or calcium sulphate against sodium chloride; keeping the marine proportions. Again we find that the magnesium and calcium salts are the most effective, although so much more dilute than the sodium salt. There is no visible clue to the explanation of these results. But we must conclude as most probable that some action is at work in the sea water and in the salt solutions which clumps or flocculates the sediment. For only by the gathering of the particles together in little aggregates can we explain their rapid fall to the bottom. It is not a question of viscosity (_i.e._ of resistance to the motion of the particles), for the salt solutions are rather more viscous than the fresh water. Still more remarkable is the fact that every dissolved substance will not bring about the result. Thus if we dissolve sugar in water we find that, if anything, the silt settles more slowly in the sugar solution than in fresh water. Now there is one effect produced by the solution of such salts as we have dealt with which is not produced by such bodies as sugar. The water is rendered a conductor of electricity. Long ago Faraday explained this as due to the presence of free atoms of the dissolved salt in the solution, carrying electric charges. We now speak of the salt as "ionised." That is it is partly split up into ions or free electrified atoms of chlorine, sodium, magnesium, etc., according to the particular salt in solution. This fact leads us to think that these electrified 57 atoms moving about in the solution may be the cause of the clumping or flocculation. Such electrified atoms are absent from the sugar solution: sugar does not become "ionised" when it is dissolved. The suspicion that the free electrified atoms play a part in the phenomenon is strengthened when we recall the remarkable difference in the action of sodium chloride and magnesium chloride. In each of the solutions of these substances there are free chlorine atoms each of which carries a single charge of negative electricity. As these atoms are alike in both solutions the different behaviour of the solutions cannot be due to the chlorine. But the metallic atom is very different in the two cases. The ionised sodium atom is known to be _monad_ or carries but _one_ positive charge; whereas the magnesium atom is _diad_ and carries _two_ positive charges. If, then, we assume that the metallic, positively electrified atom is in each case responsible, we have something to go on. It may be now stated that it has been found by experiment and supported by theory that the clumping power of an ion rises very rapidly with its valency; that is with the number of unit charges associated with it. Thus diads such as magnesium, calcium, barium, etc., are very much more efficient than monads such as sodium, potassium, etc., and again, triads such as aluminium are, similarly, very much more powerful than diad atoms. Here, in short, we have arrived at the active cause of the phenomenon. Its inner mechanism