462 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. 



compensated for by the unequal expansion of the mercury of the 

 thermometer itself, by the expansion of the glass tube, or the com- 

 pression of the liquid by the rapidly increasing tension of the vapour, 

 that their value up to 200° C. is so small that they do not produce 

 any clearly marked effect in the results. The action of water on 

 glass at temperatures above 200° C. is so powerful that I was unable 

 to determine the volume at a greater heat ; but from 25° to 200° 

 there is such a very close agreement between calculation and expe- 

 riment that I think the formula would give sufficiently accurate 

 approximations to the truth up to at least 300°. The law does not 

 hold good near the freezing-point ; but that is of no consequence in 

 the present inquiry. The following are the actual values for water 

 and various saline solutions, the volume at 0° C. being taken for 

 unity : — 



1. Water V= •0001344*+-000003245* 2 



2. „ with 10 p. c. of chloride of potassium .. V = -0001868*+-000002524* 2 



3. „ „ 25 p. c. of chloride of potassium.... V = '0003006H-'000001410* 2 



4. „ „ 12* p. c. of chloride of potassium+l v = . 0003280/+ . 0000()1422<2 



12f p. c. of chloride of sodium.. J 



5. „ „ 25 p. c. of chloride of sodium V=-0003520H--000001370* 2 



6. „ „ 25 p. c. of sulphate of soda V = '0003077*-p000001644rf 2 



7. „ „ 25 p. c. of mixed salts, being the! y== . 00 03221H-000001461<* 



mean of 3, 4, 5, & 6 J ' 



On examining this table, it will be seen that the addition of salts 

 to the water increases the value of the coefficient of t, but decreases 

 that of t 2 ; thus causing the expansion to be more uniform, by 

 making it greater at low, and at the same time considerably less at 

 high, temperatures. As far as I have been able to ascertain, the 

 most trustworthy experiments indicate that the product of these 

 coefficients is nearly constant, and that the increase in the value of 

 the coefficient of t varies in simple proportion to the quantity of salt 

 in the solution, the amount of water being taken as constant. Pro- 

 bably these laws are not strictly correct, but still sufficiently so to 

 enable us to determine the effect of variation in the strength of the 

 solution as accurately as is requisite for the purpose of the present 

 paper. 



d. Effects of pressure. 



These conclusions of course apply to those cases where the pressure 

 to which the liquid is exposed is only equal to the elastic force of 

 the vapour ; but since, in nature's laboratory, crystals have no doubt 

 often been formed under very great pressure, it will be necessary to 

 take into account the compression of the fluid. The amount of this 

 has been determined with great accuracy by a number of observers 

 (Gmelin's Handbook of Chemistry, Cavendish Society's Translation, 

 vol. ii. p. 62) ; but I would particularly refer to the paper of M. 

 Grassi (Annales de Chimie, t. xxxi. p. 437). He there shows that 

 the amount of compression of pure water, for a pressure equal to 

 one atmosphere, is '0000502, but decreases as the temperature in- 

 creases. It is less for saline solutions, but increases as the tempera- 



