139 
this result, yet allowing for all these, there seemed to be 
some other cause. In my last paper I mentioned that an 
ammonaical solution of copper, when largely diluted, became 
turbid, and that to carry out the experiment an additional 
quantity of ammonia was necessary. This small quantity 
was added at hazard, as I did not think it would have any 
influence on the result. It seemed to me afterwards to be 
a point worth examining in connection with the above 
experiments. Two solutions were made, each containing 
5 cub. c. of the previously mentioned copper solution with 
245 cub. c. of water. The solutions were placed in similar 
cylinders — to one of the cylinders more ammonia was 
added : it appeared perceptibly darker than the other. 
Hence it appears that the excess of ammonia has some 
influence on the result. I presume that ammonio-sulphate 
of copper has a tendency to be decomposed by water, and 
that some change is effected even before it becomes 
obviously marked by the formation of a turbidity; moreover 
it seems likely that the excess of ammonia has the power 
to counteract this property of water, and to restore the 
original compound. Two solutions were made, the bulk of 
each being 545 cub. c., one containing 4000 of copper 
sulphate, along with an additional 20 cub. c. of ammonia, 
the other containing 6000 of copper sulphate with 30 cub. c. 
of additional ammonia. The comparisons were made in new 
cylinders graduated to millimetres. An experiment with 
white surfaces, external, gave the following results : 
Standard solution 4000 in 545 cub. c. of water, length of 
column 23 cm. 
ABC 
6000 15-5 5955 
thus the result is very near the real quantity. I also took 
shorter lengths of the standard solution, namely 18 cm., 13 
cm., and 8 cm., the corresponding lengths of the other 
solution were 12-4 cm., 8*5 cm., and 5*3 cm. Reduced to 23 cm. 
