334 Frederick Guthrie on certain 
nearly equal except the last, which, instead of about 84-82 
grams of mercury, only held about 52, for this was the drain- 
age from the amalgam), had the composition: — 
Per cent, of K. 
0-00082 
•0038 
•0146 
•0331 
•1185 
•2061 
•2811 
•3490 
As to the comparison between Na and K, we need onlv 
contemplate the potential difference between 1*92 and 1*34 
respectively. 
"With regard to the frosted appearance mentioned in §§ 27, 
28, it can scarcely be doubted that the minute bubbles which 
compose it are hydrogen, due to the film of water or vapo.ir 
on the glass. But while this appearance travels at the rate 
of at least one foot an hour, there is no sensible quantity 
of Na to be found at even a lesser depth after fourteen days. 
The effect must therefore be a surface-effect, and be of the 
same order as the effect described in § 22, where the mercury- 
slab expands when touched by sodium amalgam, on account of 
the metals spreading almost instantaneously over its surface 
and enfeebling its skin. The condition actually set up in the 
mercury column is probably this: — A minute film of sodium 
spreads downwards between the mercury and the glass : this 
decomposes the water on the glass, and so clothes the glass 
with a film of minute hydrogen bubbles, and the mercury sur- 
face with a film of caustic soda, which latter is in absolute 
contact with the mercury surface. It is a question whether 
the sodium film is less than, equal to, or more than sufficient 
to decompose the water — probably more. At all events it is so 
minute as not to exhibit itself in any chemical reaction. The 
spectroscopic reaction here has no significance. 
The curves Na and K, Plate IV., which represent these expe- 
riments graphically, are not directly comparable with the curves 
Sn, Pb, and Zn (§ 31) in the same plate, because in the case 
of Na and K, for reasons given in § 25, it was found necessary 
to start with an amalgam, and indeed with one containing only 
about 2 per cent, of sodium. The time in the case of the Na 
and K amalgams was also only a little less than half that occu- 
pied in the diffusion of Zn, Pb, and Sn. 
§ 30. The rapid penetration of zinc by mercury suggested 
the question whether, when an amalgam of an alkaline metal 
