404 PROCEEDINGS OF THE AMERICAN ACADEMY. 



tion between the amount of gas retained by the oxide and the tempera- 

 ture used in its ignition. The material employed in these experiments 

 was prepared by dissolving pure copper wire or electrolytic copper in 

 pure nitric acid, and converting by heat the recrystallized cupricni- 

 trate into basic nitrate and finally into oxide.* Platinum vessels were 

 used throughout, whenever heating was necessary, excepting that one 

 or two of the first ignitions were conducted in a porcelain boat. Three 

 preparations of the oxide were made ; one by the rapid decomposition of 

 the nitrate, another by more gradual application of heat, and a third by 

 the very slow action of a temperature just barely sufficient to effect 

 the desired change. This temperature had been previously found by 

 a special experiment to be about 275°, if the products of the decomposi- 

 tion remain in contact with the cupric oxide. The first of these prejDara- 

 tions was the most coherent, and was capable of retaining the most gas ; 

 the second was less coherent and retained less gas, and the last was the 

 most powdery and retained even less gas. Thus it is manifest that the 

 physical condition of the solid is a factor in the question. 



These specimens of the oxide of copper were subsequently ignited 

 at varying temperatures for varying lengths of time in two different 

 forms of furnace. For the first series a large tube or small muflle of 

 Berlin porcelain capable of being heated to 1200° or more in a Fletcher 

 furnace was used, .but since it was found difficult to maintain the tube at 

 a constant temperature throughout, this arrangement was later aban- 



* It is interesting to note that the hydrate of cupric nitrate, which crystallizes 

 from hot solutions, probably has the formula 3 CufNOs)., . 8 H^O (analogous to 

 cadmio sulphate), not Cu(N03)2 . 3 HoO (Graham), as is usually supposed. Follow- 

 ing are the analyses of three separate preparations, dried over moist potash (which 

 docs not dehydrate the crystals at ordinary temperatures) proving this point. 



(1) 0.1739 gram of crystals yielded 0.0471 gram of copper = 27.08 per cent. 



(2) 0.6310 " " " 0.1705 " " = 27.01 



(3) 1.2897 " " " 0.3448 " " - 26.73 



Average = 26.94 " 



Theoretical value for Cu(N03)2 . 3 HoO = 26.20 per cent of copper. 

 3Cu(N03)2.8H.20 = 26.98 

 2Cu(N03)o.5H20 = 27.35 



The crystals analyzed in No. 3 above were large and undoubtedly contained 

 included mother liquor. Whether or not other hydrates exist, and what the tran- 

 sition temperatures may be, I did not take the trouble to discover. Much of our 

 present data regarding water of crystallization has been rounded off in the manner 

 indicated above. Undoubtedly many more such irregular ratios as 3 : 8 really 

 exist than we are prepared at present to believe. 



