342 



thrives, how far it can be brought to bear unfavourable circumstances, 

 and by what means the latter may be modified to suit its require- 

 ments, such are the "fruit-bearing" objects of the author's re- 

 searches ; he remarks, however, that in the purely scientific aspect 

 of his subject it is difficult to overrate the interest that attaches to 

 every question touching the sure production of a material that, in its 

 husbandry, manufacture, and consumption, closely concerns so im- 

 mense a proportion of the human race. In the present paper, that 

 branch of the subject which relates to the soil is taken up. 



The author in commencing gives a careful summary of all such 

 experiments on American cotton soils as have been recorded prior to 

 his own labours. These are not numerous. 



The second chapter describes, by the aid of a large geological and 

 climatological map, the precise geographical boundaries of the regions 

 of American cotton-culture, the geological or agronomic, climatal, 

 and meteorological and general superficial conditions of these great 

 surfaces, and the relations of each of these to the growth and culture 

 of the several varieties or staples, such as "Sea Island" or "long 

 staple," and "upland " or " short staple," &c. The latter sort, con- 

 stituting the vast bulk of the crop, and bearing, in 1858, the ratio of 

 eleven hundred and six and a half millions to only about twelve 

 million pounds of the " Sea Island," the author deems worthy of 

 prior investigation. In the first instance, limiting himself to this, he 

 discusses carefully in his third and fourth chapters, the choice of 

 district from which the most typical and important soils for examina- 

 tion should be collected, and decides upon the selection of " fair nor- 

 mal specimens of prairie soil " and its underlying subsoil, and to exa- 

 mine them "as minutely and accurately as possible," believing that 

 thus a clue would be more readily found to the causes of fertility as 

 dependent on soil (for which this region is remarkable), than by a 

 less careful examination of many specimens from this or that locality 

 over so vast a surface. This method, without teaching all that we 

 may want to know, is certain most readily to show us the right direc- 

 tion in which to push further inquiry. Seven specimens of soil and 

 subsoil were finally selected for comparison and analysis, mainly 

 from that class of " prairie land " known as " canebrake land," in 

 Marengo, County Alabama, in lat. 32 35' N., long. 87 36' W., the 

 points being marked upon the map accompanying the paper. 



