36 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 1, March 2012 
periods than during migration, when cover may he 
more critical as birds move through unfamiliar 
habitats and are more vulnerable to predators (Petit 
2000 ). 
Arthropod biomass varies temporally and 
spatially, but arthropod numbers typically are 
greater in mature forest than in recently disturbed 
areas (Duguay et al. 2000. Greenberg and Forrest 
2003, Ulyshcn 2005). Blake and Hoppes (1986) 
found certain types of insects to be more abundant 
in canopy gaps during migratory periods, but 
Greenberg and Forrest (2003) reported inverte¬ 
brate biomass to be greater in mature forest than 
m canopy gaps. Duguay el al. (2000) reported 
mean total invertebrate biomass was greater in 
unharvested stands than clearcut stands late in the 
breeding period when most birds had young in the 
nest. The interior of young (1 yr post-harvest) 
gaps may have greater arthropod abundance than 
adjacent mature forest, but the centers of old (7 yrs 
post-harvest) gaps , like those in Qur slud 
contained similar or fewer numbers of some 
arthropod Orders than adjacent mature forest 
(Ulyshen et al. 2004, 2005, 2006). 
Previous studies suggested birds are able to 
lol ow changes in invertebrate abundance (Gruber 
and Graber 1983, Blake and Hoppes 1986. 
Hotas et al. 1986. Gray ,993. Duguay et al. 
(MflO.H " an a Sher °' ’ 00,) . Duguay el al. 
(2000) documented a positive correlation between 
invertebrate biomass and daily nest survival rales 
of breedmg birds and faster growth rates of 
nestlings m forest stands with greater invertebrate 
biomass. Graber and Graber (1983) showed 
Ttbr L m n rati a n COincided with ,hc peak spring 
outbreak ot lep.dopteran larvae, and Holmes et ak 
1986) reported bird abundance often is related to 
outbreaks of lep.dopteran larvae, a primary 
component of the warbler diet. Supplemental^ 
fed temale Black-throated Blue Warblers (scien- 
dic names ol birds are in Table I) in New 
Hampshire produced more second broods and 
_pent less time foraging away from the nest than 
did controls (Nagy and Holmes 2005) 
xZSzzzzrji*# 
(2010) detected ncwi.ir' Kl,Lonc and DeWald 
hemlock wooly adelgid (Adelges tsugae) and 
untreated sites, even though larval Lepidopteia 
were reduced in treated sites. The removal < 
lepidopteran larvae in other studies had link 
ellect on avian site lidelity or reproductive 
success (Nagy and Smith 1997, Marshall et c 
2002). Karr and Brawn (1990) reported bird 
captures in central Panama were not consistent!) 
correlated with arthropod abundance at capture 
locations; they concluded habitat associations u 
birds were not solely food-resource mediated. 
\ itz and Rodewald (2006) documented n 
relationships between the distribution of capture 
ot forest-breeding songbirds in clearcuts are 
microhabitat characteristics, including arthropod 
abundance and vegetation structure. Kilgo(2005 
examined the relationship between Hooded War- 
bler foraging success and arthropod abundance ai 
our study site and concluded arthropods may not 
be a limiting resource for that species during the 
breeding period (i.e., they are capable of finding 
sufficient food resources across the area, even in 
areas ol relatively low' arthropod abundance). 
However, others have suggested birds can sup¬ 
press arthropod prey so that measurable arthropod 
abundance represents what birds leave rather than 
to what they respond (Marquis and Whelan 1994. 
Strong cl al. 2000). Increased bird activity in 
forest surrounding gaps during the breeding 
season in our study should have suppressed 
arthropod populations. Y'et, numbers of all 
arthropods and Coleoptera remained high in forest 
habitats during the breeding period when bird 
numbers were greatest. The lack of consistent 
relationships in our data indicate Kilgo's (20051 
conclusion may be more generally true for other 
bird species in southeastern bottomland hard¬ 
woods throughout the growing season. 
We likely captured more birds in gap than 
forest understory because of differences in habitat 
structure (Remsen and Good 1996). Birds using 
low vegetation within the gaps were more 
available for sampling with a 3-m tall net rhan 
birds in the mature forest. However, count data 
from a concurrent study of the same gaps 
corroborated our mist-net data (Bowen et al. 
-i()()7). These counts sampled both the understory 
and forest canopy and also detected more birds 
using gap habitat than mature-forest habitat 
(Bowen et al. 2007). Our focus was more on 
S ! 1 !! ts in bird use relative to gaps than on 
iffercnces in bird abundance between gap and 
orest. Capture probabilities likely were higher in 
