330 ANNUAL OF ^SCIENTIFIC DISCOVERY. 



roots. Such a relation between sigillariae and stigmarise had, it is true, 

 been already established by Mr. Binney of Manchester, and had been 

 suspected some years before on botanical grounds by M. Adolphe 

 Brongniart ; but as the fact was still doubted by some geologists both 

 in Europe and America, it was thought desirable to dig out of the 

 cliffs, and expose to view, several large trunks with their roots attached. 

 These were observed to bifurcate several times, and to send out root- 

 lets in all directions into the clays or ancient soils in which they had 

 grown. Such soils or underclays with stigmarise afford more conclu- 

 sive evidence of ancient terrestrial surfaces than even erect trees, as 

 the latter might be conceived to have been drifted and fixed like snags 

 in a river's bed. In the strata 14,000 feet thick, above mentioned, root- 

 bearing soils were observed at sixty-eight different levels ; and, like 

 the seams of coal which usually cover them, they are at present the 

 most destructible masses in the whole cliff, the sandstones and lamin- 

 ated shales being harder and more capable of resisting the action of 

 the waves and the weather. Originally the reverse was doubtless true ; 

 for in the existing delta of the Mississippi, the clays in which innu- 

 merable roots of swamp trees, such as the deciduous cypress, ramify in 

 all directions, are seen to withstand far more effectually the excavat- 

 ing power of the river or of the sea at the base of the delta, than do 

 beds of loose sand or layers of mud not supporting trees. This fact 

 may explain why seams of coal have so often escaped denudation, and 

 have remained continuous over wide areas, since the roots, now turned 

 to coal, which once traversed them would enable them to resist a cur- 

 rent of water, whilst other members of the coal formation, when in 

 their original and unconsolidated state consisting of sand and mud, 

 would be readily removed. The upright trees usually inclose in their 

 interior pillars of sandstone or shale, or both these substances alternat- 

 ing, and these do not correspond in the thickness of their layers, or in 

 their organic remains, with the external strata, or those enveloping 

 the trunks. It is clear, therefore, that the trees were reduced while 

 yet standing to hollow cylinders of mere bark, (now changed to coal) 

 into which the leaves of ferns and other plants, with fragments of stems 

 and roots, were drifted, together with mud and sand, during river in- 

 undations. The stony contents of one of these trees, nine feet high 

 and twenty-two inches in diameter, on being examined by Messrs. Daw- 

 son and Lyell, yielded, besides numerous fossil plants, some bones and 

 teeth which they believed were referable to a reptile ; but not being 

 competent to decide that osteological question, they submitted the spec- 

 imens to Dr. Wyman, of Harvard University. That eminent anato- 

 mist declared them to be allied in structure to certain perennibranclil- 

 ate batrachians of the genera J\lenobranclius and ]\Ienopoma, species of 

 which now inhabit the lakes and rivers of North America. This deter- 

 mination was afterwards confirmed by Professor Owen, of London, who 

 pointed out the resemblance of some of the associated flat and sculptured 

 bones, with the cranial plates, seen in the skull of the Archegosaurus and 

 Labyrinthodon. In the same dark-colored rock, Dr. Wyman detected 

 a series of nine vertebras, which from their form and transverse pro- 



