Wood and Perkins • CERULEAN WARBLER ACTIVITIES 
499 
TABLE 1. Number of observations (n) of male Cerulean Warblers (CERW: 20 on Hart Ridge, 9 on Snake Ridge) by 
activity and vegetative strata in Wetzel County, West Virginia, May-June 2005. 
Harl RiUgc Snake Ridge 
n 
% 
n 
% 
Activity 
Singing 
381 
64.6 
209 
77.7 
Foraging 
94 
15.9 
17 
6.3 
Perching 
36 
6.1 
15 
5.6 
Flying 
39 
6.6 
13 
4.8 
Hopping 
4 
0.7 
2 
0.7 
Preening 
21 
3.6 
11 
4.1 
Bathing 
1 
0.2 
0 
0.0 
Interaction with CERW female 
10 
1.7 
0 
0.0 
Dancing 
2 
0.3 
1 
0.4 
Aggressive interaction with CERW male 
1 
0.2 
1 
0.4 
Interspecific interaction 
1 
0.2 
0 
0.0 
Vegetative strata 
0-20% ( understory) 
10 
1.6 
1 
0.4 
21-40% (midstory) 
76 
12.2 
27 
10.3 
41-60% (lower-canopy) 
214 
34.4 
118 
44.9 
61-80% (mid-canopy) 
209 
33.6 
76 
28.9 
81-100% (upper-canopy) 
113 
18.2 
41 
15.6 
using GPS Pathfinder software and were accurate 
to 0.5-4.5 m with the majority of points accurate 
to 3 in and points on the ridgetop accurate to 0.5- 
1.5 m. We used coordinates of the male locations 
and fixed kernel methods (Barg el al. 2005) to 
calculate 95% kernel home range estimates 
(territories) and 50% kernel estimates (core areas) 
with the animal movement extension (Hooge and 
Eichenlaub 2000) in ArcView 3.2 using the least 
squares cross-validation method. Canopy gaps 
present on each study site were digitized based 
on Held sketches. GPS points collected on gap 
boundaries in the field, and 2003 aerial photo¬ 
graphs (Perkins 2006). Canopy gap, territory, and 
core area shapedles were overlaid onto 2003 
aerial photographs in UTM NAD83 State Plane 
West Virginia north using ArcGIS 9.0 (Perkins 
2006). We classified the location of each activity 
as inside or outside of the core area for each 
territory. We also classified the location of each 
activity in relation to canopy gaps. Locations that 
were within I m of a canopy gap were considered 
to be associated with a gap. 
Statistical Analyses.—We summarized loca¬ 
tions into activities, vegetative strata, and tree 
species on the two ridges and used Pearson Chi- 
square tests of independence in SAS Version 8 
ISAS Institute 1999) for statistical analyses. 
Expected values in each cell are the product of 
row observations and column observations, divid¬ 
ed by total observations (Winkler and Hays 1975). 
We used Fisher's exact test when sample size was 
loo low for Chi-square tests. We considered 
differences significant at l‘ ~ 0.05. 
We combined data from the two ridges to 
examine if male activity (singing, foraging, per¬ 
ching, preening) differed among vegetative strata, 
within versus outside of territory core areas, and 
associated versus not associated with canopy 
gaps. We also examined whether vegetative strata 
used for singing or foraging differed within versus 
outside of core areas and in association with 
versus not associated with gaps. 
We examined use of tree species separately for 
each site because the two ridges sampled had 
different tree species composition. We compared 
the distribution of singing and foraging activity 
among the nine most frequently used tree species 
on Hart Ridge to examine if different tree species 
were used for different activities. Availability of 
tree species was assumed to be equal for each 
activity. Activity was not compared among tree 
species on Snake Ridge because individual cell 
counts within the Chi-square were too low for 
a valid test. We compared singing use of tree 
species between core and non-core areas of 
territories for both sites combined. We made no 
conclusions about preferential use of tree species. 
