STEAM Se PRP ESOR SE ie ot poe, Ses fp) SS eee eee 
1913] COOPER—ISLE ROYALE 221 
Gray, Solidago uliginosa Nutt. Shrubs are scattered over the whole 
area, the principal species being Potentilla fruticosa L. and Myrica 
Gale L. The bog forest is of considerable extent and the trees are 
Thuja, Larix,and Picea mariana. Near the forest edge there is con- 
siderable sphagnum in hummocks much overgrown with grasses and 
other plants. The composite character of the vegetation in this 
locality is plain. It is certain that the change from swamp to bog 
conditions has been very gradual, and it is possible also that 
there has always existed here a slight element of bog vegetation. 
THE SECONDARY SUCCESSION 
The burn succession 
I. Causes and extent of fires upon Isle Royale 
During the period of mining activity upon Isle Royale fires 
were of frequent occurrence and many square kilometers were 
swept by them. Since the abandonment of the mines they have 
been much less frequent, so that most of the burned areas found 
today have already gone through a considerable period of forestward 
development. There is evidence that fires occurred long before 
the appearance of white men. A layer of burned wood deeply 
buried was found in the humus upon Smithwick Island, where 
otherwise absolutely no sign of burning was to be seen. Such 
fires must have been started either by Indians or by lightning. 
It is nearly certain that fire has played a part in the vegetational 
history of almost all if not the entire forested area of the island. 
II. Effect of fire upon the climax forest 
The effect of the destruction wrought by a forest fire is essen- 
tially to bring about a return to a more or less xerophytic condition, 
which is followed by a readvance leading again to the climax forest. . 
The secondary development may be along the line of the original 
Primary succession, but factors are usually present which bring 
about pronounced modifications in the process. Obviously the 
burn succession is exceedingly variable, and cannot be described 
in terms that will even approximately fit every case. The variable 
factor that is most important in creating differences in the succes- 
sion is the severity of the fire. In respect to this two cases may be 
roughly distinguished. 
