304 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 124, No. 2, June 2012 
TABLE 2. Variables examined with multinomial regression models for prey type and prey sizes 
nests in native grass field buffers in north-central Mississippi (USA). May to August 2008-2009. 
brought to Dickcissel 
2008 
2009 
Variable 
F 
df 
P 
F 
df 
p 
Prey type 
Nestling age 
2.73 
4. 600 
0.028 
1.53 
6. 901 
0.166 
Buffer 
4.91 
2. 600 
0.008 
1.15 
2. 901 
0.318 
Observer present 
7.65 
2, 600 
<0.001 
0.04 
2. 901 
0.965 
Nestling number 
2.41 
6. 600 
0.026 
0.14 
6. 901 
0.992 
Male helping 
7.59 
2. 600 
<0.001 
1.39 
2. 901 
0.250 
Prey size 
Nestling age 
3.36 
4. 640 
0.010 
6.26 
6. 1026 
<0.001 
Buffer 
1.92 
2. 640 
0.148 
0.01 
2. 1026 
0.994 
Observer present 
2.35 
2. 640 
0.096 
0.11 
2. 1026 
0.898 
Nestling number 
0.51 
8, 640 
0.847 
0.22 
6. 1026 
0.971 
Male helping 
3.89 
2. 640 
0.021 
7.28 
2, 1026 
<0.001 
et al. 1995) compared to other available arthro¬ 
pods (e.g., Lepidoptera and Araneae), and provide 
proportionally more biomass per prey item, while 
Lepidoptera are high in calcium and arachnids are 
high in phosphorus (Rebel et al. 1995). Hemipleru 
were more abundant than Orthoptera at our study 
site, and they have greater energy content and fat 
than Orthoptera (Robel et al. 1995). Dickcissels 
may have avoided Hemiptera because they are 
small and fast-moving compared to Orthoptera, 
which could increase searching and handling time. 
We excluded prey items <5 mm from analysis 
because observing adults provisioning nestlings 
I I Small prey clems 
I I Medium prey items 
U Large prey items 
S' 
c 
<D 
3 
Sf 
Nestling age (days) 
FIG. 4. Size of prey 
nestlings in north-central 
August 2008-2009. 
items brought to Dickcissel 
Mississippi (USA). May to 
can potentially miss small (<5 mm) items (E. D. 
Doxon, pers. comm.), and we may have biased 
our results against smaller arthropods. However, 
substantial bias is unlikely because we identified 
82% of the prey brought to nests (at least lo 
Order), and small (<5 mm) arthropods comprised 
a small amount of nestling diets. Grassland birds 
are also known to actively avoid small prey items 
(e.g., Kaspari and Joem 1993). 
Foraging distance was not related to cither prey 
size or nestling number as we hypothesized based 
on central place foraging theory' (Orians and 
Pearson 1979. Kacelnik 1984). although Dickcis- 
sels may have made longer foraging trips in re¬ 
sponse to nestling age. Longer travel distances can 
have negative effects on survival of broods (eg. 
Frey-Roos ct al. 1995. Brickie et al. 2000) and 
future reproductive success of parents e.g.. 
Dcerenberg and Overkamp 1999). Prey load and 
quality brought to young may compensate lor 
longer foraging distances (Krebs and Avery I9D 
Kacelnik and Cuthill 1990. Kaspari 1991), bui 
we did not observe this in our study, sugge-i 
ing Dickcissels foraged at distances (i.e.. up to 
■—200 m from nest) that did not impose undue 
energetic costs. 
Foraging trips were shorter when cloud cover or 
wind speeds were higher. This behavior nw) 
retleci an adjustment to increased stress level* and 
greater energy expenditure with increasing wind 
speeds (Wingfield et al. 1983). High winds and 
cloud cover (i.e., >75%) often occurred simulta¬ 
neously and preceded rain at our site. Thus, female* 
may have spent more time brooding and watching 
