Note to p. 33. 



With reference to the views in the text it may be said that no congelation could take place till the temperature 

 of the whole ruass was at that of fusiou , hecause unlil (hen there wouLd be a constant interchange of level between 

 the successive upper or denser layers and the lower. Without resorting to the supposition that gradual expansioa 

 as in water may begin prior to congelation , it seems clear that in such a dense fluid as mollen granite under great 

 pressure , the passage of one porlion through anolher must be efl'ected with difficulty and very slowly. Between 

 the level at which the maximum temperature ceases, and ihe rcfrigerating surface , there must, in such a fluid, 

 be an insensible and very gradual diminution of heat, and deep masses may be viewed as consisting of layers of 

 considerable thickness in any one of which the temperature is almost uniform and between which and the adjacent 

 layers the transfer of heat is exceedingly slow. It may be proper to examine Ihis further , because when the text 

 was written I overlooked the chapter in Mr. Lyell's Principles (Ch. XIX of Book II) in which he controverts the 

 doctrine of the internal fluldity of the globe by an argument which is partially opposed to some of the above 

 views, although it does not interfere with my general inferences. lts object is to prove that no consolidation at the 

 surface of a fluid mass like the globe could take place, till the whole had been reduced to a uniform heat or about 

 that of incipient fusion. It appears (note p. 440) that M. Poissos had, independently, adopted the same argument, 

 and that he imagined that if the globe ever passed from a liquid to a solid state by radiation of heat, the central 

 nucleus must have begun to cool and consolidate first. The pi incipal facts on which Mr. Ltbll relies are , that so 

 long as a fragment of ice remains in water the temperature of the water, cannot be raised above 32°, and that 

 Professor Dakiell found that white a solid piece of iron etc. remained immersed in a molten mass of the same sub- 

 stance, its temperature could not be raised above the melling point. The rematks in the text relating to the ori- 

 giu of granitic struclure do not require me to assume that the temperature of a granitic mass was ever much 

 above that of its melting point, a temperature, it should be recollecled, higher by 790° than that at which basalt can 

 exisfc in the fluid state ; and as Mr. Ltell does not seem to deny , with M. Poissoif , the possibility of a crust 

 being formed when the general temperature is near that of fusion, his remarks are not inconsistent with the hypo- 

 thesis that granite has cryslallized in successive layers. Even if the granite fluid approached to the condition of 

 water , instead of being comparalively very dense , the congelation of the first layer would tend to maintain the 

 fluidity of the next for some time, not merely by airesting the interchange of particles by which its heat had 

 thitherto been transferred to the surface, and subslituting for it the process of conduction , but by adding to it 

 the large quantity of latent calorie expelled in the act of congelation. With reference, however, to Mr. Liell's ar- 

 gument, it may be remarked that the condilions to which a fragment of ice or iron are exposod when immersed in 

 n fluid mass of the same substance , and heat constantly added from a source close to it , are very different from 

 ihose of a crust resling on the same fluid , iu contact on one side with a refrigeraling surface , and on the other 

 with one to which calorie slowly ascends from a source far below. The immersed fragment is enveloped in a ra- 

 pidly heating medium. The floating crust is between two media, one slowly giving heat, the other abstracting it. 

 The period . therefore, at which a crust can be formed is not dotermined simply by the fusing point (that is, ne- 

 «essarily postponed until the whole mass has reached its lowest fluid temperature) but is determined by the relative 

 rates at which heat ascends through the substance in its fluid state , and is conducted from it when in its solid 

 state. The latter is an indeterminate quantity, depending , as it in some raeasure does, upon the nature of the refri- 

 geraling body. But , laying that out of view , it is evident that as the heat diminishes, the rate of its passage 

 from one level to another will also decrease, because the motion of the particles of the fluid amongst themselves will 

 be impeded as the density increases. JNow long before the mass generally is reduced to the temperature of fusion, 

 the density of its higher portion may reach the point at which the rato of motion has subsided to that at which 

 heat is given off by a solid crust. Refrigcration may come to a stage at which, while the upper layers are at and 

 near the point of fusion , the lower may be far removed from it in proportion to the depth of the mass. Because 

 in a receptacle of water exposed to a cold atmosphere we see a rapid interchange of particles , and the formation 

 of ice postponed till all the lower layers have reached the temperature of about 40°, we are not entitled to con- 

 clude that in the successive layers of a deep abyss of dense mollen granite there will be a similar rapidity and 

 extent of mutual motion of particles. Even in the case supposed by Mr. Ltell of the globe consisting of water 

 having at the centre a temperature of 6400° which gradually decreased towards the circumference where a crust 

 «>f ice fifty miles in thickness existed , is it necessary that we should admit with Mr. Ltell that the ice would soon 

 melt into an atmostphere of steam ? In a layer a inile in thickness the temperature would not increase one degree. 

 At 8 miles beneath the ice the temperature would be only 40°. Below that the counter currents would first be 

 encountered, but the rate of their motion in a layer a mile in thickness in which the suramit only differed from 

 the base by 1° would be so extremely slow as to be insensible. But , to render ihe cases of a globe of water and 

 a globe of granite more analogous , the globe of water should be supplied with a refrigeraling atmosphere at a 

 temperature more than 2600° below zero , for such is the difference of temperature between that of the atmosphere 

 (even taktng that of the equator) and the fusing point of granite. With such a medium , or even one of which 

 the temperature was calculated according to the fusing point of trappean rocks , the heat given off on the upper 

 side of the icy crust might exceed that received from below even with a central temperature of 6400 ó . 



