ITS PRODUCTS OF DECOMPOSITION. 59 
REGNAULT’s own results; the highest amount obtained for the carbon being 0°63 
per cent. below the calculation, while the lowest differs by more than one per cent., 
and the mean of the whole four gives 0:77 too little carbon, involving a loss which 
could not possibly have occurred in carefully made analyses. 
Partly on account of this difference, and partly guided by his views regard- 
ing the divisibility of formule, GERHARD? was induced to doubt the exactitude of 
RecGNnavut’s formula, which presents three different deviations from his law ; the 
number of equivalents of carbon and of oxygen being uneven, and the sum of the 
equivalents of hydrogen and nitrogen also indivisible by two. He therefore 
repeated its analysis, using the crystallised codeine, and obtained the results 
contained in the table, and deduced from them the formula C,,H,, NO, for the 
anhydrous base, which gives the calculated results : 
Anhydrous. Crystallised. 


Carbon, . 5 4 72-24 68°13 
Hydrogen, ; P 7:02 7:25 
Nitrogen, . 3 é 4:68 4-41 
Oxygen, . 5 : 16:06 20°11 
100-00 100-00 
and tallies extremely well with his analysis. This formula has, however, been 
again called in question by Douurus,* who has endeavoured to determine the 
constitution and atomic weight of the alkaloids by the analysis of their hydro- 
sulphocyanates, and obtained from the codeine salt of that acid; results agree- 
ing with the formula C,,H,, NO,. Considering the known accuracy of Recnau.t, 
and of the chemists by whom his formula has been confirmed, I considered it an 
essential preliminary to my investigation to repeat its analysis with all possible 
care, so as to determine which of the two represents its true constitution. 
I. Preparation and Analysis of Codeine. 
I have little to add to the information we already possess regarding the pre- 
paration of codeine. I have obtained it, as usual, from the mother liquor from 
which morphia has been precipitated by ammonia. As the codeine forms only 
from a sixteenth to a thirtieth of the morphia, it is, of course, mixed in this fluid 
with a corresponding quantity of muriate of ammonia, which must be decomposed 
by potash, in order to obtain it. Much advantage is gained, however, by first. 
evaporating the fluid to crystallisation, and expressing the crystals deposited, as in 
this way the greater part of the muriate of ammonia, which is the more soluble 
salt of the two, is left in solution; and by repeating the crystallisation many 
* Annalen der Chimie und Pharmacie, vol. Ixv., p. 218. 
