NOTES ON ATOMIC WEIGHTS. 211 



of the purification of metallic zinc by distillation in vacuo it 

 is rather odd to find it stated that indiarubber tubing with 

 glycerine joints could not be used because the vapours of zinc 

 and of glycerine interact. What pressure of the vapour of 

 each is likely to exist at the highest temperature to which the 

 joints would ever be subjected ? The presence of gold in 

 the nitric acid distilled from a platinum still, and coming 

 from the gold solder used in it sounds also rather peculiar. 

 One knows that very finely divided gold will dissolve in 

 fuming nitric acid if kept cold, but one could hardly have 

 thought of finding it as an impurity in nitric acid prepared 

 by distillation. But the gem of all the statements comes at 

 the end of the paper when these two rising experimentalists 

 proceed to criticise Marignac's work (14), and finally to 

 teach him and us how we ought to test for oxides of nitro- 

 gen by means of starch and potassium iodide. After 

 proving to their own satisfaction by a process which cannot 

 reveal the presence of any of these oxides that they are 

 therefore obviously absent, they conclude that Marignac 

 was ignorant of the necessary precautions which must be 

 taken to exclude oxygen, especially that of keeping the 

 solution practically boiling so that the steam may keep out 

 the air. It is usually accepted as a well-established fact 

 that the delicacy of this reaction decreases rapidly with rise 

 in temperature, and that the colour goes completely before 

 the boiling point is reached, even in the presence of 

 relatively large quantities of free iodine. 



Amongst other noteworthy determinations of atomic 

 weights made recently are those of Winkler, who finds 

 the values Ni = 58*91 and Co = 59*67 by means of the 

 reaction between the chlorides and silver (15); and still 

 more recently Ni = 5871 and Co = 59*37 (16) by deter- 

 mining the amount of iodine required to unite with the 

 pure metal. Winkler uses the value Ag = 107*66, if we 

 use O = 16 or Ag = 107*93 these last values become 



Ni = 58-863 

 Co = 59*517 



The determinations of the atomic weight of boron by 



15 



