CHAP. xvi. POSSIBILITY OF VOLCANIC ACTION. 263 



country. The astonishing speed with which fire runs 

 through the moss is well described by Mr. Davies, 

 quoted in Chapter XIII., and there is no valid reason 

 why a fire should not stretch from Hudson's Bay to the 

 Gulf of St. Lawrence in a few days, as far as the com- 

 bustible nature of the fuel is concerned; but its pro- 

 gress is arrested by the presence of lakes, many and broad, 

 and the swamps by which many of them are terminated. 

 A broad sheet of flame stretching for many miles across 

 is at once divided by a lake, and as these lakes often occur 

 one after another for many miles, the fires are broken and 

 become local in their effects, except in certain cases when 

 the direction of the wind changes in such a manner as to 

 distribute them more widely. A fire in the Labrador 

 Peninsula, where the trees are few and far between, 

 very much resembles a fire in the prairies ; but owing 

 to the extraordinary dryness of the caribou moss it 

 spreads with much greater rapidity. It would be im- 

 possible to escape from an approaching sheet of flame in 

 Labrador by speed. The only plan is to scrape the moss 

 from a few square yards, which is done with the utmost 

 ease, as it adheres to the rock or soil very loosely, and 

 then to he down upon the bare earth. The smoke 

 arising from a fire made from this material is very 

 penetrating, as I experienced when our canoes were in 

 danger on one of the portages. The air is filled with fine 

 dust arising from the ashes ; and on sandy plains, where 

 the lichens and mosses are deep, and other varieties besides 

 the caribou lichens exist in abundance, the charcoal that 

 remains behind covers the soil with a uniform mantle of 

 black. If a volcanic eruption had taken place since the 



