532 
Proceedings of the Royal Society of Edinburgh. [Sess. 
They are represented by the curves, which, in order to facilitate comparison 
with the nickel salts, are printed in the third article of the series. To make 
the curves complete, the values of A for the infra-red and ultra-violet are 
added from the first and fourth articles of the series. The full considera- 
tion of the curves is postponed until the fourth article of the series. The 
values of A obtained with the spectrophotometer are represented by o s, 
the values obtained otherwise are represented by x s. 
As in the infra-red, the bromide and iodide tended to become turbid on 
standing. The following three curves for the bromide illustrate this very 
nicely. The heavy curve was taken within twenty-six hours of making 
up the solution, and in this case the solution was filtered before taking 
Fig. 1. 
the readings. The light curve was made by diluting a stronger solution 
which had been previously filtered, and the readings were taken forty-eight 
hours after the solution was made. The dotted curve was obtained from 
the same diluted solutions four days later, the solution not being filtered 
during the interval. It will be noticed that the heavy curve is both 
highest at the maximum and lowest in the red, while the dotted curve is 
lowest at the maximum and higher in the red. This is easily explained. 
When the bromide decomposes, the solution absorbs less all over; A 
decreases everywhere in the same ratio ; hence the decrease in the maximum 
value of A. But, at the same time, there is a turbidity produced which 
increases A by approximately the same amount throughout the visible 
spectrum. Hence the increase in the red. Instead of the mean, in this 
case the values represented by the highest curve were entered in the table. 
