.THE INSTABILITY OF THE HOMOGENEOUS. 423 



component element, are combinations of the simplest order 

 are but one degree less homogeneous than the elements 

 themselves. More heterogeneous than these, more decom- 

 posable by heat, and therefore later in the Earth's history, 

 are the deutoxides, tritoxides, peroxides, &c. ; in which two, 

 three, four, or more atoms of oxygen are united with one 

 atom of metal or other base. Still less able to resist heat, 

 are the salts; which present us with compound atoms each 

 made up of five, six, seven, eight, ten, twelve, or more atoms, 

 of three, if not more, kinds. Then there are the hydrated 

 salts, of a yet greater heterogeneity, which undergo par- 

 tial decomposition at much lower temperatures. After 

 them come the further-complicated supersalts and double 

 salts, having a stability again decreased; and so through- 

 out. After making a few unimportant qualifications de- 

 manded by peculiar affinities, I believe no chemist will deny 

 it to be a general law of these inorganic combinations 

 that, other things equal, the stability decreases as the com- 

 plexity increases. And then when we pass to the com- 

 pounds that make up organic bodies, we find this general 

 law still further exemplified : we find much greater complex- 

 ity and much less stability. An atom of albumen, for in- 

 stance, consists of 482 ultimate atoms of five different kinds. 

 Fibrine, still more intricate in constitution, contains in each 

 atom, 298 atoms of carbon, 49 of nitrogen, 2 of sulphur, 

 228 of hydrogen, and 92 of oxygen in all, 660 atoms; or, 

 more strictly speaking equivalents. And these two sub- 

 stances are so unstable as to decompose at quite moderate 

 temperatures; as that to which the outside of a joint of roast 

 meat is exposed. Possibly it will be objected that some inor- 

 ganic compounds, as phosphuretted hydrogen and chloride 

 of nitrogen, are more decomposable than most organic com- 

 pounds. This is true. But the admission may be made 

 without damage to the argument. The proposition is not 

 that all simple combinations are more fixed than all complex 

 ones. To establish our inference it is necessary only to show 



