Moorman et al. • SEASONAL BIRD USE OF HABITATS AND ARTHROPODS 
33 
were clustered on the target plant species groups 
(Cooper and Whitmore 1990). We collected 
foliage from ground level to about 2.3 m. We 
placed clippings immediately in plastic bags to 
avoid evasive movements of arthropods, hut 
highly mobile arthropods (a group of less interest 
lor this study) were not as effectively sampled. 
We did not sample above 2.3 m because we 
considered it appropriate to sample arthropods 
only in the same stratum in which we sampled 
birds (i.e., 3-m mist nets). Samples were placed in 
a freezer for 24 hrs to kill all arthropods. We then 
shook the foliage to collect the arthropods, placed 
them in alcohol, and identified them to Order. 
Foliage was oven-dried for 48 hrs at 40 C and 
weighed. 
Mist Netting .—We placed a single mist net 
(12 m long x 3 m tall with 30-mm mesh) at each 
of the three sampling stations at each of the 12 
study gaps, Netting was conducted once each 
week at each station during the spring migration, 
post-breeding, and fall migration periods, rotating 
between stations on a regular weekly schedule. 
Nets were operated once every 2 weeks during the 
breeding period, because birds remain fairly 
stationary during this period. Nets were opened 
at first light and operated for 4-6 hrs, depending 
on daily weather conditions. Netting was not 
conducted when wind exceeded 16 km/hr or 
during steady rainfall. We banded captured birds 
with a U.S. Geological Survey aluminum leg 
band. 
Statistical Analyses .—We assigned birds 
(Table 1) to the foliage-gleaning guild following 
Ehrlich et al. (1988) and Hamel (1992). Birds 
considered winter residents, present only front late 
fall through early spring, were not included in 
analyses. 
We used a linear mixed model (PROC MIXED) 
tSAS Institute 2000) to conduct analysis of 
variance (ANOVA) with covariates and interac¬ 
tions to analyze the effects of net location (gup. 
edge, forest understory), period, and arthropod 
abundance on bird captures. We used mean 
captures of foliage-gleaning birds/100 net hrs as 
the dependent variable. We considered net 
location and period as fixed effects with net 
location as a split plot factor and period as the 
repeated measure. Arthropod abundance was a 
continuous covariate. We included all two-way 
interactions. We used a linear mixed model to 
examine the relationship between bird captures 
and understory (0-3 m) vegetation density with 
TABLE 1. Foliage-gleaning bird species captured in 
mist nets at least once during 2001-2002 in South 
Carolina. USA. 
Species 
Scientific name 
Yellow-billed Cuckoo 
Coccyzus americanus 
Ruby-throated Hummingbird 
Archilochus coluhris 
White-eyed Vireo 
Vireo griseus 
Blue-headed Vireo 
V. solitarius 
Red-eyed Vireo 
V. olivaceus 
Carolina Chickadee 
Poecile carolinensis 
Tufted Titmouse 
Bueolophus bicolor 
Blue-gray Gnatcatcher 
Polioptila caerulea 
Gray Catbird 
Dumetella carolinensis 
Worm-eating Warbler 
Helmitheros vermivomm 
Golden-winged Warbler 
Vermivora chrysoptera 
Blue-winged Warbler 
V. cyanoptera 
Kentucky Warbler 
Geothlypis formosa 
Common Yellowthroat 
G. trie has 
Hooded Warbler 
Setophaga citrina 
American Redstart 
S. ruticilla 
Northern Bum la 
S. americana 
Magnolia Warhlcr 
S. magnolia 
Chestnut-sided Warbler 
S. pensylvanica 
Black-throated Blue Warbler 
S. caerulescens 
Pine Warbler 
S. pinus 
Prairie Warbler 
S. discolor 
Yellow-breasted Chat 
Icteria virens 
Summer Tanagcr 
Piranga rubra 
Northern Cardinal 
Cardinalis cardinalis 
vegetation as the covariate. Vegetation was only 
recorded once each year, so this model did not 
include a repeated measure. Year and gap size 
were not significant (P > 0.05) in any models, 
and these variables were not included in final 
models. Arthropod captures were standardized by 
number of arthropods/100 g of dry foliage. We 
modeled bird abundance with abundance of 
Lepidoptera because previous studies have shown 
Lepidoptera to be a primary avian food source 
(Holmes et al. 1986, McMartin et al. 2002). 
RESULTS 
The greatest understory vegetation density 
occurred in gaps (Fig. 1). Gaps had dense 
understory vegetation with no canopy, whereas 
forest had relatively open understory and midstory 
and well-developed canopy. 
We captured arthropods representing 21 Orders 
during 2001 and 2002. Total arthropod density 
(number of arthropods/100 g of foliage} was lower 
during spring migration than in the other three 
periods and greater in the forest understory than in 
gaps and at gap edges (Table 2). Total arthropod 
