208 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123, No. 2, June 2011 
2 ha in size occur throughout the floodplains of 
these sites. A third Cerulean population is in 
Lawrence County within the 71,600-ha Bankhead 
National Forest <34 20' N. 87'22' W). Bankhead 
National Forest (BNF) is along the southern 
Cumberland Plateau and is characterized by 
dissected sloping ridges and rock bluffs dominat¬ 
ed by loblolly pine (Pinus taeda), upland 
hardwood, and mixed hardwood-pine with addi¬ 
tional canopy species similar to those at the 
Jackson County sites (USDA Forest Service 
2004). Cerulean Warblers are concentrated in 
BNF along the floodplain forests of the 4,200-ha 
Sipsey Wilderness Area. 
Bird Surveys. —We surveyed Cerulean War¬ 
blers from May to June during the breeding 
seasons of 2005 and 2006 by walking the 
floodplains and adjacent slopes of Hurricane 
Creek and Larkin Fork, and along FJannigan and 
Borden creeks in Bankhead National Forest. We 
mapped male Cerulean territories on multiple 
visits using radiotolcmetry, repeated observations 
of col or-banded males, and by distinguishing song 
variability between neighbors (Woodward 1997). 
Ten-minute, fixed-radius point counts {n = 53) 
were performed within a territory and centered 
under individual, singing male Cerulean War¬ 
blers. All counts were conducted once at each 
location prior to 1030 hrs EST following Hamel 
et al. (1996) by a single observer to eliminate 
multiple surveyor bias (Sauer et al. 1994). 
Point counts were also performed once at 47 
additional locations from May to June 2005 and 
2006 in an effort to locate new breeding 
populations (Fig. 1). We concentrated our sam¬ 
pling effort in Skyline WMA and Bankhead NF to 
reduce spatial variability; however, our only 
requirements for these locations were that they 
must occur in deciduous/mixed forest and the 
Cerulean Warbler’s historic breeding range in 
Alabama (Imhof 1976). We used the ArcGIS 
(Version 9.1, ESRI 2005) extension Hawth’s 
Tools (Beyer 2004) to generate 26 random points: 
12 in Skyline WMA, six in Bankhead NF. and 
eight in state parks and nature preserves. An 
additional 19 locations were selected from 122 
pre-existing point count stations in Bankhead NF 
USDA Forest Service 1995). The remaining two 
locations were based on Cerulean Warbler 
observations from Alabama Breeding Bird Atlas 
surveys in 1999 and 2001 (R. L. West, pers 
comm.) in an effort to verify the continued 
existence of breeding Cerulean Warblers. Play¬ 
back of a conspecific song was broadcast for 5 min 
at the conclusion of these counts to ensure that no 
Cerulean Warblers were present. 
Microhabitat Characteristics. —We compared 
microhabitat front used habitat (// = 52) centered 
at point counts conducted under Cerulean Warbler 
males, and in unused habitat (n = 47) defined as 
point count locations where Cerulean Warblers 
were not detected. Habitat measurements for one 
used point count were not collected due to logistic 
constraints. The median distance between used 
locations in the year of sampling was 192.9 m 
(interquartile range = 112.4-383.6) and 2.1 km 
(interquartile range = 1.(1-2.8) between unused 
locations. Vegetation was measured by one 
observer within 0.04-ha (11.3-m radius) circular 
plots following James and Shugart (1970) and 
Noon (1981). Plot measurements included basal 
area, total live stems 2-3-cm diameter at breast 
height (DBH). total snags >8-cm DBH. tree 
height, slope, aspect, understory density, and 
distance to and size of nearest canopy gap. 
Percent canopy cover was estimated from 40 ± 
vettical readings along transects in the cardinal 
directions using an ocular densitometer tube. Each 
reading was assigned one of four height intervals 
(< 5, 5-15. > 15-25. > 25 m, or no cover) to 
estimate canopy structure complexity. Slope was 
measured in degrees using a clinometer and aspect 
was transformed to a value ranging from 0.0 to 2.0 
(Beers et al. 1966). This distinguished less 
productive, southwest facing slopes (value = 
0.0) from more productive, mesic northeast slopes 
(value =■ 2.0) (van Manen et al. 2005). We 
assigned flat plots a neutral value of 1.0. Distance 
to and size of the nearest canopy gap <50 m from 
plot center and >10 nr were measured following 
Runkle (1992). 
Analysis of Point Counts. —All nocturnal, 
colonial, and raptor species including birds with 
restricted vocalizations (e.g., hummingbirds), 
were excluded from the analysis because of the 
difficulty of reliably detecting them during diurnal 
point counts (Bibby et al. 2000). Species richness 
was calculated as die total number of species 
detected during each count, and total number of 
individuals counted at each location was used as 
bird abundance. Species diversity was estimated 
with the Shannon-Weiner index using the Micro¬ 
soft ® Office Excel (Microsoft Inc. 2003) macro 
Biological Tools Version 0.2 (Hanks 1995). We 
constructed a conservation concern value using 
designations developed by the Alabama Nongame 
