146 BULLETIN OF THE ESSEX INSTITUTE. 



type in amygdaloid on the siiore of Lake Superior, near 

 Duluth, and I have described and figured equally striking 

 "root " dikes of sandstone and conglomerate in the granite 

 of the Boston Basin'^, which have been exposed by the 

 erosion of the Carboniferous conglomerate of the Nan- 

 tasket area. No student of onr coastal geology can 

 doubt that other fine examples are now forming where the 

 numerous deep and narrow chasms due to the erosion of 

 trap dikes or of the rock between contiguous joint planes, 

 as in the case of the chasm called Purgatory at Newport, 

 R. I., are being filled with sand and gravel, during the 

 gradual subsidence of the land ; and unquestionably dikes 

 of this type were developed on a grand scale when the 

 rugged topography due to pre-glacial and early glacial 

 erosion was wiped out by the drift deposits. 



It appears, then, that although the term sandstone dike 

 is a comparatively recent coinage, four distinct types now 

 claim recognition, and may be classified as follows : — 



1. Dikes formed in antecedent or original fissures 

 (erosion crevices and gorges) by sedimentation. The 

 "root" dikes of Duluth, Boston Basin, and doubtless, 

 many other localities, belong here. 



2. Dikes formed in subsequent or secondary fissures : 

 (a) By earthquake movements forcing quicksand 



upward from a bed below the horizon of the dikes. This 

 type includes the dikes described by Diller^ in California 

 and Hay^ in Nebraska ; and the fact that the ejection 

 of quicksand from fissures is a common accompaniment 

 of earthquakes indicates that many other examples must 

 await discovery. 



(6) By the settling down of sand from overlying 



• Occas. Papers, Boston Soc. Nnt. History, iv, pt. 1, 76-77. 

 ' Bull. Geol. Soc. Aiiicrica, J, 411. 



• Bull. Geol. Soc. America, 3, SO-S."). 



