BULLETIN OF THE BUSSEY INSTITUTION. 189 
Nitrogen was estimated by the soda-lime process, care being taken to 
powder the materials very finely, and to use large quantities of material 
for those determinations, where the proportion of nitrogen in the shells 
was small. 
To estimate magnesia in the sea-shells proper, ten grammes or more of 
the powdered shell that had been calcined to destroy organic matters were 
dissolved in chlorhydric acid, and a mixture of chloride of ammonium, 
carbonate of ammonia, and caustic ammonia, was added directly to the 
solution, in order to precipitate the lime, as well as any iron or alumina 
that might be present, and the trace of phosphoric acid also. The filtrate 
from the lime precipitate was evaporated to dryness, to expel ammonium 
salts; and the residue, after having been taken up with acid, was treated 
with oxalate of ammonia to remove the last traces of lime. In the filtrate 
from oxalate of lime, magnesia was precipitated as phosphate of magnesia 
and ammonia, in the usual way. In the case of the lobster-shell, it was 
found necessary to remove phosphoric acid in the first place by precipita- 
tion as phosphate of iron, before proceeding to remove the lime, as above. 
For the sake of completeness, I have compiled the following table of 
analyses of shells, published hitherto, that seem to be trustworthy. 
Several well-known analyses have been purposely omitted from the 
table either wholly or in part since they are manifestly incorrect, par- 
ticularly in respect to the item phosphoric acid. An analysis of mus- 
sel-shells by Payen, for example, that has been often cited,* has been 
excluded on this account; and so has Reichardt’s estimation of phos- 
phoric acid (analysis No. XI. of the table). The old analysis of oyster- 
shell by Bucholz and Brandes (Watts’s “ Dictionary of Chemistry,” 4, 
322), which is still often cited, must be regarded as incorrect, in so far 
as phosphoric acid is concerned, since the percentage of this constituent 
is given as 0.55, ¢.e., the equivalent of 1.2% of phosphate of lime; 
98% of carbonate of lime was reported by these chemists, and 0.5% 
of albuminous substance. 
It will be noticed that the Roman numerals in the first column 
of the table correspond with those of the paragraphs below it in which 
the names of the materials and the authorities are given. 
* Compare Wilda’s “ Centrallblatt,” 1858, 6. (II.) 33. 
