314 Scientific Intelligence. 



unity, was, in the vapor of sulphur at 440° 7*30, 7*50 and 7*40; 

 in that of phosphoric sulphide at 518°, 7'1 6 and 7*20; in that of 

 stannous chloride at 606°, 6'34; in the furnace at 758° 4*80; at 

 835°, 4-54 ; at 943°, 4'56 ; at 1117°, 4-27 ; at 1244°, 4'25 ; at 1260°, 

 4-28 ; about 1400°, 4*26 ; about 1600°, 4*08. It appears from these 

 results that between 440° and 758° the vapor of aluminum chloride 

 is in a condition of permanent dissociation and that it does not 

 attain a normal gaseous condition until the temperature reaches 

 800°, when the vapor-density becomes 4*60 and the molecular mass 

 133*15 corresponding to the formula A1C1 3 . Between 800° and 

 1000° there can be no doubt that aluminum chloride has a constant 

 density agreeing with the composition A1C1 3 . For the Dumas 

 method, the aluminum chloride, formed by passing hydrogen 

 chloride gas over the metal, was sublimed directly into the cylin- 

 drical bulb, of about 100 cc capacity. After heating so long as 

 vapor was evolved the tube was sealed, cooled and weighed. It 

 was then opened under mercury and the volume of residual gas 

 measured. After washing and drying the tube it was again 

 weighed, the difference being the apparent mass of the aluminum 

 chloride and of the residual gas. It was then filled with dry air, 

 sealed and weighed a third time. The difference between the 

 second and third weighing gave the apparent mass of the air. 

 It was then filled with water and weighed for the purpose of 

 calculating its capacity. The vapors employed for heating were 

 those # of nitrobenzene, 209°, engenol, 250°, diphenylamine, 301°, 

 mercury, 357°, antimonous iodide, 401°, and sulphur, 440°. The 

 vapor-densities obtained varied from 9*92 at the lowest temperature 

 to 8*78 at the highest; being 9*62, 9-55, 9*34, and 9*02 at the inter- 

 mediate ones. These results, therefore, confirm those obtained by 

 Friedel and Crafts as to the general fact that there is a vapor- 

 density of 9*20 within these limits of temperature. But this value 

 is not constant and happens to exist at about the boiling point of 

 mercury ; while within a range of 50° above and below this point 

 it varies considerably, and between 209° and 440° it changes by 

 an entire unit. The authors conclude, therefore, (1) that starting 

 from the boiling point, aluminum chloride is continually disso- 

 ciated as the temperature rises, not becoming a perfect gas until 

 800° is reached; at which point the vapor-density 4*55 is attained 

 corresponding to the formula A1C1 3 and the molecular mass 

 133*15; (2) that this vapcr-density is departed from only very 

 slightly even at the highest temperatures attainable; and (3) 

 that hence aluminum is trivalent. 



In the following paper Friedel regards the conclusions of 

 Nilson and Pettersso:^ as too absolute. If there is disso- 

 ciation between 200° and 400° there must exist a vapor capable 

 of dissociating, and this can be only A1 2 C1 6 . Moreover, if 

 curves be constructed representing the results given by the 

 two methods, these curves do not unite with each other to 

 form a single curve. The determination of density between 200° 

 and 400° is regular, as is also that above 600°; while the portion 



