As suggested above, it is probable that the extent of the zones of compression is determined by another factor in addition to the great development of surface. That factor is connected if not identical with that power which makes the system lyophile, and is evidently connected also with the solubility of the disperse phase, and may indeed be electrochemical forces tending to form a series of hydrates, or at least to cause an orientation or definite arrangements of the water molecules in the zone of compression. This idea receives some support from the hydrate theory of solution, and the zones of compression and orientation are the colloid analogue of the hydrates supposed to exist in solutions of electrolytes. The extension of such zones on cooling are then analogous with the series of hydrates formed, for instance, by manganese chloride with 2, 4, 6, 11, or 12 molecules of water when crystallized at temperatures of 20°, 15°, -21°, -30°, and -48° C. respectively, the idea being that the salts most hydrated in solution crystallize with most water. As the compression is the result of two factors, one of which depends upon the nature of the disperse phase, we expect--and find--in other lyophile systems a considerable variation in their power of gelation. Some indeed, though very viscous, _e.g._ egg albumin, never quite set like gelatine, and others (_e.g._ agar-agar) set to a stiff gel from a much weaker sol than gelatine. When the zones of compression are large, as in gelatine, the magnitude of the compressing force on the outermost part of the zone is relatively small, and it is not surprising that time is necessary for the victory of this force over the kinetic energy of the water molecules. Hence we find a 5 per cent. jelly sets readily on cooling, but its elasticity increases steadily for many hours after it has set. This phenomenon, known as hysteresis, we should expect--and find--to be much more marked in a case where the zone of compression is unusually large (_e.g._ an agar gel). We should also expect--and find--that hysteresis is more marked in a high-grade gelatine than in a low-grade gelatine where both eventually form gels of equal elasticity. We should expect too--and we find--that hysteresis is more prominent in weak gels than in strong. These points are of obvious importance in testing gelatine by its elasticity, _e.g._ the well-known "finger test." There are also other facts and considerations which have an important bearing upon the point under discussion. It is necessary ultimately to regard true solutions of electrolytes and other bodies as heterogeneous, though perhaps of a rather different order. From this point of view molecules and ions existing in an aqueous solution will present a surface and have associated zones of compression analogous with those suggested for the minute particles of gelatine.