MIX ERA L TRANSFORMATIONS. 277 



be cited to illustrate the doubt which often prevails as to the 

 original source of pseudomorphs. 



The process by which they have been produced is often 

 even more uncertain ; and yet those who have attempted a 

 systematic treatment of pseudomorphs have usually devoted 

 their attention mainly to the processes by which they may 

 have been formed, with the object of elucidating the pro- 

 blems of chemical geology. 



The two best known and most exhaustive treatises on 

 the subject are the elaborate work of Blum, Die Psendo- 

 morphoscu des Miner air etc ks, 1843 (with four appendices, 

 1 847-1879), and the first volume of Roth's Allgemeine und 

 Chemische Geologie, 1879, in which all the previously de- 

 scribed pseudomorphs are collected and classified. 



Thus in the classification of Blum, they are considered 

 according to the relation between the initial and final sub- 

 stance as due ( 1 ) to loss of constituents, as when Gypsum 

 is converted into Anhydrite by dehydration ; (2) to gain of 

 constituents, as when Anhydrite is converted into Gypsum 

 by hydration ; (3) to exchange of constituents, as when 

 barium carbonate is converted into sulphate ; and (4) to 

 total replacement of one substance by another. 



la the classification of Roth, a distinction is drawn be- 

 tween those which are due to the simple and direct action 

 of air, water and carbon dioxide, and those which result 

 from the action of solutions containing mineral matter. 



In the present state of our knowledge there is here 

 necessarily much that is mere speculation, since, even if the 

 initial and final products are certain, the intermediate steps 

 and processes are rarely known to us. 



Taking these and similar difficulties into account, it is 

 clearly necessary in any arguments based upon pseudomorphs 

 to select only those cases in which there is absolutely no 

 doubt as to the nature of the original mineral. A complete 

 revision of the work of Blum and Roth is sorely needed, if 

 only to correct downright errors and to weed the doubtful 

 from the certain. 



In no application of such arguments is this more neces- 

 sary than in those speculations with which the present 



