STEVENSON— INTERRELATIONS OF THE FOSSIL FUELS. 115 



It is seemingly probable that the pause at the close of the bench- 

 formation was long enough to permit complete decay of exposed 

 parts of the stems. Usually, however, the decay was complete before 

 the end of the peat- forming period and one generally finds the upper 

 part of the bench continuous over the tops of the stumps. The dura- 

 bility of stumps and of some woods is remarkable even when they 

 are exposed, and it is much greater when they are buried in peat or 

 in loose material containing much moisture. But decay of the wood 

 is much more rapid than that of the bark; a stem may become hollow 

 and the space may be filled with silt or sand holding leaves or remains 

 of small animals, as Lesquereux, De la Beche and Potonie have 

 shown. 



12. The roof or toit of a peat deposit may be as variable as the 

 floor. As in the latter one finds usually a gradual passage from 

 the rock to the peat, giving a faux-mur, so on top one finds similarly 

 a gradual passage from peat to rock, a veritable faux-toit. In this, 

 one recognizes frequently roots, stumps and prostrate stems, remains 

 of the forest which covered the peat. But the forest was not always 

 present ; the deposit was buried before the cycle had been completed, 

 so that one finds, mingled with the silt or sand, leaves only of upland 

 vegetation. The roof may be of freshwater, marine or aeolian origin, 

 it may be sand, clay, marl or limestone ; the calcareous beds accumu- 

 late slowly, the others slowly or rapidly. Erect trees, rooted, are 

 found in the roof material, but unlike those in the peat, they are not 

 all broken off at the same height. Where engulfing material is 

 seolian sand not deposited continuously, the trees may adjust them- 

 selves to the conditions and a long period may elapse before their 

 death, so that the buried forest may remain in normal position and 

 the erect trees may penetrate a notable thickness of rock, as in the 

 Baltic provinces. When the material has been transported by run- 

 ning water, the accumulation may be less rapid, but enough so to 

 kill the trees by rendering the cover too wet. The erect stems may 

 be of any height, from mere stumps to a score or more feet and 

 they may be surrounded by rocks of various kinds, sands or clays, 

 in mass or in alternating layers. 



Under the term " roof '' may be comprehended all rocks between 



