EXPERIMENTAL KNOWLEDGE OF THE PKOPEllTIES OF MATTER. 133 



Troost ' gave FeoClg, AlaClg, AlaBr^, AlaTg, from about 400° to 1040° (the 

 boiling-point (?) of zinc). These formulfe had led V. Meyer to expect 

 In2Cl6, which was not given by any vapour- density determination of 

 indium chloride. So V. and C. Meyer find formuloe Sn2Cl4, ZnOia for 

 temperatures between 620° and 700° as determined by a block of pla- 

 tinum and a calorimeter. 



Mitscherlich had found at 571° AS4O6, and V. and C. Meyer find for 

 a much higher temperature — about 1000° — the same formula AsjO^, and 

 for even higher temperatures SbjOg, CU2CI4, and at over 900° CdBr, ; S2 

 at about 1500° (?), while at temperatures below a bright red heat the 

 vapour-density gave Sg. They tried potassium, sodium, and then 

 chlorine, but found that these attacked porcelain. 



The temperatures in these experiments with the calorimeter and the 

 heated block of platinum were not very accurate when very high tempe- 

 ratures -were to be measured. The highest estimated temperature 1567° 

 gave O2 (from AgoO), ^2) 83 as the molecular formute of oxygen, 

 nitrogen, and sulphur. For chloi'ine at the highest temperature of their 

 furnace they obtained a molecular formula f Clj, that is, from Pt2Cl4 the 

 chlorine given off, which at as high a temperature as about 620° had given 

 density corresponding to formula CI2, had given smaller and smaller 

 values for the densities at higher temperatures, till at the highest tem- 

 perature it had a density 1-60, 1-62, a little less than 1-63 calculated for 

 §Cl2 ; admitting that chlorine was undergoing dissociation it was not 

 clear that it would not at higher temperatures give still lower densities 

 (always compared with air). 



The results thus given for chlorine naturally led to speculation as to 

 the behaviour of bromine and iodine in the same circumstances ; and as 

 Deviile and Troost had {loc. cit.) found for iodine a normal vapour- 

 density corresponding to I2 at the bright-red heat required for reaching 

 the boihng-point of zinc (1040°, as found by Deviile and Troost) the 

 result with chlorine was considered doubtful ; this taken in conjunction 

 with the fact of the porcelain being attacked by alkali metals and 

 chlorides led to a revision of the arrangement of the apparatus. In 

 succeeding investigations V. Meyer, in conjunction with Ziiblin, used 

 a porcelain tube glazed inside and out, and placed in the furnace so 

 as to be heated by it when necessary to the highest temperatures. The 

 gases of the furnace were thus entirely unable to diffuse into the interior 

 of the porcelain tube, and thus the platinum tube which was inside the 

 porcelain tube was absolutely guarded against the action of these gases, 

 and operations which would be vitiated by the action, either of the gases 

 of the furnace on porcelain, or of the substance which was the subject of 

 the experiment, could be heated with safety to the highest attainable 

 furnace temperatures in the jjlatinum tube. The experimental tube was 

 filled with nitrogen, and the vapour-density determined in an atmosphere 

 of this gas. The temperature was now accurately determined (it having 

 been previously found that up to very high temperatures nitrogen, 

 oxygen, mercury, and, as afterwards shown, hydrogen, when compared at 

 the same temperature, gave always vapour-densities corresponding to the 

 same formula; in fact, that in all these cases the absolute densities 

 diminished as temperature rose always in the same ratio) by measuring 

 the nitrogen which filled the tube before the experiment, and the nitrogen 

 ■which filled it after the experiment. 



' Annaleg de Chimie et de Physique, 1860 (3), Iviii. p. 257. 



