6 BULLETIN 621, U. S. DEPARTMENT OF AGRICULTURE. 
THE ROOSTING HABIT. 1 
While crows, even in the nesting season, are more or less clannish, 
their gregarious habit is most highly developed during the colder 
months. Soon after the nesting season one may expect to see evi- 
dences of it, but in the latitude of Washington, D. C, roosts are not 
well established until the end of September. At this time their 
migratory habits have brought together in a comparatively small 
area the bulk of the crow population of North America, so that the 
area lying between the thirty-seventh and forty-second parallels of 
latitude — that is, from Connecticut to Virginia — and extending west- 
ward from the Atlantic coast to beyond the Mississippi Eiver har- 
bors these birds in extremely large numbers. Their roosts are occu- 
pied with considerable fluctuation in population until the advent of 
milder weather in March, when the numbers rapidly decrease. 
A variety of situations, differing widely in the character of vege- 
tation, are acceptable as sites for crow roosts. Pine and other ever- 
greens are most frequently chosen, though records of crows passing 
the night in groves of deciduous trees, as oaks and maples, are com- 
mon. A large roost in Crawford County, Kans., was in a heavy 
stand of catalpa. That crows roost among such low vegetation as 
reeds or tall grass has been noted, while in some cases even in severe 
weather the birds have been known to gather on the ground in open 
fields or on exposed sand bars. 
Many attempts have been made to estimate the number of birds 
which gather at some of these roosts, but the daily fluctuation, caused 
by changes in weather and by birds stopping at some local roost 
when they have been overtaken by darkness, makes the computing 
of their number difficult and, in large measure, unsatisfactory. The 
wide variation of the estimates made by several observers at the 
same roost readily shows the uncertainty of results. Furthermore, 
the impression made upon a person not very familiar with the sight 
of the gathering thousands is quite likely to be an exaggerated one. 
A roost at Arlington, Va., was supposed to contain at the height 
of its occupancy from 150,000 to 200,000 birds. These figures have 
been averaged from the records of a number of observers and may 
be regarded as reliable. The "Arbutus" roost, near Baltimore, 
contained in 1888, according to the account of Mf, C. L. Edwards, 2 
a population of more than 200,000. The St. Louis roosts, about 1886, 
contained from 70,000 to 90,000 crows. One at Peru, Nebr., at the 
same time had 100,000 to 200,000. Other roosts numbering approxi- 
mately 200,000 birds were recorded about the same year in New 
1 More fully treated by the author in Winter Crow Roosts : Yearbook for 1915, U. S. 
Dept. Agr., pp. 83-100 (Sep. 659), 1916. 
2 Edwards, C. L., Amer. Journ. Psychol. I, No. 3, p. 454, May, 1888. 
