SPARROWS IX THE FIELD. 43 
from a false buckwheat vine. But these were isolated instances, rare 
exceptions to the general rule of ground feeding. 
It should be stated, however, that there was snow on the ground 
during only one of the visits, and then it was barely an inch deep, 
while there was a big bare place along the tributary where spar- 
rows spent much of their time feeding on crab-grass and pigeon- 
grass. Had the snow been a foot in depth all of this food supply and 
much of the ragweed would have been rendered unavailable, in which 
ease it seems probable that the birds would, for a time at least, have 
been obliged to take ragweed akenes from the plants, and perhaps, 
like the snowflake (see p. 53), to resort to the seeds of amaranth and 
lanib's-quarters. which appeared to be less palatable to them. 
The snow afforded very useful records of the actions of the spar- 
rows. A few footprints far out in the middle of the ragweed field and 
many in the weedy garden above showed where the birds had been 
feeding. In many places much chaffy debris gave additional evidence. 
It appeared more frequently under plants of lanib's-quarters than 
under those of the more abundant amaranth, showing the birds' pref- 
erence for the former species: and in most instances bending plants 
just grazing the ground had been resorted to, while sometimes the 
birds had apparently been feeding from the stalks. Similar debris, 
consisting of parts of flowers and broken shells of akenes. were found 
under many ragweed plants. It might be supposed that weed seed 
destruction by mice would be of more value than that by sparrows, 
but the records left in the snow did not justify any such conclusion. 
Not only were mice tracks extremely few compared with those of the 
birds, but no evidence appeared of the eating of weed seeds by the 
few mice that had been in the fields. One track that I carefully 
inspected passed directly through the ragweed in the wheat stubble, 
but nowhere could I find any indications that the mouse had eaten 
ds of any kind. 
Some notes were made of the destruction of ragweed akenes. In 
the snow beneath a ragweed 3± feet high a song sparrow was feeding. 
Subsequent examination showed about 50 imprints of its feet, 3 whole 
akenes, and the broken pieces of the shells of about 2< » more, together 
with some of the debris of floral parts. On another day a hundred 
juncos. white-throats, and song sparrows had been feeding busily for 
half an hour in a patch of ragweed, and a few minutes' careful inspec- 
tion of the ground disclosed only 6 whole akenes among a hundred 
empty half akenes. At another time a dozen fox sparrows were busily 
scratching on the ground under a patch of giant ragweed and picking 
up the large akenes. Here I found a large number of empty entire 
akenes. at the smaller end of each of which was a rent through which 
the meaty seed had probably been squeezed by pressure from the 
birds' beaks. This method of manipulation seems to account for the 
