354 MERRILL — DISINTEGRATIOX AND DECOMPOSITION OF DIABASE. 



hydrochloric acid and sodium carbonate sokitions. The insohible por- 

 tion, constituting 22.13 per cent of the silt, consists of unaltered feldspars 

 and iron, lime and magnesian silicates, which are easil}'' recognizable 

 under the microscope in minute, sliarply angular particles. This portion 

 of the separation is of interest as showing the minute stage of subdivision 

 attainable without complete decomposition in deposits where no mechan- 

 ical forces are operative other than those involved in hydration and tem- 

 perature changes. The bearing of this upon the origin of the materials 

 forming loess and such secondary rocks as the shales and finer sand- 

 stones, I cannot but regard as of some importance.* 



It is, however, obvious to all who have given the subject attention that 

 chemical analyses alone, as ordinarily set forth, convey a very imperfect 

 idea of the actual changes which have gone on in the {)rocess of weather- 

 ing. Indeed, the proper interpretation of such analyses is attended with 

 very great difficulty. In the present instance it is de3iral)le to learn, so 

 far as possible, not merely tlie difference in composition between the 

 fresh and altered rock, but, what is of greater importance, the propor- 

 tional loss or gain of the various constituent elements, as well as the 

 change in their method of coml)ination. The problem is rendered one of 

 ver>' great difficulty from the fact that we do not know with any degree 

 of precision just what may have been the actual processes involved. 

 From the manner in Avhich the decomposed rock conducts itself toward 

 solvents and from a study of the percentage figures in the table, it is, 

 however, at once evident that hydration is aU'important factor, and that 

 a very appreciable amount of material has been carried away in solu- 

 tion ; also, that of the material remaining a consideral)le amount exists 

 in combinations quite different from those in the fresh rock. These are 

 facts that have long been recognized. 



In seeking to determine what proportion of material has actually be- 

 come lost, it is necessary to select some one of the constituents which 

 shall serve as a means of comparison. It is self-evident that the con- 

 stituent thus selected should be the one which sutlers least change, or a 

 known change, during the process of decomposition. 



It is a commonly accepted fact that of all those constituents occurring 

 in considerable quantities in silicious, crystalline rocks, the alumina is 

 the most refractory and least liable to be carried away through the sol- 

 vent action of water. Hence Justus Roth f and others have not infre- 

 quently selected this as the constant factor, and by recalculating the re- 

 sults of analyses on the basis that all the alumina of the original rock 



*Tlie petrology of the clastic rocks must evidently begin with a study of the products of rock 

 degeneration, 

 t Allegemeiue u. Chemische Geologie, vol. 3, 1893. 



