356 MERRILL — DISINTEGRATION AND DECOMPOSITION OF DIABASE. 



small amount of organic matter. This increase, it should be stated, is 

 greater than may at first be apparent, for the reason that the fresh rock 

 contains a considerable amount of secondary calcite, which is quite lack- 

 ing in the residual sand. A large part of the ignition in columns 1 and 

 8 is therefore to be accredited to carbonic acid and not to water of h^ydra- 

 tion. 



We have as yet, however, fallen far short of the desired results, and 

 in columns 11 and 12 I have attempted to show the actual percentage 

 amounts of the original constituents that have becii saved or lost during 

 this process of decomposition. In these calculations, as before, the alu- 

 mina is considered a constant factor. The method of calculation is not 

 new, or at least is so onl}^ in its mode of application. A similar method 

 was employed by Dr R. A. L. Penrose in considering the origin of manga- 

 nese deposits through the decom[)osition of manganiferous limestones,* 

 though in his case the silica was considered the constant factor.f 



From the figures thus obtained it api:)ears that of all the essential con- 

 stituents the lime and potash salts have suffered the most, though the iron 

 oxides have also been carried away in amounts equaling 18.10 per cent 

 of their original percentages. ^Magnesia has also proven ver}'' suscei)tible 

 to the solvent action, disappearing to the amount of 21.70 per cent; and, 

 lastly, silica to the amount of 18.08 per cent. The small original amounts 

 of manganese and j^hosphoric acid render the results obtained by these 

 calculations as of doubtful value, since it is possible they may be due in 

 part to errors of analysis. 



In order to further test the method I have recalculated analyses of 

 (1) fresh and highly weathered diabase from Venezuela, as given by G. 

 Atwood,J and also (2) the granitic analyses given by myself in the pub- 

 lication already alluded to.§ In both cases the iron (Fe.^Og) was found 

 to be the most stable factor. The results thus obtained are given on the 

 opposite ])age. 



From these tables it appears that in the case of diabase the total loss 

 equals 39.51 per cent, with an actual gain in volatile matter. Of the 

 individual constituents, 83.23 per cent of the original lime ; 61.37 per cent 



*Ann. Rep. Geol. Survey of Arkansas, vol. 1, 1890. 



fThe formulae employed in the calculation is as follows: 



£^C ~ ^'' ^"^ 100 — X = F, in which A = the percentage of any given constituent in the re- 

 sidual material ; B = the percentage of the same constituent in the fresh rock, and C= the quotient 

 obtained by dividing the alumina of the residue by that of the fresh rock (in this particular case 

 1.155), the final quotient being multiplied by 100. X then equals the percentage of the original 

 constituent saved in the residue and Fthe percentage of the same constituent lost. 



I Quart. Jour. Geol. Soc. of London, vol. xxxv, 1879, part 2, p. 586. 



g In this paper, it will he remembered, the writer contented himself with calculating the results 

 of analyses back to a water-free basis, whereby an "apparent loss" of only some 3 per cent of 

 material was indicated as against 14.93 per cent by this method. 



