32 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 1. March 2012 
METHODS 
Study Area .—We sampled foliage-gleaning 
birds and foliage-dwelling arthropods within 
forest canopy gaps, gap edges, and mature 
bottomland forest understory during 2001 and 
2002 at the Savannah River Site, a 78.000-ha 
National Environmental Research Park owned by 
the U.S. Department of Energy. The site was a 
mature stand of bottomland hardwoods, 120 ha in 
size, in Barnwell County in the Upper Coastal 
Plain Region of South Carolina. Birds, arthropods, 
and vegetation structure were surveyed in 12 
group-selection gaps harvested in December 1094 
and in the mature forest understory adjacent to 
gaps. Minimum spacing between gap centers was 
100 m. and the mean distance between a gap's 
edge and the edge of its nearest neighbor was 
102.7 m (range = 44-230 m). The gaps were in 
their seventh and eighth growing seasons post- 
harvest during the study. They were of three sizes 
(0.13. 0.26. and 0.50 ha) with four replicates of 
each size. Previous research within this size range 
in these gaps identified a threshold in response by 
breeding (Moorman and Guynn 2001) and fall 
migrant birds (Kilgo et al. 1999). The mature 
lorest canopy was dominated by laurel oak 
(Quercus laurifolia), cherry bark oak (Q. falcuta 
var. pagodaefolia), sweetgum (Uquidamhar styr- 
aciflua ), and loblolly pine (Pinus taeda). The 
midstory was patchily developed, consisting 
primarily of red mulberry {Monts rubra), iron- 
wood < Carp inns caroUniamis). and American 
holly (Ilex opaca). The understorv contained 
patches of dwarf palmetto (Saba!'minor) and 
switchcane (Arundinaria gigantea ). Vegetation in 
the gaps varied from I to 8 m in height and was 
dominated by regenerating trees (primarily sweet- 
gum, loblolly pine, sycamore [Platanus occiden- 
tahs). green ash [Fraxinus pennsy/vanica], oaks, 
and black willow [Salix nigra]) and patches of 
blackberry (Rubus spp.). dwarf palmetto, and 
switchcane. 
Sampling Design .-We surveyed birds and 
arthropods during four avian activity periods in 
, °°' and 2002; . s P ,in g migration (25 Mar through 
15 May), breeding (16 May through 30 Jun), post- 
breeding (1 j u l through 31 Aug), and fall 
migration (I Sep through 18 Oct). These begin- 
n>ng and endmg dates are estimates of biologi- 
Ca ly mea nuigf U | periods, and each overlaps 
SStLd ith ' he Ull ’ Cr Ma " y individuaJs 
initiated breeding on our study area before 16 
May, but transient species that bred to the north 
continued to migrate through South Carolina 
until mid-May. Similarly, some individuals mi¬ 
grated from or through our study area before 1 
September, but most fall migration occurred after 
I September. Wc established a sampling transect 
radiating southward from the center of each gap to 
investigate bird-arthropod relationships within 
each period with three bird and arthropod 
sampling stations along each transect: one in the 
gap center, one at the southern edge of the gap. 
and one 50 m into the forest. 
Vegetation Measurements .—We measured veg¬ 
etation structure during June 2001 and 2002 along 
10-m transects on each side of and parallel to all 
mist-net stations. 1,5 m from each net. We 
measured vertical distribution of vegetation mod¬ 
ified Irom Karr (1971) at I -m intervals along each 
10-m transect (total 20 points). We recorded the 
number of times vegetation touched a 2-m pole or 
the height intervals directly above the pole at 12 
height intervals (0-0.25, 0.26-0.50. 0.51-075. 
0.76-1, 1.1-1.5. 1.6-2, 2.1-3. 3.1-5.5.1-10. II- 
20, 21-25, and 26-30 in). Touches >2 ni high 
were estimated visually. The percent cover for 
each height interval was calculated from the 
percentage of the 20 sampling points with 
vegetation touches in that interval. We calculated 
the mean number of pole touches for height 
intervals <3 m as an index of foliage density for 
understory vegetation. 
Arthropod Collection.—Wc sampled foliage¬ 
dwelling arthropods at each station during each 
avian activity period in 2001 and 2002. We used 
foliage clipping (Cooper and Whitmore 1990) to 
sample foliage-dwelling arthropods on each oi 
live target plant species groups: (I) white oaks 
(white oak 1 Quercus alba), swamp chestnut oak 
[(?• michauxif). overcup oak | Q. lyrata). Durand 
oak ]Q. durandii J), (2) lobed red oaks (cherrybark 
oak). (3) unlobed red oaks (water oak [ Q. m§ro\. 
laurel oak. willow oak [<9. phellos]). (4) sweet- 
gum. and (5) switchcane. This suite of species was 
selected to represent dominant members of the 
understory and overstory, as well as species 
important as avian foraging substrates (Buffington 
et al. 2000. Kilgo 2005). Each sample consisted of 
25 branch tips from each target species group 
(total sample = 125 branch tips) collected in the 
vicinity of each sampling station (i.e.. staying 
within the target habitat type). Each branch-tip 
dipping was 2.54-15.24 cm in length and usually 
came from the end of a branch where most leaves 
