80 STEVENSON— FORMATION OF COAL BEDS. [April i8, 



as much dolomite as calcite while in three others dolomite is absent 

 or insignificant. Dolomite is unimportant in the Yorkshire balls, 

 but it is in large proportion in Stur's specimen, while it is shown in 

 small proportion by the roof balls of the same area. It would ap- 

 pear that sea-water can hardly be regarded as the source, in view of 

 the marked variations found within short distances. And this sug- 

 gestion is strengthened by the fact that Gothan's Torf-dolomite 

 closely resembles in composition the coal balls described by Stur. 



It may be preferable to seek the source in the materials them- 

 selves, the inorganic matter forming the shales and the ash of the 

 coals. Carbonate of magnesium is found in most of the coals as 

 well as in peats and it is often an important constituent of wood 

 ash. The varying proportion in the balls may indicate merely a 

 varying proportion in the shales, depending on the nature of the 

 rock whence they were derived. And this seems to be reasonable, 

 when one considers the composition of limestones. McCreath'^^ 

 made many analyses for the Pennsylvania survey, which illustrate 

 the conditions. The Vanport limestone of the Allegheny formation 

 is of marine origin throughout and is one of the widely extended 

 deposits. Carbonate of magnesium rarely exceeds 2 per cent, and 

 very often is less than i per cent. ; but on the northern border, where 

 it extends into old valleys and is mingled with land material, the 

 percentage increases, attaining 6.65 at one locality. A similar 

 change appears in the Ames limestone. In Harrison county of West 

 Virginia, that limestone is approaching its southern limit as a marine 

 deposit. It contains in its upper division 25 per cent, of alumina 

 and in the lower, 18 per cent, of silica. The influx of land ma- 

 terial is very marked, though the marine fossils persist in great 

 numbers; the carbonate of calcium varies from 40 to 48 per cent, 

 and carbonate of magnesium from 15 to 21 per cent.'^*' McCreath's 

 analyses of Monongahela and other limestones, which from their 

 relations must be regarded as non-marine, show that in some cases 

 they are markedly dolomitic and with few exceptions they have a 



"A. S. McCreath, Sec. Geol. Surv. Penn., Rep. MM, 1879, pp. 281-362; 

 Rep. M:^, 1881, pp. 79-94. 



" B. W. Hite, in West Va. Geol. Surv., County Reps., 1912, p. 251. 



