1904-5.] Mr Cameron on the Constitution of Complex Salts. 725 
From the first five mixtures perfectly definite crystals were 
obtained, and these were in each case recrystallised. The 
crystals finally obtained were indistinguishable in appearance. 
The sixth solution yielded no crystals on evaporation, but left a 
gummy mass. 
The density of the crystals was determined in accordance with 
the method described by Retgers,* using a mixture of bromoform 
and turpentine, which was found to give good results. In each 
case I selected the densest crystals — those most free from air- 
cavities — by separating such as sank first in liquid of almost the 
same density, and determined the density of the crystals by 
adding bromoform to the same liquid till the crystals neither 
rose nor sank, and ascertaining the density of this liquid by 
the pyknometer. The temperature at which these determinations 
were made was about 10° C., though it may be pointed out that 
slight differences of temperature should produce no appreciable 
difference of density in the solid crystals, and hence the determina- 
tions should be unaffected (provided no transition temperature be 
passed through). 
The following results were obtained : — 
Crystals from 
Mixture - 
Density. 
Specific Volume. 
1 
2T33 
•4689 
2 
2-111 
•4737 
3 
2T09 
•4742 
4 
2-099 
•4764 
5 
2-054 
•4868 
The percentage of ammonia was determined by distilling very 
dilute solutions with alkaline carbonate and Nesslerising the 
distillate. The second and third mixtures by themselves seem to 
show that the crystals resulting from such mixed solutions have 
no definite composition. The percentages are as follows : f — 
* Zeit. fur phys. Chem ., 3, p. 289. 
t It will be noticed that in each case the percentage of ammonia found is 
less than half that present in the corresponding solution. The results were 
confirmed by a determination of the potassium (as sulphate) and of the 
ammonium (by titration method) in the case of the fifth mixture. 
