502 



SCIENCE 



[N. S. Vol. XLII. No. 1084 



expenditure for fertilizers for tlie year pre- 

 ceding that in whicli tlie census was taken. 

 The ratio between that and the total acreage 

 of improved land is certainly a function of 

 the natural fertility of the soil, although it is 

 of course governed by some other factors as 

 well.^ If we determine the relative fertility of 

 different states or counties by striking an 

 average between the results obtained from a 

 study of the percentage of improved land and 

 the amount of fertilizer used per acre perhaps 

 we shall not be far wrong; and still greater 

 accuracy might be obtained by using addi- 

 tional criteria of an analogous nature. 



The present study is limited to the eastern 

 half of the United States ; because in the West 

 farming is chiefly confined to areas originally 

 treeless, and there is no necessary connection 

 between the fertility of such areas and the 

 composition of the forests in the same states, 

 which may occupy very different soils. Massa- 

 chusetts, Ehode Island and Connecticut are 

 omitted because their large urban population, 

 deriving most of its food from the farms of 

 other states, complicates matters too much; 

 New Jersey for a similar reason and also be- 

 cause its most abundant evergreen, Pinus 

 rigida, is of little value for lumber and there- 

 fore does not figure very largely in the re- 

 turns; and Illinois, Iowa and Minnesota for 

 the same reason as the western states. 



In the table below the 24 remaining states 

 are arranged in the order of their evergreen 

 percentages, as determined in the manner 

 above described, with the highest first. The 

 first column of figures contains these percent- 

 ages, the second the percentage of the total 

 area " improved " in 1880, and the third the 

 amount spent for fertilizer in 1909 for each 

 acre of improved land reported in April, 1910. 

 The last column gives the average rank of the 

 states as determined by columns 2 and 3. 

 E. g., Florida ranks first in proportion of un- 



3 Some persons with whom the writer has dis- 

 cussed the matter during the past year pretend to 

 believe that the amount of fertilizer used depends 

 mainly on the enterprise of the farmers, but they 

 would hardly contend that the farmers of South 

 Carolina and Florida are ten times as enterprising 

 as those of Ohio and Indiana! 



improved land, and second in the use of fer- 

 tilizer per acre, making its average rank IJ. 



Per 

 Cent, of 

 Ever- 

 greens 



Florida 



South Carolina .. 



Georgia 



New Hampshire 



Alabama 



Maine 



Louisiana 



Mississippi 



North Carolina.. 



Delaware 



Virginia 



Vermont 



Wisconsin 



Arkansas 



New York 



Pennsylvania 



Michigan 



Maryland 



West Virginia . . , 



Missouri 



Tennessee 



Kentucky 



Ohio 



Indiana 



91.5 



80.0 



77.4 



72.2 



68.2 



68.0 



65.6 



55.3 



51.5 



41.5 



36.1 



;55.9 



34.6 



30.3 



26.8 



24.2 



19.8 



19.0 



13.5 



6.9 



4.7 



2.3 



0.5 



0.1 



1880 



Ferti- 



lizerper 



Acre 



1909-10 



2.7 

 21.2 



21.8 

 40.0 

 19.5 

 18-3 

 9.4 

 17.6 

 20.8 

 59.4 

 33.0 

 56.3 

 25.9 

 10.7 

 58.1 

 46.8 

 22.6 

 52.0 

 24.7 

 38.1 

 31.9 

 41.8 

 69.3 

 60.6 



?2.00 



2.49 



1.87 



.55 



.79 



1.72 



.38 



.30 



1.39 



1.21 



.70 



.35 



.01 



.07 



.48 



.54 



.07 



1.01 



.10 



.03 



.11 



.09 



.22 



.13 



7 



13 



7 



4 



7J 



14 



lU 



17 



18 



12 



17J 



14J 



16 



13 



15 



19 



15J 



18J 



20 



20 



The departure of the last column from a 

 numerical sequence indicates how far the 

 evergreen percentages and soil fertility, as de- 

 termined from census statistics, from which 

 the personal equation is practically eliminated, 

 fail to correspond. The average difference be- 

 tween these figures and what they should be if 

 the correlation were perfect is only 2|, whereas 

 if there were no correlation at all the probable 

 average difference would be two or three times 

 as great. The correspondence is as close as 

 could reasonably be expected under the cir- 

 cumstances; and moreover, some of the dis- 

 crepancies are easily explained. 



For example, in Maryland, North Carolina 

 and Tennessee a large proportion of the de- 

 ciduous forest is in mountainous regions, too 

 rough for farming but not for lumbering, so 

 that the true evergreen percentages are prob- 

 ably higher than the figures indicate. In 

 Mississippi and Louisiana the deciduous trees 

 are mostly in swamps, which also repel farmers 

 more than lumbermen. In New Hampshire, 

 Vermont, New York and Wisconsin conditions 



