6 BULLETIN" 187, U. S. DEPARTMENT OF AGRICULTURE. 



farm and affording the largest amount of food per acre. There is no 

 doubt that the average 5 acres immediately surrounding farm build- 

 ings contain more birds' nests than any other 5 acres on a farm. As 

 the farms selected contained orchards larger than the average, and 

 as orchards are especially preferred by many kinds of birds for nest- 

 ing sites, the areas selected evidently have a bird population denser 

 than that of farm lands as a whole. 



The letter of instructions asked specifically that censuses of wood- 

 land be made separately from the rest of the farm land, and many 

 such censuses were received the average of which was 175 pairs to 

 each 100 acres, whUe the farm land contained 119 pairs per 100 acres. 

 It is evident, then, that comparatively small areas of woodland con- 

 tiguous to cleared and cultivated land offer superior attractions as 

 nesting sites and contain a correspondingly larger number of birds' 

 nests than land which is whoUy under cultivation. 



The 58 acres on the average census-covered area of farm land is 

 divided into 17 acres of plowed land, 8 acres of hay land, 4 acres of 

 orchard, 3 acres of woodland, and, after deducting the land imme- 

 diately surrounding buildings, 26 acres, about one-third of which is 

 permanent meadow and two-thirds pasture land. These 58 acres 

 support a bird population of 69 nesting pairs. 



FAKM LANDS NOT COVERED BY CENSUS. 



The 50 acres of each farm not covered by the census consist on 

 the average of 15 acres of plowed ground, 7 acres of hay land, 18 

 acres of woodland, and 10 acres divided between meadow and pasture. 

 If the 18 acres of woodland contain on the average 31 pairs of birds, 

 the other 32 acres would have to support a bird population of only 

 28 pairs to give these 50 acres the same bird population per acre as 

 the part where the buildings are situated. It is not probable that 

 •the 32 acres would have quite 28 pairs, but they may easily have half 

 that number, or a total of 114 pairs on 108 acres, practically one pair 

 to the acre for the land in farms in the Northeastern States. 



NONTARMED LANDS. 



There remains the question of the bird population on the areas not 

 included under the head of farm land. Again reverting to the returns 

 of the census of 1910, in the great farming States of Iowa and Ohio 

 this land represents less than 10 per cent of the whole area, almost a 

 neghgible quantity, while for New England it is just half the total 

 area, and there is nearly as large a proportion in Wisconsin and Min- 

 nesota. 



This nonfarmed land is largely forest, and unfortunately not a 

 single bird census has been contributed from an area typical of forest 

 conditions in the Northeastern States. The only census on a large 



