From a further discussion of the mathematical condition under which the subsidence of the particles takes place, Barus is of the opinion that Durham’s theory of suspension being only a lower limit of solution is rapidly gaining ground, yet without being attended with concise experimental evidence which will account for the differences in the rate of subsidence. On the contrary, Brewer’s hypothesis of colloidal hydrates is more easily subjected to experimental proof. The test shows that the particles retain their normal density, no matter how they are suspended or circumstanced. Further, in the explanation of the phenomenon of sedimentation, the following principle may be regarded as determined; namely, if particles of a comminuted solid are shaken up in a liquid, the distribution of parts after shaking will tend to take place in such a way that the potential energy of the system of solid particles and liquid, at every stage of subsidence, is the minimum compatible with the given conditions. According to Barus it is necessary, in order to pass judgment on the validity of any of the given hypotheses, to have in hand better statistics of the size of the particles relatively to the water molecule, than are now available. Inasmuch as the particles in pure water are individualized and granular, it is apparently at once permissible to infer the size of the particles from the observed rates of subsidence. His observations show that the said rate decreases in marked degree with the turbidity of the mixture. Hence the known formulæ for single particles are not rigorously applicable, though it cannot be asserted whether the cause of discrepancy is physical or mathematical in kind. It follows that special deductions must be made for the subsidence of stated groups of particles before an estimate of their mean size can fairly be obtained. Rowland[127] reaches a closer approximation for the fall of a single particle by showing that the liquid, even at a large distance from the particle, is not at rest. In the case of water, however, it is noticed that despite the large surface energy of the liquid, subsidence takes place in such a way that for a given mass of suspended sediment the surfaces of separation are a maximum. On the other hand, in case of subsidence in ether or in salt solutions, the solid particles behave much like the capillary spherules of a heavy liquid shaken up in a lighter liquid with which it does not mix. In other words, the tendency here is to reduce surfaces of separation to the least possible value, large particles growing in mass and bulk mechanically at the expense of smaller particles; in other words, exhibiting the phenomenon of flocculation.