SHORT COMMUNICATIONS 
387 
that were open on the ground, but with a layer of 
vegetation between 1 and 2 m above the ground. Egg 
success was related to arrival dale, density of the highest 
vegetation layer, density of nests in a given area, and 
percentage of nests on the ground. Brown pelicans that 
arrived earliest appeared to choose more optimal ncst- 
sites and had greater egg and hatching success. Received 
12 August 2010. Accepted 21 December 2010. 
Animals tire distributed non-randomly within a 
given habitat due to the pressures of natural 
selection (Southwood 1977). Non-random distri¬ 
bution is the result of organisms selecting 
particular patches within their habitats and 
selection is presumed to be adaptive: organisms 
choosing optimal habitats will be more successful 
than those that do not (Martin 1998). Selecting an 
optimal nest-site by birds can have implications 
tor nestling fitness and overall survival. 
Brown Pelicans ( Pelecanus occidental^ ) are 
now reaching population levels along the Gulf 
Coast that were achieved prior to the widespread 
use of DDT as a pesticide (Schreiber and 
Risebrough 1972, Wilkinson el al. 1994). This 
species is now a common breeder in the 
southeastern United States, but little has been 
reported on its breeding biology along the Gulf 
Coast (Sachs and Jodice 2009) with no reports for 
Alabama. 
Brown Pelicans, while historically commonly 
observed on the coast of Alabama, had not nested 
in the state (Imhof 1976) until four nests were 
discovered on Gaillard Island in 1983 (Wilkinson 
et al. 1994). As many its 5,000 breeding pairs have 
returned to Gaillard Island each year since 2003 
(Roger Clay, pers. comm.). Our objectives were 
to: (I) investigate nest-site selection, and (2) the 
implications of nest-site selection on hatching and 
egg success of Brown Pelicans. 
METHODS 
Study Site. —This study occurred over two 
breeding seasons (2007-2008) at Gaillard Island. 
Alabama (30 30'N, 88 02'W). Gaillard Island 
is a man-made, dredge spoil island in Mobile Bay 
that is 2.6 km at its greatest width and 3.6 km at 
its greatest length. It is just east of Dog River and 
17.7 km south of downtown Mobile (Robinson 
and Dindo (2009). There is a 6.1-m dirt berm 
perimeter completely around the island that 
protects it from storm surge and provides 
protected nesting hubitats tor numerous bird 
species. The dominant vegetation on the island 
is marsh elder ( Iva frutescens) and cord grass 
(Sportina spp.). Brown Pelicans commonly nest in 
marsh elder on the island and many nest on the 
ground near vegetation. The southern end of 
Gaillard Island, comprising about 20% of the 
island, is used by Brown Pelicans for nesting, as 
much of the island is non-vegetated dredge spoil. 
Brown Pelicans typically arrive on Gaillard Island 
in late March and begin building nests in April. 
New nests can be found through June. Gaillard 
Island is the only known nesting site in Alabama 
for Brown Pelicans. 
Field Methods— Eleven sites (and quadrats) on 
Gaillard Island were selected in February 2007 
and February 2008. prior to arrival of Brown 
Pelicans. Quadrats were chosen based on vegeta¬ 
tion cover and the sites’ positions on the island. 
Vegetation density was ranked as: 0 = no 
vegetation (<2% cover above 0.5 in from the 
ground), 1 = low vegetation density (<30% 
cover), 2 = moderate vegetation density (30-60% 
cover), and 3 = high vegetation density (>60% 
cover). Sites were selected for each category 
inside and outside of the berm. Percent vegetation 
cover was estimated following Sneddon (1993). 
Quadrats (20 nr ) were established at each of the 
sites in locations that represented the diverse 
vegetation cover differences and position on the 
island. 
Nests and young were counted in each marked 
quadrat at least four times each month starting in 
February 2007 and ending in September 2008. 
Nests on the border of each quadrat were counted 
if they were touching the border ol that quadrat 
(Sutherland et al. 2004). 
Statistical Methods. —Egg success is the prob¬ 
ability of an egg surviving from date of initial 
laying to time of fledging (Mayfield 1961) and 
was calculated for each quadrat rather than each 
nest. Hatching success was defined as the number 
of young hatched per egg laid in a given quadrat 
(Schreiber and Risebrough 1972, Mayfield 1975), 
assuming that all eggs laid in a given nest have the 
same chance to hatch. 
Each quadrat was divided into four vegetation 
layers: from the ground to 1 m. 1-2 tn, 2-3.5 m, 
and 3.5 m to the top of the vegetation (Sneddon 
1993). Percent cover at each layer for each 
quadrat was used in the statistical analysis. Seven 
variables were measured to examine their effects 
on egg and hatching success: date of arrival 
(Julian date), vegetation density (DV) of the layer 
<1 m (ground height DV), vegetation density of 
