480 Dr. Tycho E :son Auren on the 



The agreement between the values of X. m.k. found in the 

 various concentrations (which in other experiments have, as 

 a rule, been greater than in the above instance) , is, as is shown 

 here, rather good. The special values, from which the 

 average values of Table IV. are calculated, generally differ 

 at the most by 5 per cent, from the average numbers. The 

 values of X. m. k., measured after this method, are, as shown 

 in Table III., within experimental error independent of 

 concentration in the solutions to be examined, apart from 

 colloidal solutions. The fact is that for solutions that are 

 strongly hydrolyzed, and then partly pass into a colloidal 

 state, e. g. FeS0 4 that has been oxidized by the contact of 

 air, Na 2 HAs0 4 , &c, absorption increases with increased 

 dilution, about which see further on (p. 485). The concen- 

 tration of the examined solutions has mainly been chosen so 

 that (d 2 —d 1 ) has, at the utmost, reached about 1°5 cm. by 

 the way of avoiding the correction necessitated if the posi- 

 tion of the comparator is too much changed relatively to the 

 bulb. 



When elements insoluble in water have been investigated, 

 especially organic fluids whose absorption is less than that of 

 water, another absorption vessel containing a water sheet of 

 1 cm. thickness has been placed before the absorption vessel. 

 Then absorption has been examined when this vessel has 

 been empty, as well as when it has been filled with a liquid. 

 After reading the thickness of the water sheet (d 2 — d 1 ) in 

 the comparator, we directly find the thickness of the water 

 sheet that corresponds to a sheet of liquid in question 1 cm. 

 thick. In examining absorption for metals, the procedure 

 has been similar. 



III. Results of Experiments, 



The following table gives a survey of the values of X. m. k. 

 I have found when examining the solutions named in the 

 table. For each substance, as a rule, I have carried out 6-S 

 determinations with various concentrations or, if this has not 

 proved possible, several determinations for the concentration, 

 and the values seen in the table are average values of these 

 determinations. 



