170 
Analyses of the Ashes of Plants. 
Potash and Soda. 
The arguments and experiments wliicli apply to chlorine apply 
also to these ingredients of plant-ashes, which could only be lost 
in the form of chlorides. 
The chlorides of potassium and sodium (especially the first) 
are volatile at very high temperatures, but they may be heated 
to dull redness with perfect security, as is done every day in the 
separation and estimation of these alkalies. We have shown that 
at a temperatui'e adequate to the burning of the most troublesome 
vegetable no appreciable amount of chlorine is lost, and under 
the same circumstances, therefore, no loss of the alkalies is to be 
feared. 
It is, after all, a matter of temperature, and the discrepancies 
in the results obtained by different ash analysts may, no doubt, 
in many cases be referrlble to this cause. The burning of a 
plant for analysis is plainly a process requiring considerable care 
— a process that should in no instance be hurried forward ; — but 
we think that, on the whole, we have demonstrated that (with 
the exception of the sulphur, which is obtained by a separate 
process) our method of burning crops affords an ash really and 
correctly representing the mineral matter of the plants under in- 
vestigation. 
It only now remains that the method of analysing the ash 
so prepared should be shortly adverted to. In the last Report 
on plant-ashes a short notice of the methods of analysis was 
inserted, and we shall in this place notice the only particular 
upon which any doubt could possibly arise, namely, the estima- 
tion of the phosphoric acid. The processes employed in the 
estimation of the other constituents are those in common use 
amongst chemists, and universally allowed to be accurate. 
In the estimation of phosphoric acid a measured quantity of 
iron solution is employed. It is sometimes thought that a suffi- 
cient amount of accuracy cannot be ensured by this measuring of 
the test solution, in which a long delicately graduated measure 
capable of containing 1000 grains of liquid is employed. The 
following trials were made of different quantities of the same iron 
solution, the oxide of iron being precipitated by ammonia, burnt, 
and weighed : — 
Oxide Grains in 
of Iron. 1000 of liquid. 
Istcxp. 660-0 measures gave 2-140 = 3-242 
2nd „ 972-0 3-150 = 3-241 
3rd „ 24-4 ,, -790 = 3-238 
These experiments were made by one of us. The following 
are results of the measurement of the same iron liquids by the 
other author of this paper : — 
