559 



My experiiiienis lead to the following values lor this radius: 



z\ ^0.157 /*, = 0.410 r, = 1.24X 10-*^ ram. 

 (beginning of the 

 flat piece) 



i, = 0.362 A, = 0.517 r, = 1.67 X 10-g mm. 



i, = 0.491 /i, = 0.596 / •, = 2.13 X 10-^ mm. 



i^ = 0.57 h, = 0.66 r, = 2.57 X 10-« mm. 

 (end of the 

 flat piece"). 



The values found for the radius of the micro-capillaries, are in 

 such close agreement as regards order of magnitude with the values 

 of ZsiGMONDY and Anderson, and with those of Bachmann, that it is 

 astonishinti that always this order of magnitude is again met with. 

 (The second system of capillaries which Bachmann thinks that he 

 can derive from his curves, seems questionable to me). 



The agreement in the form of the curves for the heats of sorption 

 with Iheir typically flattened [)iece corroborates that the flat pai't 

 of the isotherm for carbon and for silicic acid has the same cause. 



It is the more striking under these circumstances that Bkri, and 

 Andhkss have found that the same carbon which gives a flat middle 

 piece in the isotherm with water, lias (/ curve wit/iotit any Jiat 

 middle piece, and with a much longer horizontal initial part (for 

 small I) loit/t organic liquids (as benzene or methyl alcohol). If the 

 correctness of these experiments is confirmed, tiiey furnish the proof 

 that Zsiomondy's explanation, cannot be the true one, at least for 

 carbon. I am, therefore, occupied with a repetition of these experi- 

 ments, and also with a determination of the heats of sorption. 



Since Zsigmondy's explanation is inadequate to account for the 

 flat piece in the isotherm and for the flattened piece in the heats 

 of sorption, it is in ray opinion natural to see a connection between 

 the deviating form of the isotherm of walei- anil the fact that ivater 

 7)ioisten.s solid bodies, as cai bon, much less easily than organic liquids, 

 as benzene or methyl alcohol, do. We should then have to do in 

 water and carbon with surface adsorption at a surface that is not 

 easily moistened, a phenomenon of which so far only one example 

 has been studied somewhat more closely '), viz. the adsorption of 

 watervapour to glass-wool which has been Ihorougiily dried before- 

 hand, investigated by Trouton'). The glass-wool had been previously 



') Freundlich, Kapillarchemie, 2"^ edition, p, 223. Possibly there is solid 

 solution present as a complication in tfie boundary layer also here. 

 ') Freundlich, loo cit. 



