We have now added the pressure of the 2 miles of water, because there could be no water at the depth of 9 miles; for the critical temperature of water is known to be 412°, beyond which temperature water cannot be maintained in its liquid state by any amount of pressure, however great; and 9 miles would give 483° temperature at 1° for each 30 metres. At that depth there might be steam, although it is difficult to see how it could penetrate so far, because the only force to help it to penetrate would be gravitation, and that would have to act against the increasing repulsion of heat. There is another circumstance to be considered which would tend to increase the density of the outer portion of the crust, if there be a crust, and if not, of the outer portion of the earth itself. When the earth was in the molten liquid state, it is generally supposed to have been surrounded by vapours of a great proportion of the metals and of some of the metalloids, in addition to the vapour of water, air, and other gases, which floated above them higher up in the atmosphere. In that case when the crust began to be formed through cooling, these vapours would be precipitated on the surface and mixed with the half-liquid half-solid matter there, but the proportion of condensed vapours would be very small compared with what they fell upon, and the specific gravity of the mixture would not be great enough to cause it to sink much below the surface, because it would soon meet with matter as dense as itself; consequently we must consider that all these metals would remain near the surface--most likely much nearer to it than the 9 miles which we have as yet descended to--and whatever may have been the proportion of their density it ought to be added to the weights and pressures that have been taken into account above. We believe that it will be shown later on that this estimate of a density of three times that of water at 9 miles deep in the earth is very much lower than it should be; because, when the pressure upon the matter there came to be greater than its crushing strain, compression would go on more rapidly than shortly afterwards, and it might so be that with a strain of very much less than four times that of crushing, compression would be reduced to its utmost limit. But more of this hereafter.