3'21 



together as one group, we find an average of 1755 birds to the 

 square mile of orchard, more than three fourths of this number 

 English sparrows; 11S6 per square mile in pasture, nearly one 

 third of them English sparrows; 394 to the square mile of 

 plowed ground, 230 of these being liorned larks; 373 to the 

 square mile of corn, three fourths of these English sparrows; 

 308 to the square mile of meadow-lands, where meadow-larks 

 and cowbirds made each about a third of the number; 231 to 

 the square mile of stul)l)le. about two fifths of them meadow- 

 larks; and 148 to the square mile of young wheat, of which 

 meadow-larks made nearly three fifths. This statement may i)e 

 still further generalized and simplihed by saying that the number 

 of these birds per square mile varies in round numbers from 

 150 in young wheat to eight times that number in pastures, 

 and to neai'ly 12 times the same number in orchards; and that 

 the intervening ratios were 230 pers(juare mile in stubble, 300 

 in meadows, 375 in corn, and 400 on plowed ground. 



The wide differences of tlieir nmnl)ers in these several sit- 

 uations can not be taken to demonstrate corresponding differ- 

 ences in the local or ecological preferences of these birds, 

 although they do indicate something of the eft'ects which birds 

 may be producing on equal areas in these crops. If sparrows 

 resort to orchards largely for resting places and for protection 

 again.st the wind, they would tend to accumulate there in 

 much greater numbers to the unit of area in a country con- 

 taining only scattering small orchards than in one where 

 many large orchards were within their reach; and if horned 

 larks decidedly prefer bare ground to a grassy turf, there will 

 be a larger numiier of them in plowed Kelds to the .square mile 

 when but few Helds have been lately ])lowed than when the 

 larger part of the agricultural area has just been broken up. 



Ratios of Fkecu'ency and Preference. 



Bearing in mind the necessity thus shown for an intelligent 

 analysis and interpretation of certain of the facts, the follow- 

 ing tables of frcipirnci/ n/fios. and (■twfficiciits of p reference may 



