STEVENSON— INTERRELATIONS OF THE FOSSIL FUELS. 77 



heath society. An indication of the mode in which peat formation 

 began was observed on the north side of this island, where some 

 spaces of otherwise bare rock were covered with a mat of Sphagnum 

 with other mosses and lichens. As soon as these form their de- 

 posit, other plants obtain foothold and thenceforward accumulation 

 of peat is continuous, if atmospheric conditions remain favorable. 

 The growth of peat in a region where rain, at times, is wanting for 

 weeks and where the soil is a coarse peat, only an inch or two deep 

 and resting on a smooth rock, is due to condensation of atmospheric 

 moisture in fogs. The plant society of this high, dry peat bed is 

 very closely allied to that which characterizes the older and more 

 mature portions of peat beds with a rock substratum — a confirma- 

 tion of the belief that both are xerophytic habitats. 



Chevalier's^'' observations in the Niger area, between 5° and 9° 

 N.L., show that in that region, at 200 to 400 meters above sea-level, 

 a sedge grows luxuriantly on the bare granite and gneiss, where it 

 attaches itself so firmly as to resist the winds and the tropical rains. 

 There, in hundreds of square miles within French West Africa, this 

 sedge-growth has caused an accumulation of peat, 5 to 30 cen- 

 timeters thick. The conditions are wholly unfavorable to increase 

 of peat, as there is a dry season, during which the plants wither and 

 the loss is accentuated by fires ; yet the surface is covered with a 

 fibrous material, described as very humic. 



Ordinarily, however, some organic film is necessary, if the 

 growth is to be rapid. As already stated, those engaged in the peat 

 industry learned long ago that, if peat be removed wholly so as to 

 lay bare the underclay, regeneration of the bog is very slow ; but if 

 a thin cover of peat be left on the floor, it is more rapid. The 

 passage from the floor to clean peat may be gradual or abrupt. If 

 the accumulation be in a pond, the transition may be marked by a 

 faux-mur, showing laminae of sand, clay or marl and peat, or it 

 may consist more or less of the Lebertorf or Sapropel mud. In 

 other types of deposits this faux-mur may consist of alternating 

 peat and silt or sand, evidence of repeated Hooding before the peat- 

 forming plants gained the mastery. At times the passage is abrupt, 



99 A. Chevalier, " Les tourbieres de rocher de I'Af rique tropicale," 

 Comptes Rendns, Vol. 149, 1909, pp. 134-136. 



