OCTOBER, 1902. 
THE MIGRATION OF BIRDS. 
201 
which shows the observations of three evenings combined into a 
single plot, brings out this feature most strikingly. 
In the first observations collected here the general conclusion 
suggested itself that birds discontinue their migrations soon after 
midnight, but the comparison of these observations with those 
from other places points to the conclusion that the topography 
rather than the time of night determines the number of birds 
flying. 
[f the curves drawn from data collected at Detroit, Michigan 
(Plate I.) are compared with the map of MichigaUj it will be seen 
that the maximum of the curve is reached as the telescope points 
more nearly along the course of the Detroit River. The telescope 
used was situated about three-fifths of a mile from the river, on 
the western side. During the early part of the evening the insru- 
ment pointed over the river at an angle ; toward 10 o'clock it grad- 
ually came to point along the course of the river, and then grad- 
ually moved away from the river again. The curve shows a more? 
or less general increase toward 10 o'clock and a gradual decrease 
after that time, until it finally falls even lower than in the early 
part of the evening, when the telescope pointed across the river. 
At Ann Arbor the map of Michigan shows the river so situated 
with reference to the place of observation that we might expect 
the highest point in the curve in the earlv part of the evening. It 
begins high and falls gradually as the telescope moves away from 
this area (Plate I.). 
The topographv surrounding Lake Forest, Illinois, makes the 
conditions influencing the mis^ratory movement much the same 
as those for Detroit. Lake Forest is situated on the southwest 
shore of Lake Michigan. During the early part of the evening 
the telescope cuts the shore line of the lake at an angle; later it 
points more nearlv in a line with the shore, after which it gradu- 
ally moves away from this area again. These are in general the 
same conditions that prevail at Detroit, and the curves for the two 
places are, as we should expect, almost precisely the same. The 
time of observation at Lake Forest extended onlv to 11 o'clock; 
we might predict, however, that were they continued the curve 
would fall considerablv lower (Plate I.). 
The Detroit data show other interesting features th-^t conform 
entirely with the theorv here advanced. Out of 111 birds observed 
74 were traveling northeast, 13 north-northeast, 11 east-north- 
east, 4 north, 5 north-northwest, 3 northwest, and 1 southeast. In 
other words, 66 2-3 per cent of all the birds recorded were flyino: 
northeast and about 21 2-3 nearly northeast, leaving only about 12 
