470 STEVENSON— THE FORMATION OF COAL BEDS. [Nov. i, 



probably an Allegheny sandstone on the same river, was not less 

 than 50 feet long. As a rule, the fragments mentioned by observers 

 are much smaller, rarely exceeding 7 or 8 feet ; they are broken at 

 the ends, without trace of root or branch ; some appear to have 

 retained their bark when entombed, but, in a great proportion of 

 the instances, the rock contains indeterminate casts of the stem 

 with scattered impressions of the bark. The battered fragments 

 are unquestionably those of floating wood which had long endured 

 exposure, while those which show no injury, aside from loss of 

 roots and branches, were probably prostrate stems carried from a 

 flooded plain. There seems to be no general distribution of stems 

 in any sandstone ; in each case the occurrence is noted by the ob- 

 server as an interesting local phenomenon. If the materials had 

 been carried out to a great basin, where the plant remains could 

 float until water-logged, there would have been a general distribu- 

 tion over a wide area, such as that described by A. Agassiz as 

 existing off the coast of Lower California and in the Caribbean; 

 but the condition is wholly different in the Coal Measures sand- 

 stones and bears close resemblance to that observed in the rivers of 

 this day, where logs and undermined trees are stranded on the 

 banks or on gravelly islands to be attacked by successive floods and, 

 after removal of the fragile portions, to be buried in the accumulat- 

 ing deposits. 



The less injured fragments, retaining the bark, differ much 

 from those found in coal beds in that commonly they are but 

 slightly deformed by pressure- — and this in spite of the fact that' 

 ordinarily they are merely casts surrounded by brilliant coal from 

 the bark. Prostrate trunks in coal beds are flattened, while stumps 

 in the coal retain the wood, converted into mineral charcoal and 

 surrounded by the brilliant coal. It may be that the uncompressed 

 stems were those thrown upon sandy bars to decay slowly and to 

 have the interior replaced with sand, while the battered fragments 

 may have remained long above reach of ordinary floods, to be 

 swept away only after loss of the interior. But in any event, one 

 must recognize that the small quantity of the wood found in sand- 

 stones is a very important matter; the sandstones and con- 



