185 



nounced than the differences. Thus a succession on granite, 

 such as may be studied to advantage on the numerous islands 

 that rise out of the salt marshes in the vicinity of New Haven, 

 is strikingly similar, except in minor details, to that observed 

 on trap.* 



The effect of the chemical nature of the substratum on vege- 

 tation has been the occasion of much debate. A few years since, 

 Fernaldf made extensive investigations of arctic and alpine 

 plants in the northern United States and Canada, and found that 

 their distribution is controlled very largely by the preponderance 

 in the soil of potassium, calcium, and magnesium. The plants 

 which he studied grow mainly on the faces of cliffs, in rock 

 crevices, or on talus slopes, localities where the soil is derived 

 primarily from the rock in place, so that his conclusions possess 

 unusual value. So far as the rate of succession is concerned, the 

 chemical nature of the underlying rock would appear to be most 

 important as it affects the rapidity of erosion. But it cannot be 

 doubted that to a certain extent the specific composition, especi- 

 ally of the earlier rock face and crevice stages, is modified directly 

 by the abundance or scarcity of certain chemical elements in the 

 soil. In Connecticut the problem is complicated by the fact 

 that so much of the soil is of glacial or alluvial origin, and except 

 in a superficial way+ this problem has not been investigated. 



In physical structure unconsolidated rocks like gravel, sand, 

 and clay contrast sharply with consolidated rocks like granite 

 and trap. That this dissimilarity in the nature of the substratum 

 should be reflected in the character of the vegetation is naturally 

 to be expected; yet the differences are less marked than one might 

 be led to anticipate. As a concrete illustration of a succession on 

 uncompacted rock the sand plains succession has been selected. 

 W. E. Britton§ has given an interesting account of the vegetation 

 of the North Haven sand plains, devoting special attention to 

 certain desert-like areas and the structural peculiarities of the 



* Compare Cowles, Bull. Amer. Bur. Geog. ii: 14. 1901. 



fFernald, M. L. The soil preferences of certain alpine and subalpine plants. 

 Rhodorag: 149-193. 1907. 



t See Torreya 13: 109-110. 1913. 



§ Bull. Torrey Bot. Club 30: 571-620. pi. 23-28. 1903. 



