330 Scientific Proceedings, Royal Dublin Society. 
with pure carbon [prepared from sugar], kaolin, quartz, obsidian, 
basalt, silica, or alumina :— 
1. Settlement is effected much more rapidly in solutions of 
diads such as MgCl., BaCl,, CaCl,, than in equimolecular solutions 
of monads, as NaCl or KCl. 
2. Between diads and monads the effects appear about equal 
when the molecular equivalent concentrations are as | : 82. 
3. In the experiments the concentrations ranged from 1 gram 
molecule in 6 litres to 1in 100 litres. Concentrations of 1 gram 
molecule in 2 litres showed little or no difference between equi- 
molecular solutions of monad and diad salts.' 
4. When A1,Cl, was used as the triad salt, and compared with 
solutions of monads and diads, all made up to the concentrations 
of Linder and Picton’s experiments, the triad was apparently 
of insufficient concentration, showing a marked lag in precipitation. 
Al, (SO,); did not reveal this insufficiency. 
5. The suspension generally reveals the valency effects only 
after 12 to 24 hours standing, the earlier settlement being effected 
without notable differences. 
6. Suspensions in distilled water lag in precipitation behind 
those containing even small quantities of salts; but concentra- 
tions less than 0-04 gram-equivalents per litre of monads or diads 
produce very little effect. 
7. Suspensions in acidified water clear much more rapidly 
than those in alkaline water. The latter may remain up in- 
definitely. The acidity of certain salts hence enters as a com- 
plication in the experiments. 
8. Flocculation ig earlier and more markedly visible to the 
eye or lens in the case of diads and triads than in the case of 
monads. 
‘Later experiments have elucidated move fully the cause of this. There is an 
optimum concentration, for example, in the case of MgCl, ; less or greater concentra- 
tions precipitate more slowly. Compared with NaCl it is found that the curve of 
precipitating power of the latter crosses that of the former salt at a concentration 
of about 0°75 gram-equiyalent molecules per litre. At about this concentration, 
therefore, both salts act alike. At further concentrations the monad salt has even 
the advantage. These curves cannot be explained as merely due to viscosity, but 
inyolye, apparently, the theoretical views given above as to the swamping of the 
valency effects at high concentrations. This is effected at lower concentrations for 
diads than for monads, as might be expected. Hence the curves cross. 
