Location of Strikes in Relation to Terrain and Wind Direction 
The locations of 247 albatross strikes which occurred on 
runway 6-24 were recorded by personnel in the Air Control Tower. Of 
this total, 220 (89 percent, or 44 per 1,000 feet of runway) occurred 
along the western 5,000 feet of the runway, adjacent to high trees, 
dunes, and revetments, which created updrafts, while only 27 (11 
percent, or 9 per 1,000 feet of runway) occurred along the eastern 
3,000 feet of the runway, where the shoulders have been leveled. 
Expressed in another way, albatross strikes are five times as likely 
to occur along portions of the runway where the adjacent terrain 
causes favorable soaring conditions, than they are along portions 
with wide level shoulders. 
Figure 1 illustrates the relative proportion of strikes 
occurring at each 1,000-foot section of this runway. It can be seen 
that the highest number of strikes occur at the eastern end of the 
line of dunes along the south edge of this runway. The updrafts 
caused by this line of dunes result in its being a major flyway, and 
its eastern end is an "intersection" from which albatross flight lines 
fan out over the entire eastern half of the island. During periods of 
high wind in November, a veritable cloud of hundreds of albatrosses 
could be seen wheeling about this point. 
Data on these same 247 albatross strikes indicated that 158, 
or 64 percent,occurred under wind directions which were more favorable 
for soaring (see Studies of Control Methods -- Terrain Modification), 
while 89, or 36 percent, occurred under wind directions which were 
less favorable for soaring. Albatross strikes are thus twice as likely 
to occur when wind directions are such as to create the strongest up- 
drafts. 
D. Studies of Control Methods 
Large-scale Killing Program 
In order to test the effectiveness of albatross elimination 
as a method of reducing the aircraft hazard, and also in hopes that it 
would provide some immediate alleviation of this hazard, the Navy re- 
quested that an experimental killing program be undertaken in the 
breeding season of 1957-58. Although the previous season's relatively 
small-scale experimental killing program (Kenyon, et al. 1958) 
indicated that no immediate practical results were likely to accrue 
therefrom, it was thought that a more extensive experiment was neces- 
sary. ' 
Methods.--The killing program was limited to the area within 
750 feet of the centerline of the western half of runway 6-24 (west 
of its junction with runway 15-33), with the exception that the entire 
24 
