130 BOTANICAL GAZETTE [FEBRUARY 
edge, and a fringe of Alnus crispa overhangs the bank (fig. 23). 
Upon the cliff at the left (fig. 22) the forest has also reached the 
lake level, but the trees are necessarily far apart, growing only 
where ledges or large crevices give them foothold. Wave action 
being at a minimum, the limit of possible forest extension is at the 
water’s edge, and the conditions of surface, slope, and exposure 
to wind being favorable, all of the forestable territory has been 
occupied. 
B represents the phase which is perhaps the commonest upon 
Isle Royale, and which shows to best advantage the various stages 
Fic. 23.—Thoroughly sheltered shore of type A: forest to the water's edge; 
southeast side of Blake Point peninsula. 
that make the complete rock shore succession. The limit of possible 
forest extension may be at the water’s edge or at any line above 
it, according to the degree of exposure. Back of this limit there 1s 4 
zone where occupation has been more or less incomplete, because 
of one or more of the retarding influences, unfavorable character of 
the rock, steep slope, and exposure to winds. The manner ol 
invasion is plainly seen upon such shores. As we pass downward 
from the climax forest the trees become continually sparser and 
smaller, and the undergrowth more xerophytic. As we g0 farther 
only scattered trees are seen, closely confined to the crevices, and 
