

Station 3. Middle of runway 6-24, immediately west of 15-33. 
Observer on north side. Revetments and medium-height trees on south 
side, create favorable updrafts. 
Station 4. West end of runway 6-24. Observer on north side. 
Revetments and medium-height trees on south side, create favorable 
updrafts. Area very similar to area 3. 
Station 5W. On eastern half of runway 6-24. Observer on north 
side. Area level, except for a few small scattered piles of sand and 
one block house covered by low bushes along south side. These create 
minor updrafts, but the area is much less favorable to soaring than all 
areas except 5K. 
Station 5E. East end of rumvay 6-24, Observer on north side. 
Area absolutely level. No updrafts created. 
The average number of birds per hour passing the index points 
is presented in Figure 22. The decided influence of terrain features on 
the abumdance of soaring birds is obvious. Under certain wind conditions, 
the number of tirds over the area with the highest trees and dunes is 165 
times as great as it is over the level area, or expressed in another way 
the birds are increased by over 16,000 percent where soaring conditions 
are favorable. 
In order to establish with certainty that the updrafts result 
ing from the presence of dunes and tall vegetation were responsible for 
increased soaring the figures were broken down and analyzed according 
to wind direction. Winds blowing across the runway toward the wall 
of dunes and vegetation would produce the strongest updrafts; these we 
have termed "more favorable" wind directions. Winds blowing onto the 
runway from over the top of the obstructions would create poorer soar- 
ing conditions; these we have termed "less favorable" winds. Figure 22 
illustrates the marked difference in the abundance of birds resulting 
from different wind directions. 
Statistical tests were used to compare counts made at stations 
1, 2, 3, and 4 under favorable wind directions with those made at these 
same stations under unfavorable wind directions. They showed a highly 
significant difference due to wind direction. In fact, the probability 
that the difference could have resulted from chance alone is consider- 
ably less than one in a thousand. 
In order to evaluate the effect of terrain modification, which 
removed favorable soaring conditions, two counting sectors of equal size 
were chosen near runway 15-33. During many observations prior to mid- 
April it was noted that numerous albatrosses soared along the embank- 
ment of sand overgrown with Scaevola and ironwood trees that bordered 
15-33. A number of these birds turned and banked out over the runway. 
In late April and early May about one-third of this area was flattened 
for use in water catchment. Counting station 6 was chosen opposite 
43 
