HEAT AND HEAT OF TRANSFORMATION OF IODIDE OF SILVER. H75 
II. This specific heat for the alloys of silver- and copper-iodide diminishes when the 
percentage of silver*iodide increases, but if we calculate the specific heat of the alloys 
from the specific heats and the proportion of components, we approach only roughly to 
the true values. It seems that the coefficient of T+£ generally increases with the 
percentage of Agl. The coefficient 0*000028, for Cu 3 I 2 .3AgI, appears to be less than 
that for Cu 3 I 3 .2AGI ; but, perhaps, that is due to some uncertainty in the experiments, 
notably Nos. 25 and 26. The coefficient for Cu 3 I 3 .12AgI was not determined, owing 
to the low temperature of change of this substance. 
III. The value of the heat of change, X, for PbI 3 .AgI is less , and for the alloys of 
silver- and copper-iodide is greater than for pure iodide of silver. This singular result 
leads to the assumption that some or all of the alloys of copper- and silver-iodide are 
real chemical or molecular compounds. Starting from the values of X for Agl and for 
Cu 3 I 3 .AgI, and calculating the values of X for the other alloys, considered as mixtures 
of Agl and Cu 3 I 3 .AgI, we find the numbers 7*99, 7*62, 7*38, 6*72 respectively for 
Cu 3 I 3 .2 Agl, Cu 3 I 3 .3AgI, Cu 3 I 3 .4AgI, Cu 3 I 3 .12AgI. These numbers differ from those 
given above, but making allowance for errors of observation, for want of the values of c Y 
respecting Cu 3 I 3 .AgI and Cu 3 I 3 .2AgI, and for the uncertainty of the temperature for 
which X is computed, it is perhaps probable that at least Cu 3 I 3 .2AgI and Cu 3 I 3 .3AgI 
might be considered as mixtures of Agl and Cu 3 I 3 .AgI. The uncertainty in the 
temperature of change is chiefly influential in Cu 3 I 3 .4AgI and Cu 3 I 3 .12AgI, for 
which bodies the interval of temperature between the commencement and end of 
the change of structure exceeds 100° C. Moreover, for these substances c 1 is small, 
and the coefficient of T+£ in the value of c has not been determined for Cu 3 I 3 .12AgI, 
and is somewhat uncertain for Cu 3 I 2 .4AgI 3 owing to discrepancy in the experiments 
Nos. 16 and 17. For these reasons the values of X deduced from observation for those 
two substances do not pretend to great accuracy. 
IY. The values of c 1 increase with the percentage of Cu 3 I 3 ; and in the case of Agl 
and those alloys of Agl and Cu 3 I 3 for which c 1 was determined, the values of c 1 are 
smaller than the values of c computed for corresponding temperatures. The difference 
between c 1 and c, extended beyond the temperature of change, diminishes as the 
percentage of Cu 3 I 3 increases. It is therefore probable that c 1 does Dot much differ 
from c in Cu 3 I 3 .2AgI and Cu 3 I 3 .AgI, for which the values of X are computed on this 
assumption. 
Y. For PbI 3 .AgI the value of c l is greater than the value of c for corresponding 
temperatures. 
Perhaps other conclusions may be derived by comparing our calorimetric results 
with those of thermic expansion and contraction obtained by Mr. Rod well ; but 
MDCCCLXXXII. 7 L 
