June 28, 1873.] 
THE PHARMACEUTICAL JOURNAL AND TRANSACTIONS. 
1033 
FRANGULIN AND FRANGULIC ACID.* 
BY A. FAUST. 
The bark of the Rhamnus franguZa is digested for three 
days with 90 per cent, alcohol at 25°~30°; from the ex¬ 
tract the alcohol is partially distilled off, and then lead 
acetate is added to the residue as long as a precipitate 
falls : the filtrate from this is precipitated by basic lead 
acetate, and the precipitate disseminated through alcohol 
and decomposed by sulphuretted hydrogen ; the resulting- 
liquid yields crystals of frangulin on cooling after filtration 
while boiling hot. 
Casselmann describes frangulin as crystallizing under 
the microscope in quadratic tables ; the author could not 
discern these tables either in his own preparation or in a 
specimen prepared by Casselmann, the two samples melt¬ 
ing at 226 Q and 225° respectively, instead of 249° as 
stated by Casselmann. It forms a lemon-yellow crystal¬ 
line mass after recrystallization from hot alcohol, and is 
soluble in alkalies with a deep cherry-red colour; when 
boiled with hydrochloric acid it splits up into glucose and 
frangulic acid: 
C 20 H 20 Oio — Ci 4 H 8 0 4 + C 6 H 12 0 6 . 
Hence the author ascribes the formula C 20 H 20 O 10 to 
frangulin, instead of the sub-polymeric formula CgHyOg 
attributed to it by Casselmann. 
Frangulic acid is most conveniently prepared by boiling 
the bark with rain-water for an hour, then adding caustic 
soda equal to one-twentieth or one-thirtieth of the weight 
of bark taken, and boiling again for several hours ; the 
resulting liquid extract is supersaturated with hydrochloric 
acid, and boiled for some hours, whereupon the frangulic 
acid separates and can be collected on a cloth filter ; the 
dried and pulverized precipitate is boiled with alcohol, 
and the hot alcoholic solution precipitated by lead acetate 
(which scarcely precipitates frangulic acid); the filtrate is 
precipitated with basic lead acetate, and the precipitate 
disseminated through alcohol and decomposed by sul¬ 
phuretted hydrogen ; and the resulting liquid is heated to 
boiling and filtered hot, the acid then separating from the 
filtrate on cooling. This treatment with lead salts, etc., 
is repeated several times ; and the purified acid, after 
being dried and boiled with benzene to separate sulphur, 
is finally recrystallized several times from hot alcohol. 
Different samples of bark differ much in the quantity of 
frangulic acid obtainable from them; in one case, 50 lbs. 
yielded only a few grams. 
Frangulic acid melts at 252°~254°, and sublimes par¬ 
tially at lower temperatures ; it is but little soluble in hot 
water, and is insoluble in cold water and alum-solution ; 
alkalies dissolve it, with a red colour, removed on boiling 
with zinc-dust; on treating it with red-hot zinc-dust an¬ 
thracene is produced in small quantities ; it crystallizes 
with 14 proportions of water, the last gH 2 0 being only 
removed at 180°. 
THE MEANS OF DETECTING AND ESTIMATING 
BROMIDE IN IODIDE OF POTASSIUM.+ 
BY ALFRED E. TANNER. 
It occurred tome that the above subject would be well 
worthy attention at the present time, for not only are the 
processes for detecting and estimating bromide in iodide 
few and imperfect, but it also seems to me a very pro¬ 
bable adulterant, inasmuch as there is a great difference 
in price between these two salts, and the difficulties 
attending the detection of bromide when mixed with 
iodide are considerable. 
Cl, Br, and I, so resemble one another in their chemical 
characters and reactions, that it becomes a difficulty by 
* Ann. Chem. Pharm, clxv. 229; from the Journal of 
the Chemical Society. 
f Paper read before the Liverpool Chemists’Association, 
May 8th, 1873. 
no means easily surmounted to distinguish them when in 
the presence of one another, and this is especially the case 
wii/h the two latter, and as a sample of KI may contain 
75 or more per cent, of IvBr and y e t be indistinguishable 
from the pure article when tried by the pharmacopoeia 
tests, it needs little further to point out the desirability of 
investigating this subject ; and before I go further, I 
must confess that I fear I have accomplished little towards 
doing away with the difficulties. What we require is a 
test presenting no great difficulties of application by the 
ordinary pharmacist, and one which shall indicate with a 
fair degree of accuracy the object sought to be attained. 
Of course it is well known to chemists that PdCL, when 
added to a neutral solution of an iodide containing 
bromide, will remove the whole of the I without affecting 
the Br, but PdCl 2 is a rare and most expensive reagent to 
use, and would scarcely pay the pharmacist who examines 
usually but small parcels of iodide at a time. This, 
although I believe to be the most accurate, we must con¬ 
sider out of the question. Recent chemical works tell us 
that a mixture of FeS0 4 , two parts, and CuS0 4 , one part, 
added to a neutral solution of an iodide, in the presence of 
bromide and chloride, and the mixture neutralized with 
NH 3 , will remove the whole of the I without affect¬ 
ing the Br or Cl. In my hands, at least, the practice of 
this process has been attended with only partial success, 
for I have found it impossible to remove the whole of the 
I ; the difficulty therefore remains as great as ever ; it is 
probable, however, that further experiments with this test 
may yet prove it adequate to the purpose. I rather sus¬ 
pect the Cu 2 I 2 to be slightly soluble in the solution from 
which it is precipitated ; we must therefore seek some 
salt to add to the mixture to prevent this. The following 
is the reaction stated to occur :— 
2 KI-f-2 CuS 0 4 +2 FeS0 4 =Cu 2 I 2 +K 2 S0 4 +Fe 2 3S0 4 
A test proposed by Van Melekebeke {Journ. Pharm. 
d’Anvers, xxviii., 49, 1872) seemed to promise well, and 
from its great simplicity would have been a valuable one 
if successful. It depended on the fact that a saturated 
solution of one salt is capable of dissolving appreciable 
quantities of another salt. A saturated solution of KBr 
was therefore used, and to this the sample of KI, in 
powder, was added in small quantities at a time, when, if 
pure, it dissolved readily, but if KBr were present, the 
liquid being already saturated with this salt, it would 
remain undissolved. Repeated trials with this test have 
proved to me that it is quite useless. The author recom¬ 
mends you to take 10 c.c. of the saturated KBr 
solution, and to add to this 10 drops of distilled water ; 
1 gram, of the suspected salt, in powder, is then added, 
small portions at a time, which, if the iodide be pure, 
shoidd at once dissolve ; but 10 drops of distilled water 
is quite sufficient to dissolve 5 or more grains of KBr, and 
if no water be added, some KBr is very liable to be thrown 
out of solution by the shaking necessary. The next test I 
tried was one by M. Personne, published in the Journal 
de Pharmacie. It depends on the property possessed by 
HgCl 2 of precipitating a solution of iodide but not one of 
bromide, bromide of mercury being soluble. It is neces¬ 
sary to the success of this test that the iodide be free from 
KI0 3 , KC1, and K 2 Co 3 . 
I may mention here that KI0 3 is much more fre¬ 
quently present than is generally supposed, and traces of 
it may genex-ally be detected in the best samples of iodide, 
and as this salt (IvI0 3 ) is stated on pretty good authority 
to be of a poisonous nature, it behoves us to be on our 
guard against it; it is fortunately easy to detect by add¬ 
ing a little starch solution to the iodide to be tested, and 
then adding a small quantity of tartaric acid; a blue 
colour is developed more or less rapidly, by the liberated 
iodide acting on the starch, if the KI contains KI0 3 , 
thus :— 
5 KI+KI0 3 +H 2 C 4 H 4 0 6 =K 2 C 4 H 4 0 6 +5 HI+HI0 3 
and 
5 HI+HI0 3 =T 2 4-3H 2 0. 
