SHORT COMMUNICATIONS 
177 
tality are additive to natural causes, and may pose 
a threat to many declining species. Avian 
mortality during fall migration has been attributed 
to multiple anthropogenic structures, including 
vehicles, communication structures, buildings and 
windows, powerlines, and wind turbines (Erick¬ 
son et al. 2001). Avian mortality during migration 
has also been documented at off-shore oil derricks 
(Hope-Jones 1980) and at navigational lightships 
(Bullis 1954). The common element among most 
of these sources of avian mortality is lighting. The 
bright lights may attract night-migrating birds 
(Gauthreaux and Belser 2006, Gehring et al. 
2009). An additional source of modem avian 
mortality during fall migration is impact and death 
by exhaustion on the open decks of recreational 
cruise ships. 
OBSERVATIONS 
On 28 September 2003, while aboard a 14- 
story, 3,114-passenger cruise liner in the Carib¬ 
bean Sea (~80 km south of Miami, Florida, 
USA), I observed a massive, mixed-species flock 
of migratory songbirds and egrets flying around 
the ship. Flock size could not be estimated due to 
the erratic, non-directional flight behavior of the 
flock, but a visual snapshot would suggest the 
flock was in the magnitude of thousands. During a 
45-min sweep (0015 to 0100 hrs, EDT) of parts of 
the open area of two decks, I found eight dead 
Common Yellowthroats ( Geothlypis trichas). The 
mortality was due to impacts with glass windows 
of upper decks (based on location of two carcasses 
directly below windows) and exhaustion from 
flight within the wind drafts of these open decks 
(4 deaths witnessed). Many birds were trapped 
within partially enclosed portions of these decks. 
The most common species observed within these 
partial enclosures was the Common Yellowthroat. 
Other species observed resting on the ship 
included Louisiana Waterthrush ( Parkesia mota- 
cilla), American Redstart ( Setophaga ruticilla ), 
Tree Swallow ( Tachycineta bicolor ), and Barn 
Swallow ( Hirundo rustica). Cattle Egrets ( Bubul- 
cus ibis ) were observed flying around the ship, but 
they did not land on the vessel. Cattle Egrets were 
likely preying upon the aggregated flock of 
songbirds (P. W. Sykes Jr., pers. comm.), adding 
an additional source of mortality. It appeared 
the birds were attracted to the lights of the ship, 
and then became confused and caught in the 
wind draft associated with the ship’s movement 
( 22 knots/hr). 
The weather condition was overcast with air 
temperature —23° C. About 3 hrs prior to the 
observation period, it rained hard for about 1.5 hrs. 
Migrants may have lowered the altitude of their 
migratory path due to weather conditions of that 
evening. However, songbirds were observed in 
partially enclosed portions of the ship during the 
daytime of the previous 4 days. The weather 
conditions of preceding nights were clear skies 
with similar air temperatures, suggesting the rain 
event was not what brought the songbirds into 
close proximity of the ship. Egrets were also 
observed flying in the wind draft of the ship on 
three previous nights, during clear sky conditions. 
Unfortunately, I did not visit the open decks or 
watch carefully for songbird migrants on any of 
the other nights during the cruise. 
Considering that in 2003 there were ~50 large 
cruise ships that carry 1,200 or more passengers 
operating in the Caribbean Sea (with destinations 
in the Bahamas, East Caribbean, South Caribbean, 
West Caribbean, and East coast of Mexico only) 
during the fall migration (Aug-Oct), this source of 
avian mortality may not be trivial (Table 1). 
These ships are advertised as 207 to 311 m in 
length, and are 10-14 stories in height. They are 
all well lit at night. The general design of these 
vacation cruise ships is to have 3-5 decks with 
open areas, a few of which have partial enclo¬ 
sures. On the ship I observed, the area I searched 
covered ~ 1 /4th of the open areas on the ship, 
excluding balconies (wind draft of ship does not 
swirl into these areas). My 45-min observation 
period represented —1/15th of the dark hours of 
the night. To extrapolate from the mortality I 
directly observed (4 deaths, excluding the 4 birds 
that were found already dead); this single ship on 
this single night may have resulted in four bird 
deaths X 4 X 15 = 240 bird deaths. The cleaning 
staff of the cruise ship acknowledged removal of 
songbirds from the deck during early morning 
hours prior to guest activities, although they did 
not admit how often or how many dead birds 
were removed. Based on advertised sail dates 
and itineraries, the 50 large (>1,200 passengers) 
cruise ships operating in the Caribbean Sea during 
August through October result in an estimated 
2,981 ship-nights (Table 1). Assuming 240 bird 
deaths per ship-night, an estimated 715,440 bird 
deaths may have occurred on large cruise ships in 
the Caribbean Sea during fall migration 2003. 
Given the small sample size and low sampling 
intensity, I am uncertain whether the mortality I 
