Vegetation of Northern Cape Breton. 429 



A somewhat puzzling modification in floating mat formation 

 is exhibited in particular by many of the small ponds in the 

 barrens. Commonly the encroaching banks advance at a more 

 or less uniform rate into the pond from one or several sides, but 

 frequently the rate of advance varies locally, and to such an 

 extent that the marginal bog comes to project out into the pond 

 in triangular or tongue-shaped masses (Fig. 64). Through the 

 continued spread or coalescence of such masses a relatively large 

 pond may become subdivided into several smaller ones : in one 

 case noted a group of nine small ponds had thus originated. 

 This singular behavior is not correlated with any differences in 

 the depth of the water, and for a long time the author was at 

 a loss for an explanation. The solution, however, appears to be 

 somewhat as follows. Attention has already been called to the 

 fact that in the small, undrained ponds of the barrens, the sur- 

 face during summer is commonly occupied by a floating mass 

 of aquatic sphagnums. In winter this mass sinks to the bottom. 

 Here the loose tangle may become intergrown with various fila- 

 mentous algae to such an extent as to render it impervious to 

 gases, and the following spring it may be floated toward the 

 surface as a result of gas accumulation underneath or within 

 the mass. Cases of this sort have been frequently noted. As 

 a rule, only portions of the mass actually reach the surface, and 

 on the substratum thus presented sedges or shrubs may gain a 

 foothold, thus inaugurating what essentially is a floating mat. 

 Very often, as might be expected if this explanation is correct, 

 peninsulas and islands of bog are encountered in these ponds. 

 In this connection, see Powers' paper on "Floating Islands" 



Ci4). 



Encroachment from without in combination with filling from 

 within. — It commonly happens that the conversion of a pon.d 

 into a swamp is accomplished through a combination of the two 

 methods of filling just described. The filling in up to the sur- 

 face level may be due largely to the activity of aquatic vegetation, 

 and it is on the substratum thus formed that the mat advances. 

 In this connection the conditions observed around a small, name- 

 less lake, near the upper limits of the forested region, and studied 

 with some care will serve as an illustration. The lake covers an 

 area of perhaps two acres, has roughly the shape of a rounded 

 equilateral triangle, and is drained by a small, sluggish stream. 



