24 THE RIGHT HON. LORD KELVIN, G.C.V.O., ON 



few hours or days, or any longer time, is taken, the sohd 

 formed lias the well known rough crystalline structure ot 

 basaltic rocks found in all parts of the world. Now Carl 

 Barus finds that basaltic diabase is 14 per cent, denser than 

 melted diabase, and 10 per cent, denser than the glass pro- 

 duced by quick freezing of the liquid. He gives no 

 data, nor do Rucker and Roberts- Austen, who have also 

 experimented on the thermodynamic properties of melted 

 basalt, give any data, as to the latent heat evolved in the 

 consolidation of li(][uid lava into rock of basaltic quality. 

 Guessing it as three times the latent heat of fusion of the 

 diabase pitch-stone, I estimate a million cubic centimetres ot 

 liquid frozen per square centimetre per centimetre per three 

 years. This would diminish the depth of tlie liquid at the 

 rate of a million centimetres per three years, or 40 kilometres 

 in twelve years. 



§ 25. Let us now consider in what manner tin's diminution 

 of depth of the lava ocean must have proceeded, by tlie 

 freezing of portions of it ; all having been at temperatures 

 very little below the assumed 1420'^ melting temperature of 

 the bottom, when the depth was 40 kilometres. The loss of 

 heat from the white-hot surface (temperatures from 1420° to 

 perhaps 1380° m difteicut parts) at our assumed rate of 2 

 (gramme-water Centigrade) thermal units per sq. cm. per 

 sec. produces very rapid cooling of the liquid within a few 

 centimetres of the surface (thermal capacity -36 per gramme, 

 according to Barus) and in consequence great downward 

 rushes of this cooled liquid, and iqDwards of hot liquid, 

 spreading out horizontally in all directions when it reaches 

 the surface. When the sinking liquid gets within perhape 

 2Q or 10 or 5 kilometres of the bottom, its temperature* 

 becomes the freezing point as raised by the increased 

 pressure ; <)r, perhaps more c<n'rectly stated, a temperature 

 at which some of its ingredients crystallise out of it. Hence, 

 beginning a few kilometres above the bottom, we have a 

 snow shower of S(4idiiied lava or of crystalline flakes, or 

 prisms, or granules of felspar, mica, hornblende, (puirtz. and 

 other ingredients: each little crystal gaining mass and falling 

 somewhat faster than tlie descending liquid ai'oimd it, till it 



* The, temperature of the sinkius,' ]iqui4 rork ri-^es in virtue of the 

 increasing pressure: but much 1 ^s,s than doe.-^ the fieezing ])oint of the 

 liquid or of some of its ingredients. (See Kelvin, Mitiii. ciid I'hys. 

 Papers, vol. iii, pp. G!), 70.) 



