors. 



The composition of the base is C 20 H 21 NO 4 ; and as its forma- 

 tion from berberine is unattended by the production of any 

 other essential product, its generation must be very simple. 

 Hydrogen could only have acted as a reducing agent, or have 

 been simply assimilated. To express the formation of the new 

 substance, Hlasiwetz and Gilm have been compelled to assume 

 for the formula of berberine € 20 H 17 NO 4 , which agrees with the 

 published analyses, and which Pen-ins has also recently adopted*. 

 The formation of the new base, which the authors name hydro- 

 berberine, is then thus simply expressed : — 



€ 20 H 17 NO 4 + 4 H = C 20 H 21 NO 4 . 



Berberine. Hydroberberine. 



From hydroberberine, berberine may be regenerated by the 

 process by which Fritzsche succeeded in transforming harmaline 

 into harmine. Hydroberberine is dissolved in a warm mixture 

 of hydrochloric acid and alcohol, and some nitric acid gradually 

 added ; the liquid acquires the colour of the berberine solution, 

 and, on warming, red vapours are disengaged. On cooling, 

 crystals are deposited, which are hydrochlorate of berberine. 



XXII. On the Hexatomicity ofFerricum and Aluminium. 

 % William Odling, M.B., F.R.S., Sec. Chem. Soc.f 



UNTIL lately the chlorides of ferricum and aluminium were 

 ordinarily represented as sesquichlorides by the formulae 

 Fe 2 Cl 3 andAPCl 3 respectively. Doubling the atomic weights 

 of the constituent metals, as advocated by Messrs. Cannizzaro 

 and Wurtz and now generally admitted to be necessary, these 

 formulae become of course Fe CI 3 and Al CI 3 . But there are 

 undoubtedly very strong reasons for believing that both for- 

 mulae express only half molecules of the respective compounds, 

 and that the entire molecules ought to be represented by the 

 formulae Fe 2 CI 6 ' or Ffe^ CI 6 ', and Al 2 CI 6 or All vi CI 6 ', in which the 



* Journal of the Chemical Society, vol. xv. p. 339. 

 t Communicated by the Author. 

 12 



