THE ONYX MARBLES. 559 



that the most constant distinction between those of the onyx marbles 

 which are spring deposits and those which are formed in caves is the 

 absence, in the hitter, of appreciable quantities of metallic oxides. This 

 is presumably' to be accounted for on the supposition that the cave mar- 

 bles result from the solvent action of cold carbonated water on limestone 

 containing, aside from the iron oxides, only mechanically included 

 impurities which do not enter at all into solution, but remain in the form 

 of the ochreous residual clays which are so characteristic of limestone 

 caverns the world over. The travertines, on the other hand, result from 

 the solvent action of heated solutions on deep-seated siliceous rock, and 

 which as a result carry not merely lime, but a considerable proportion 

 of rarer constituents as well. That these rarer constituents are not 

 more abundant in the deposits themselves is due to their unequal solu- 

 l)ility and the consequent fractional separation which takes place on 

 evaporation. This separation has already been alluded to, on p. 549. 

 The reverse of the above-stated rule does not always hold good, since, 

 as above noted, the Suisin City deposit, which is a travertine, contains 

 scarcely a trace of iron. The percentage of manganese, as will be 

 noticed, is, with but two exceptions less than one-half of one per cent. 

 These exceptions are (I) a faintly greenish stone from Lower California, 

 and (-) a pure milk white variety from Lake Oroomiah, the latter yield 

 ing 4.34 per cent, of this material calculated as a carbonate, (MnCOs) 

 or 2.68 per cent, when calculated as oxide (MnO). The magnesium car- 

 bonate is almost invariably present in small amounts, and singularly 

 enough is highest in the cold-water (cave) deposit from Syout, Egypt 

 (01336), where it reaches 0.88 per cent. Careful tests were made for the 

 rarer elements, but with negative results in the most cases, the Suisin 

 City stone showing 1.59 per cent, of strontium carbonate; that from the 

 Hacienda del Carmen, Mexico (01337), 1.34 per cent, of calcium sulphate, 

 and that from San Luis Obispo 0.25 per cent, of tricalcic phosphate, 

 Ca:,(P04)2. The milk white stone from Lake Oroomiah yielded 2.30 per 

 cent, of calcium sulphate and 0.24 per cent, of tricalcic phosphate. 



In order to illustrate the possible changes in color from secondary 

 oxidation, pis. 5, 0, and 7, are given. The slabs were all sawn from the 

 same block, not above 10 inches in thickness, and which was at first 

 supposed to be nearly uniform throughout. In pi. 5, it will be observed, 

 there is a wide vein of ocherous brown extending somewhat diagonally 

 from top to bottom, with smaller veins cutting it from left to right. In 

 the second slab the vein has extended so as to include the whole upper 

 left section, while in the third the original green has been almost wholly 

 obliterated. It is easy to perceive that this change in color has been 

 brought about wholly through the oxidizing influence of percolating 

 solutions which followed the lines of existing flaws. In no case that 

 has thus far come under my observation am I inclined to regard these 

 veins and dashes of red and brown as original, but as results of secon- 

 dary oxidation. 



