106 EEPOET — 1886. 



The pressure of Hg-vapour as determined by Hertz ' at 10° is 

 0-0005 mm., and at 20° is 0-0013 mm. ; Bobr's results recorded in bis 

 paper are for pressures of 0-1 mm. and upwards, and therefore could not 

 be seriously affected by any error due to pressure of mercury vapour ; 

 moreover, bis measurements of pressure being in all cases differential 

 ■with a mercury-barometer, this pressure would be almost if not entirely 

 eliminated. 



The power which the inner surface of the glass tube may have m 

 condensing the oxygen in the conditions of these experiments must not 

 be disregarded when such very small pressures are concerned ; it may be 

 that the walls of the tube and the surface of the mercury are not wholly 

 without action on the gas ; if any action were exerted it might well be 

 quite imperceptible in relation to pressures of, say, over lOOmm. ;_ any 

 source of error which might be so trifling as to be ignored at higher 

 pressures might be important at abnormally low pressures ; in these 

 extreme cases the relation of pressure to volume as observed may possibly 

 be not wholly determined by the molecules of oxygen, but may be partly 

 influenced by the surface of the glass or of the mercury ; cf. Amagat 

 ('Ann. de Chim. et de Phys.' 5, xxviii. p. 499). 



It must be allowed that Bohr has gone a long way to meet any doubt 

 on this head by using two tubes of very different diameters ; but further 

 experiments with still wider tubes, and with tubes of different varieties 

 of glass, would probably show whether the surface of the glass in contact 

 with the oxygen is in any way concerned. 



Some remarks of Regnault ^ in reference to Dalton's law of the 

 mixture of gases with vapours as a ' theoretical law,' and showing devia- 

 tions from it, due to the inner surface of the glass, are suggestive of such 

 an influence in other cases by analogy. 



Certainly, on the face of them, the observations of Bohr are remark- 

 able, and his singular point diflicult to account for in any other way than 

 the direct and obvious one ; and his experiments show the greatest pams 

 taken to minimise the effects of the possible sources of error. 



Volume and Temperature — SegnauU. 



Regnault determined the dilatation of dry air by many series of 

 experiments and by different methods. Representing by 100« the whole 

 expansion of unit volume of dry air at 760 mm. between 0° and 100°, his 

 different series gave, taking the mean of the numerous determinations of 



1. 100a=-36623 with possible error -00140. 



2. 100a=-36633 „ „ „ -00101. 



3. 100«=-36679 „ „ „ .00079. 



4. 100a=-36650 „ „ „ -00130. 



In the above experiments the dilatation was arrived at indirectly by 

 observing the increased pressure of air at constant vohime between 0° and 

 100°, and deducing by Boyle's laws the volume which the air at 100° 

 would then occupy. 



The mean of a fifth set of experiments, by which the value of 100a is 

 determined for expansion at constant j^ressure, gave — 



6. 100a =-36706, with possible error -00025. 



' Wiedemann's Annalen der Phijsik nnd Chemie, xvii, p. 199. 

 2 Memoires de VAcademie, t. xxvi. iDp. 694, 69.5. 



