16 PROCEEDINGS OP THE AMERICAN ACADEMY. 



In the case of the double bromides, a singularly beautiful phe- 

 nomenon took place. The antimouious bromide would sublime and 

 cr^-stallize all over the edge of the powdered mass of salt in those 

 characteristic fine colorless needles measured and described by Pro- 

 fessor Cooke.* These crystals were for the most part of microscopic 

 size, but some attained a length of two or three millimeters. On fur- 

 ther desiccation they soon disappeared. As would be expected, a 

 direct experiment showed that the salt SbCls . CaCL . 8 HoO decom- 

 posed much more rapidly than the salt SbBrg . CaBrj . 8 HgO, under 

 precisely similar conditions. 



Richards, t in 1890, described a tetra ammon-cupriammonium bro- 

 mide with a formula CuBro . 6 NH3, which loses four molecules of 

 ammonia in a desiccator, forming CuBra . 2 NH3. Although the 

 escape of ammonia from its compounds is not uncommon, yet we 

 believe this is the nearest case to the one in point, inasmuch as the 

 loss here is perfectly definite. 



The escape of antimony haloids from these compounds at a low 

 temperature has a very important bearing 011 the question of molecu- 

 lar combination. It is almost universally admitted that the relation 

 of the water of crystallization to a salt is different from that of its 

 other components. A study of these compounds, however, shows 

 that there is apparently no greater affinity between the two haloids 

 than between the compound and the water of crystallization. Hence, 

 in whatever manner the latter is combined, it is reasonable to suppose 

 a similar union of the two haloids. If, therefore, the crystal water of 

 a salt is held in molecular combination, the evidence as set forth in 

 this research would be interpreted as signifying a molecular union of 

 the two haloids. 



* Tliese Proceedings, XIII. 76. 



t Berichte der deutsch. Cliem. Gesell., XXIII. 3790. 



