I9I3-] STEVENSON— FORMATION OF COAL BEDS. 79 



rule, however, the plant remains are less well preserved than in the 

 coal .balls ; they are so disintegrated that in many cases they are not 

 identifiable. Roots are best preserved, probably because they entered 

 when the surrounding mass had already become peat. They show 

 no trace whatever of compression. Some fragments of stems have 

 great lacunae, indicating that they are of plants belonging to a moist 

 habitat. The great variety in the plants suggests that the deposit is 

 a typical Waldtorf, which accords with the belief that the brown 

 coals were deposited as Waldmoors. 



This conclusion is very similar to that reached by Stopes and 

 Watson, who recognize a swamp vegetation in the coal balls, as, 

 indeed, Stur had done long before. Stur had noted the difference 

 in condition of the vegetable material in the two types of balls, and 

 this difference is emphasized by Stopes and Watson. Scott had ob- 

 served that the roof-ball flora, though of Lower Coal Measures age, 

 has no slight resemblance to that of the Permian, and those authors 

 think that it is comparable to an upland flora, so that it may be more 

 characteristic of the widespread vegetation than is that of the coal 

 balls. 



Dolomite, calcite and carbonate of iron are not the only minerals 

 replacing plant material in concretionary fashion. E. B. Andrews 

 and Lesquereux found wood in coal wholly replaced with sulphide 

 of iron, the form being uncompressed ; but no microscopic study was 

 made to ascertain whether or not any trace of structure remained. 

 The Grand'Croix flints are of the same type as the coal balls and they 

 yielded interesting results to Renault, who recognized that they are 

 petrified peat. Near Salem in Oregon there are fossil stems, w'hich 

 show all gradations between lignite and silica within a few feet. 



The source of the material forming the balls has been subject for 

 speculation. Balls from the more celebrated localities are in coal 

 beds with marine deposits as the roof. Binney thought the shells 

 provided the material, but objection was made that the shells are 

 not dolomite and that one should look to sea-water as the source. 

 As the roof shales in the Coal Measures localities are marine, sea- 

 water must have covered them all alike ; yet in Lancashire, the balls 

 are dolomite at Bacup, whereas at Shore only one specimen showed 



