84 Lord Kelvin on the Age of the 



the liquor is less, by something like 15 per cent., than the 

 specific gravity of the solid crystals, it must tend to find its 

 way upwards, and will actually do so, however slowly, until 

 stopped by the already solidified impermeable crust, or until 

 itself becomes solid on account of loss of heat by conduction 

 outwards. If the upper crust were everywhere continuous 

 and perfectly rigid the mother liquor must, inevitably, if 

 sufficient time be given, find its way to the highest places of 

 the lower boundary of the crust, and there form gigantic 

 pockets of liquid lava tending to break the crust above it 

 and burst up through it. 



§ 37. But in reality the upper crust cannot have been 

 infinitely strong ; and, judging alone from what we know of 

 properties of matter, we should expect gigantic cracks to 

 occur from time to time in the upper crust tending to shrink 

 as it cools and prevented from lateral shrinkage by the non- 

 shrinking uncooled solid below it. When any such crack 

 extends downwards as far as a pocket of mother liquor 

 underlying the wholly solidified crust, we should have an 

 outburst of trap rock or of volcanic lava just such as have 

 been discovered by geologists in great abundance in many 

 parts of the world. We might even have comparatively 

 small portions of high plateaus of the primitive solid earth 

 raised still higher by outbursts of the mother liquor squeezed 

 out from below them in virtue of the pressure of large sur- 

 rounding portions of the superincumbent crust. In any such 

 action, due to purely gravitational energy, the centre of 

 gravity of all the material concerned must sink, although 

 portions of the matter may be raised to greater heights ; but 

 we must leave these large questions of geological dynamics, 

 having been only brought to think of them at all just now by 

 our consideration of the earth, antecedent to life upon it. 



§ 38. The temperature to which the earth's surface cooled 

 within a few years after the solidification reached it, must 

 have been, as it is now, such that the temperature at which 

 heat radiated into space during the night exceeds that re- 

 ceived from the sun during the day, by the small difference 

 due to heat conducted outwards from within *, One year 



* Suppose, for example, the cooling and thickening of the upper crust 

 has proceeded so far, that at the surface and therefore approximately for 

 a few decimetres below the surface, the rate of augmentation of tem- 

 perature downwards is one degree per centimetre. Taking as a rough 

 average -005 c.G.s. as the thermal conductivity of the surface rock, we 

 should have for the heat conducted outwards •005 of a gramme water 

 thermal unit centigrade per sq. cm. per sec. (Kelvin, Math, and Phys. 

 Papers, vol. iii. p. 226). Hence if (ibid. p. 223) we take -g-^ as the 



