September 14,1872.] THE PHARMACEUTICAL JOURNAL AND TRANSACTIONS. 
203 
ILS was given off in sufficient quantity readily to react 
upon lead paper, and to discolour metallic solutions. 
A mixture composed of dilute acid and this glycerine 
and water became highly offensive within a couple of 
hours at the ordinary temperature. As glycerine and 
dilute acid frequently enter into the composition of mix¬ 
tures and lotions, the occurrence of sulphur in the 
former is a possibility which dispensers should bear in 
mind. Its presence may readily be detected as above 
indicated. 
Bismuth .—The common impurities of metallic bismuth 
are arsenic, antimony, copper, and lead. I have very 
rarely met with lead, but my experience is perhaps 
somewhat limited. The amount of arsenic and antimony 
is generally small, and will not require here any 
especial attention, since the greater portion (if not quite 
all) of these two metals is eliminated during the subse¬ 
quent treatment for the removal of copper, which latter 
is the most difficult metal to get rid of. 
In the purification of commercial bismuth, instead of 
following the B. P. method of fusing with nitre, I have 
adopted the process of Hugo Tamm,* which has proved 
in my hands exceedingly efficient and satisfactory. The 
treatment consists simply in fusing the coppery bismuth 
with potassium sulphocyanide. Tamm says, “ The 
sulphocyanide which I use is prepared by mixing eight 
parts of cyanide of potassium and three parts of flowers 
of sulphur. One part of this mixture is thrown over 
sixteen parts of the metal melted at a low temperature.’ 
A bright red heat is sufficient, such as may be readily 
obtained by almost any Bunsen burner. 
In order to satisfy myself of the thorough elimination 
of the copper by this process, a sample of coppery metal 
and the resulting purifled button were carefully examined 
as follows : 
3*25S5 grains of the impure bismuth were dissolved in 
dilute nitric acid, ammonium chloride added, and the bis¬ 
muth precipitated as oxychloride by the addition of a large 
volume of water, excess of ammonia was now added, the 
precipitate thrown in a filter, well washed, and the 
acidified filtrate divided into two equal portions, a 13. 
Through a, IUS was passed to excess, the precipitate 
so formed collected, washed, and ignited. The residue 
thus obtained is a mixture of copper sulphide, oxide, and 
sulphate. To remove the latter, it was again carefully 
ignited with ammonium carbonate, and the resulting 
mixture of copper sulphide and oxide weighed. The 
weight = -0115 grains, this x 2 =-0230 grains in quan¬ 
tity taken (3*2585 grains) and *0230 gr. = *01836 grain 
copper=*563 per cent, copper in sample examined. As 
the percentage of copper in copper sulphide and oxide is 
identical, a mixture of the two is of no consequence in 
estimating the amount of metallic copper. 
The portion (3 of the filtrate was treated in a similar 
manner, by passing H 2 S through, but the resulting pre¬ 
cipitate was dissolved in nitrohydrochloric acid, eva¬ 
porated to dryness, and the residue dissolved in dilute 
hydrochloric acid. From this solution the copper was 
pracipitated as metal by zinc in a platinum dish in the 
usual way. The weight of metal obtained was equal 
to *558 per cent, of quantity taken. The mean of the 
two experiments = *563-}-*558-7- 2 = *560 per cent. 
48*45 grains of impure metal were then fused with a 
mixture of 2*5 grains pure K Cy, and 1*0 grain S for 
fifteen minutes. The fire was then removed, and after a 
few moments, to allow the slag to separate, the crucible 
being gently tapped to agglomerate the metal, the latter 
was poured out, and the button when cold weighed : 
Weight of button = 44*25 grains, the loss was there¬ 
fore 4*20 grains = 8*6. per cent. 
This loss is, however, much too great, the fact being 
that I did not succeed in obtaining the whole of the metal 
in the button, brilliant metallic specks being visible after 
cooling, disseminated through the slag. A second fusion 
of the latter yielded a small button, but on examination 
it was found to be contaminated with particles of slag, 
and consequently gave evidence of the presence of 
copper. 
Subsequent operations have proved that the loss 
should never exceed five per cent. 
I must here remark, that the proportions laid down by 
Tamm work well if pure cyanide be used, but if ordinary 
commercial fused cyanide be employed, then a larger pro¬ 
portion is necessary. Practically, I have found that 
50 parts of impure metal required from 3*5 to 4 of 
cyanide with 1 of sulphur. If a deficiency of cyanide be 
used, sulphide of bismuth is formed, thus involving a 
second fusion, or entailing a serious percentage loss of 
metal in the button, but this need not happen if the 
above precautions be observed. 
A weighed portion of the button first obtained was dis¬ 
solved, and after the separation of Bi. as before, treated 
with H 2 S, but not a trace of Cu. was revealed. Excess 
of ammonia to the acid solution did not produce the 
slightest coloration, nor did any of the usual specific 
tests for Cu. indicate the slightest trace of that metal. 
The coppery bismuth had, therefore, been completely 
freed from the Cu. by the very simple process of Tamm. 
Since writing the above, I have met with a paper by^ 
Mr. Schacht,* wherein he gives the results of some, of 
his experiments in removing Cu. from Bi. by fusion with 
nitre. It is stated that the whole of the Cu. cannot be 
removed by the B. P. process. This statement is doubt¬ 
less correct, repeated fusions are certainly necessary. 
The process of Tamm, so far as my experience goes, is 
unquestionably by far the more perfect, and I think 
leaves nothing to be desired, A little sulphur may be 
left with the metal, but this is practically of no conse¬ 
quence. With regard to percentage loss, Mr. Schacht 
states “thatthe loss varies with the duration of the pro¬ 
cess from 7 to 17 per cent.” In his experiment, 1000 
grains of coppery bismuth, after three fusions with nitre, 
lost 170 grains, and “ still yielded abundant evidence of 
copper.” In this respect, therefore, as well as in the 
more thorough efficiency, the process of Tamm may well 
supersede that of the present Pharmacopoeia. 
NEW DERIVATIVES FROM MORPHINE AND 
CODEINE, f 
BY C. It. A. WRIGHT, D.SC. BOND. 
Lecturer on Chemistry in St. Mary’s Hospital 
Medical School. 
During the past year further experiments have been 
made on the derivatives of these two alkaloids, in con¬ 
tinuation of the results briefly described at the last meet¬ 
ing of the Conference. The principal results are as 
follows 
It was shown previously that compounds are obtainable 
from codeine (by the action of hydrobromic acid) which 
may be regarded as formed by a polymerizing action, the 
resulting products containing four times as much carbon 
in their formulae as the original base; . and that analo¬ 
gous substances are formed by the action of hydriodic 
acid in presence of phosphorus, hydrogen being also 
added on in this case. Further examination has confirmed 
these results in the main, with this difference, however, 
that the action of hydrochloric acid on codeine and 
morphine appears to indicate that the formula) of these 
bases are double of those usually ascribed to them , 
while polymerides exist containing respectively twice, 
three times, and four times as many atoms in their 
formulae as the original bases; so that the following 
series may be written:— 
^Pharmaceutical Journal, April, 1868. 
t Read at the Brighton meeting of the British Pharma¬ 
ceutical Conference, August 4, 1872. 
* ‘ Chemical News,’ vol. xxv. p. 100. 
