Even in those rocks which merely supply springs, the amount of disseminated water must be enormous ; for they so far resemble filters, that are necessarily charged with the fluid before they permit it to pass out.

De La Beche has advanced the opinion that capillary attraction has great power, both in mechanically disseminating water among rocks, and in retaining it in them when so disseminated, and that it therefore keeps them, to a certain extent, saturated with moisture, and assists in promoting a more equal flow of water in springs.

Capillary attraction and gravity probably carry water down far beyond those situations where it can be returned in springs, at least cold springs, for there are certain

circumstances connected with those which are thermal, which go to prove, that the water thrown up by them may have percolated to considerable depths. It is very

evident that most rocks contain disseminated moisture, for there are few which,

when exposed to heat, do not give water. Sulphate of lime, for example, or plaster of paris, contains about 20 per cent., and common serpentine, as much as 15 per cent, of it. Soap-stone has 4 per cent,, and even quartz 2 per cent, of water, in their composition. This fluid exists in minerals either as water of crystallization, or combined as a hydrate.

But though water is thus generally diffused over the surface of the globe, yet it is not found perfectly pure in any place ; even the rain and the snow that descend from the clouds, the condensation, as it were, of a natural distillation, are slightly tainted by saline matters ; which circumstance can only arise from the great solvent power of water enabling it to take up a portion of most substances with which it comes into contact, in its natural condition.