( 275 ) 
From this it might be deduced that the difference of free energy 
between red and yellow mercuric oxides was of the order of 7 milli- 
volts, whilst OstTwaLD gives it as smaller than 1 millivolt. 
Here also a careful and exact investigation appears desirable. 
4. In the first place it is rather surprising that neither OstwaLp 
nor GLAZEBROOK and SKINNER say anything about the purity of the 
materials, which is nevertheless of the greatest importance owing 
to the great sensitiveness of the electrometric method. In determi- 
nations made by this method, traces of impurity may have an 
enormous effect on the E.M. F. measured. 
Further, the preparation of absolutely pure mercuric oxide, whether 
red or yellow, is to be counted among the more difficult tasks of 
preparative chemistry, as may be seen from what follows. 
Four specimens (from different makers) of red mercuric oxide 
(hydrargyrum oxydatum rubrum praecipitatum pro analysi) and yellow 
oxide (hydrargyrum oxydatum via humida paratum pro analysi) 
could not be used on account of the many impurities which they 
contained. 
In the red and yellow oxides from Merck of Darmstadt, I could 
not discover any traces of impurity by analytical means; these 
preparations were therefore used as the starting point of the in- 
vestigation. 
The water which served for making the solutions employed, was 
very pure and possessed a conductivity of 1 x 10-®. It was 
distilled from a heavily tinned copper vessel a trace of phosphoric 
acid being added; the middle fraction was used, carbonic acid being 
removed by means of an air current in the well known way. 
The solution of potash used was made from potassium and this 
water, the carbonic anhydride of the air being excluded. 
The mercury was, after a preliminary purification, twice distilled 
in vacuo. All glass vessels, flasks etc. with which the substances 
came in contact were steamed out. 
5. <A cell was now put together of the form shown in the figure ; 
aa and bb are glass tubes 7} em. long and 2 em. wide, connected 
below with capillaries, which are bent round and carried upwards 
to 2 cm. above the caoutchouc stoppers k k. In this way it is 
possible to submerge the whole apparatus in water without the 
platinum wires P, and P>, which must be heated to redness before 
each experiment, coming in contact with it. These wires terminate 
in the mercury which is placed on the bottom of aa and 45. Upon 
