10 MINNESOTA BOTANICAL STUDIES. 



elapsed since the completion of this dam would be at all suffi- 

 cient for the development of such an atoll as that of Anderson. 

 The growth of Picea would perhaps indicate a greater age 

 than this hypothesis would permit. I was at first inclined to 

 think that the fact of lake Whitman being not directly con- 

 nected with the Gull lake chain, but emptying into a rapid in- 

 let stream of Gull lake, was sufficient to disprove the hypothe- 

 sis that the dam on Gull river could have at all effected the sep- 

 aration of the atoll-rings, but Mr. Upham in conversation 

 expresses the opinion that even a disconnected lake in the 

 neighborhood of one, the level of which had been increased by 

 a dam across its outlet, might vary in level through changes in 

 the saturation plane of the surrounding soil. Ballard's atoll 

 could scarcely be connected with the changes in level of Gull 

 lake except upon the assumption of such modifications in the 

 general saturation plane of the district. From the facts at 

 hand neither Mr. Warren Upham nor the writer is inclined to 

 attribute the atoll- formations to the artificial increase in the 

 level of Gull lake and its tributaries. It is however a tenable 

 hypothesis and should it be the correct one furnishes a new 

 and highly interesting example of a method of plant-distribu- 

 tion and the appearance of a unique plaat-physiognomic feature 

 through the modification of natural conditions, as a result of 

 human activity. 



Particular conditions of atoll formation. It is evident that 

 the mere fact of decrease in the level of a pond, followed by 

 increase, does not explain fully the formation of the sphagnum 

 atolls which have been described in this paper. Were these 

 the only factors, apparently such atolls would be much more 

 common phenomena. Instead of being, so far as known, limi- 

 ted to two obscure ponds, they might be looked for in most of 

 the small bodies of fresh water in central Minnesota. The 

 rarity of atoll-producing ponds indicates that there must be a 

 concurrence of several favorable conditions else the atoll will 

 not be formed. A few of these probably necessary conditions 

 will be noted here. 



Height and regularity of pond banks. Both atolls studied 

 were situated in small ponds with high and regular hills sur- 

 rounding them. They were thus sheltered from the winds and 

 this, I believe, is an important condition for the formation of a 

 sphagnum atoll. Had the winds free sweep across the pond 

 it is evident that the atoll in process of formation would, while 



