88 SANDSTONE CONCRETIONS— MERRILL. 



to be attacked by atmospberic agencies, tbey have undergone rapid 

 decomposition. Wbeu lying on the immediate surface tbis decompo- 

 sition (so far as the sulpbide is concerned) consists mainly in tbe pro- 

 duction of sulphates which are rapidly removed in solution, or wbicb 

 during the dry part of the year accumulate in the form of a thin, sul- 

 phur-yellow coating on the surrounding surface. When, however, buried 

 in the loose siliceous sand the result is noticeably different. Here, owing 

 presumably to an insufficient supply of sulphuric acid, a considerable 

 portion of the sulpbide passes into the condition of sesquioxide, whicli 

 segregates in a narrow zone about the nucleal pyrite, cementing together 

 the granules of siliceous sand and forming a crust or shell-like coating 

 which is often quite dense and hard. All stages of the process are to be 

 found, from those in which there is merely a thin crust of oxide (figs. 

 ij and C) to those in which the sulphide has nearly disappeared (fig. 7). 

 As the original concretionary mass rarely consisted of pure pyrite, but 

 inclosed more or less sandy material, this last becomes liberated and 

 not infrequently remains as loose sand partially filling the geode-like 

 cavity. 



The chemical processes involved in this change are presumably sim- 

 ple, though as we do not know for a certainty the exact conditions 

 attending either solution or precipitation we can not be expected to 

 describe them in detail. On the assumption that the iron was orig- 

 inally in solution as a ferrous sulphate, we can readily account for tbe 

 presence of the pyrite concretions through the reducing action of 

 gases given off by the decomposing wood. If, however, the iron 

 existed, as at first seemed more probable, as a ferrous carbonate, the 

 precipitation is less readily accounted for, since it seems doubtful if 

 the small amount of sulphuretted hydrogen liberated would be suffi- 

 cient for the production of so large a (juantity of pyrite as is here 

 found. 



EXPLANATION OF PLATE VL 



Figs. 1,2, 3 and 4. Characteristic forms of concretions formed of granules of sili- 

 ceous sand cemented by marcasite. In fig. 4, there has been internal shrinkage, caus- 

 ing cracks suggestive of an intermediate stage in the formation of septaria. 



Figs. 5, 6 and 7. Noduk'S showing stages of oxidation. In fig. 5, the oxidation has 

 barely commenced, giving a red brown coating perhaps one-eighth inch thick on 

 the outer surface. This coating has been removed from the lower end, exposing the 

 marcasite. In fig. 6, the nucleal mass carries so much sand as to be distinctly 

 granular, but the line of demarkation between the oxidized and unoxidized portion 

 is plainly evident. In fig. 7, the loose sand fell away in process of cutting, leaving 

 the unoxidized portion as shown. 



