59 ] ANALYSIS AND INTERPRETATION OF DATA 59 



It might possibly be inferred from a certain degree of 

 correlation between different kinds of geographic environ- 

 ment and different types of literary talent, brought out by 

 Table VII, and also by the facts discussed in the last few 

 pages, that geographic environment was causally related to 

 the distribution of men of letters. A little study will show, 

 however, that so far as correlation exists it was due chiefly 

 to other factors. One might infer, perhaps, that the south- 

 ern states were handicapped by heat, humidity and dis- 

 ease. This is probably true to a considerable extent. It is 

 evident, however, that climatic conditions cannot be con- 

 sidered the predominant influence, because adjacent states 

 possessing practically identical topographical and meteoro- 

 logical conditions varied widely in literary fecundity. Some 

 other influence must be sought to explain why, for example, 

 Alabama ranked far below Georgia, and Rhode Island below 

 both Massachusetts and Connecticut, while the District of 

 Columbia stood far above either Maryland or Virginia. 

 Thus one is led to conclude that though no doubt geographic 

 environment did play a part in the production of literary 

 talent, it was far less important than other factors. 1 



Density, also, might appear to have had an important in- 

 fluence, since, for example, the relatively crowded New 

 England and Middle Atlantic states were relatively more 

 productive of men of letters. A superficial examination of 

 the census reports, however, shows that there was not a 

 uniform relation between the two conditions. If there had 

 been such a uniform relation, Rhode Island, for instance, 

 would have headed the list of states, instead of ranking 1 

 seventh, and Delaware would have been above Maine, New 



1 This conclusion is identical with that of Odin. He conceded that 

 geographic environment has some influence in the production of men 

 of letters, but considered that its influence is too small to be measured. 

 Odin, op. cit., pp. 439 et seq. 



