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



Comparison of analyses means not much unless one knows the 

 method of choosing the samples ; no definite conclusions respecting 

 conditions under which coal was deposited can be based on a mass 

 of analyses gathered indiscriminately from all quarters of the globe. 

 It might be that the ash would tell much, if all portions of a bed 

 were studied, each by itself ; but even then the information might be 

 too localized. Another difficulty is that published analyses, with a 

 small proportion of exceptions, are of coal supposed to have com- 

 mercial value, so that they do not give a proper conception of the 

 character of the greater part of coal beds. Thus, Stainier^"" com- 

 pared 2,568 analyses, gathered from reports on coal areas in Europe 

 and America. Of these only 15 per cent, showed more than 10 per 

 cent, and less than 2 per cent, had more than 20 per cent, of ash. 

 It is very certain that the 2,250 analyses, giving less than 10 per cent., 

 were not all made of prisms representing the whole bed ; and equally 

 that the coal was taken from localities which were promising from 

 the commercial standpoint. This is beyond dispute, since more than 

 one half of the analyses report 5 per cent, or less of ash. It is 

 unsafe to take the average of such analyses as representing a prob- 

 able average condition. Most of the coal horizons show extreme 

 variations which at times are abrupt, so that, while a sample from 

 one locality may have but 5 per cent., another, only a short distance 

 away, may have 25 per cent. It is quite probable that if analyses 

 were made of all the coal beds in some small areas of southwestern 

 Pennsylvania, where the column is long, the results would show that 

 more than half of the beds have more than 10 per cent, of ash. One 

 must recognize that in many localities the conditions did not favor 

 the accumulation of clean coal ; in the higher portions of the Coal 

 Measures, within the bituminous region, the beds are all poor, broken 

 by thin slates, no analysis showing less than 12 per cent, and most of 

 them above 16 up to almost 32. 



The difference in ash-content of the benches of a coal bed may be 

 very great. A. S. McCreath's analyses of several beds in Pennsyl- 

 vania show, for those in the Allegheny, differences between the 



^"' X. Stainier, " Notes sur la formation des couches de charbon," Bull. 

 Soc. Beige de GcoL, Vol. XXV., 191 1, P. V., pp. 73-91. 



