196 BULLETIN OF THE BUSSEY INSTITUTION. 
in presence of acetic acid, in the usual way ; and in the filtrate from the 
ferric phosphate, the rest of the phosphoric acid was estimated by means 
of a standard solution of acetate of uranium.* This process, though 
not absolutely accurate when employed for the analysis of products liable 
to contain alumina, was deemed sufficient for the purposes of this re- 
search, and has in fact approved itself to be so. It will be noticed that, 
besides a tendency on the part of the operator to add a little too much 
of the uranium salt in the process of titration, the error most likely to 
occur in applying this process to the analysis of wood-ashes is that due 
to the presence in the ashes of more or less alumina derived from earth 
that had adhered to the wood from which the ashes were produced. One, 
certainly, and probably both of these sources of error, would tend to in- 
. dicate more phosphoric acid than was actually contained in the specimens 
of ashes under examination rather than less. The alumina would be 
thrown down in the form of a phosphate, together with the phosphate of 
iron, and would be weighed and computed as if the entire precipitate 
were Fe,O,, P, O;. But while a gramme of normal phosphate of iron 
contains 0.47 grm. of phosphoric acid, a gramme of Al ,O;, P,O;,in which 
form the alumina is probably precipitated, contains 0.58 grm. of P,O;. 
As will be seen directly, the amounts of phosphoric acid found in the 
ashes were unexpectedly small. In order therefore to make sure that the 
results obtained by the foregoing method were correct, other processes of 
analysis were resorted to. ‘Thus, for the sake of controlling the first set 
of results, the phosphoric acid was estimated in samples Nos. I. and XI. 
by precipitating and weighing the whole of it as a highly basic phos- 
phate of iron, after Fresenius’s ‘‘ Quantitative Analysis,’? New York 
edition of 1871, p. 273, and determining the amount of iron in the pre- 
cipitate by titrating with permanganate of potash. These estimations 
were made, many weeks after those of the first set, by another, analyst, 
who had no knowledge as to the amounts of phosphoric acid that had 
been found by his predecessor, and who was moreover perfectly familiar 
with the iron process. There was found by the iron process 3.87% of 
P.O; in No. I., and 0.41% in No. XI.,in the samples from which the first 
‘operator, using uranium, had got 3.16% and 0.430, respectively. The 
agreement between these results is complete, since the soapboiler’s ashes 
(No. I.) unquestionably contained a little alumina, all of which would 
necessarily be counted as phosphoric acid in the estimation by the iron 
process. Sample No. XI., on the other hand, contained no alumina, as 
has been already remarked. 
* The value of the uranium solution was determined by means of pure 
diphosphate of soda that had been dried at 100° C. Phosphate of soda thus dried 
is a compound of perfectly constant composition (2 Na,O, H,0, P,O;), that can 
be weighed without the least trouble, or risk of alteration. 
+ My assistant, Mr. W. T. Leman. For the first set of determinations, and 
for most of the other analytical work, I am indebted to my assistant, Mr. James 
A. Beatley. 
