172 
Analyses of the Ashes of Plants. 
lime, equal to 6'16 per cent, of lime. The liquid divided into 
two portions ; the magnesia thrown down hy the addition of 
ammonia (there being an abundance of phosphoric acid already 
present) gave 1"92 phosphate of magnesia, equal to 7 '03 per 
cent, of magnesia. 
The phosphoric acid separated by the addition of ammonia 
and of sulphate of magnesia from the other portion of the liquid, 
yielded 6 "05 phosphate of magnesia, equal to 38-31 per cent, 
phosphoric acid. The results on 100 grains of the ash of peas 
by these different methods are thus staled : — 
riiosphoric 
Lime. Magnesia. Acid. 
The phospholic acid, being separated by oxide 1 q.|q 7' 14 37-&9 
of iron ) 
Tlie phosphoric acid separated after the lime] g.jg 7'03 38'31 
as jjhospliate of magnesia / 
This experiment proves satisfactorily that lime may be entirely 
precipitated as oxalate in the presence of free acetic acid, and 
that phosphoric acid will, in the same circumstances, remain in 
the liquid for subsequent separation. 
It is necessary to make a few remarks in reference to the com- 
position of the phosphate of iron which is thrown down in the 
first stages of plant-ash analysis — that is, when the acid solution 
of the ash is neutralized by acetate of ammonia. To this phos- 
phate of iron Fresenius ascribes a definite composition, with a 
percentage of 
42 "78 peroxide of iron. 
57*22 phosphoric acid. 
We have been led, however, from several causes to doubt the 
uniformity of this compound, and to institute some experiments 
on the subject. If an acid liquid containing a persalt of iron 
and an alkaline phosphate in excess be neutralized by ammonia, 
phosphate of iron is thrown down, more or less coloured, and of 
variable composition. The neutral phosphate of iron is (as is 
well known) robbed of phosphoric acid by a fixed alkali, such as 
potash, and reduced to the red-coloured sub-phosphate. The 
same occurs, although in a less degree, with caustic ammonia, and 
the yellowish-white phosphate of iron is soon reddened in an 
ammoniacal liquid. Indeed so unstable is this compound that it 
will part with phosphoric acid even to some ammoniacal salts: 
thus if the phosphate of iron when almost colourless be dissolved 
in hydrochloric acid and re-precipilatcd by acetate of ammonia, it 
will be found to have acquired a very decided red tint, and phos- 
phoric acid may be detected in the filtered liquid. This circum- 
stance, it is plain, must materially affect the estimation of oxide 
of iron, where it is calculated from the amount of phosphate pre- 
