106 Trans. Acad. Sci. of St. Louis. 



failed to discover any trace of implements or of wrought 

 flints on the surface of the barrens. The most obvious 

 and natural explanation would be that the knolls are 

 merely the result of surface erosion, remnants of the soil 

 mantle that once covered the entire area; but when seen 

 in the field this hypothesis can scarcely be entertained 

 as a satisfactory one, in view of the peculiar appearance 

 of the mounds, as mentioned above. Rejecting these ex- 

 planations as unsatisfactory, a third possible one sug- 

 gests itself: May not these low mounds have been con- 

 structed by some animal long since extinct, that once 

 found a congenial habitat in the rocky barrens along the 

 banks of these clear, swift flowing streams? However 

 their origin may be accounted for, their frequency and 

 uniform size render them a singular and striking feature 

 of the region. 



The range of most of the peculiar species of plants is, 

 as has been stated, co-extensive with the chert outcrop; 

 and it is a curious fact that nearly all of those found at 

 the Shoal Creek localities should reappear with the out- 

 cropping of the Grand Falls chert north of Turkey 

 Creek. Although only about six miles distant this area 

 is entirely isolated from the main outcrops, and no trace 

 of most of the barrens plants can be found in the inter- 

 vening or surrounding country. A few of the species, 

 such as Selenia aurea, Cyperus inflexus, Specularia lepto- 

 carpa, Portulaca pilosa, Talinum calycinum and Linaria 

 canadensis, are also found occasionally in sandy soil or 

 in limestone barrens throughout Jasper and Newton 

 Counties. 



The occurrence in the barrens of a number of plants of 

 such restricted range suggests interesting questions in 

 regard to the survival and distribution of species. The 

 soil of the uplands in this part of the State is entirely 

 residual, resulting directly from the disintegration of the 

 underlying rocks. Under such conditions a closer core- 

 lation between geological formations and plant distri- 

 bution could naturally be expected, although of course, as 



