310 DR. KANE ON THE CHEMICAL HISTORY OF ARCHIL AND LITMUS. 
being' so unusual a direction for the analysis of an organic substance to vary from the 
truth, a modification of the formula should be adopted. I consider, however, that 
where the formula gives, in the other instances, results so very consonant to experi- 
ment, there may be in that one case some little inaccuracy of weighing suspected, 
particularly as the excess of hydrogen is also in some degree greater than should be 
found in the analysis of a substance that had been so well dried. 
Another reason for the adoption of the formula C 18 H 7 0 16 is, that it accounts most 
easily for the formation of this substance at the expense of the azolitmine ; the azo- 
litmine losing an equivalent of ammonia and gaining six equivalents of oxygen. 
Thus 
Azolitmine. . . C 18 H 10 N O 10 
- H 3 — N + O 6 
Spaniolitmine . . C ]8 II 7 0 16 
the replacement being regular; for as the hydrogen and azote do not exist as ammonia 
in azolitmine, they are, although perhaps evolved as ammonia, separately replaced, 
the three equivalents of hydrogen making way for three of oxygen, and the nitrogen 
being equivalent to three equivalents of oxygen also, as has been fully proved by the 
Researches of Laurent and of others. Indeed, if azolitmine be considered as the source 
from whence the spaniolitmine is derived, no other formula can well be taken, for any 
change in the number of equivalents of carbon would render the decomposition un- 
necessarily complicated. 
If, however, the spaniolitmine be supposed to originate in the decomposition of ery- 
throlitmine, a formula may be adopted at once deducible by the simplest considera- 
tions of substitution, and agreeing fully as well as that previously given with the 
analytical results. Thus 
Erythrolitmine . . . = C 26 H 22 0 12 
— H n + O u 
Becomes spaniolitmine = C 26 H n 0 23 
The calculated results of this formula I subjoin, in order that they may be compared 
with the experimental numbers, and with those given by the other theory. 
C 26 H n 0 23 . C 26 H n 0 23 + 4 AgO. 
= 158'6 
44-85 
^26 
= 158-6 
19-40 
= ll'O 
3-11 
H n 
= 11*0 
1-34 
= 184-0 
52-04 
^27 
= 216-0 
26*42 
353-6 
100-00 
A gi 
= 432-0 
52-84 
817*6 
100-00 
These results agree better with experiment than those given by the other view, but 
the constitution of the purple lead salt is opposed to it, as, the organic substance in 
the purple lead salt being C 26 I-I U 0 23 , the oxide of lead amounts to seven and a half 
