540 THE WILSON JOURNAL OF ORNITHOLOGY • Vol 123, No. 3. September 2011 
FIG. 3. Mean (± SE) number of birds (of species that 
breed only north of the study area) detected per point as a 
function of season along the northern Lake Huron shoreline. 
1993-1994 pooled. 
0.50) whereas the difference between shoreline 
and 3.2 km approached significance (Tukey P = 
0.056). There were significant differences be¬ 
tween birds counted at the immediate shoreline 
and those counted at distances of 0.8 and 1.6 km 
inland (Tukey P < 0.004). Significantly more 
birds were observed al 0.4 km than at 0.8 and 
1.6 km (Tukey P < 0.001) while the difference 
between 0.4 and 3.2 km approached significance 
(Tukey P = 0.060). Long-distance migrant 
abundance was associated with midge abundance 
after controlling for date (r k -= 2 24 n = 4?0 P = 
0.025). 
Abundance of spring short-distance migrants 
did not vary by year (F,. 807 = 0.242. P = 0.623), 
but did vary by distance from shore (F 4807 = 
10.89, P < 0.001), transect (F K9l01 = 5.21, P < 
0.001 >» and date (F, 3 . 807 = 9.42,’ P < 0.001). The 
immediate shoreline had more short distance 
migrants than all other distances sampled (Tukey 
P < 0.002; Fig. 2A). There were no other 
differences in numbers of birds observed relative 
to distance from the shoreline (Tukey P > 0.34). 
Short-distance migrant abundance was associated 
with midge abundance after controlling for date 
(?k = 7.31, n = 420, P = 0.012). 
Species that breed north of the study area 
concentrated near the shoreline in spring (Fig. 3). 
There was an effect of year (F K809 = 5 . 13 , p = 
0.024), distance from the shoreline (F 4SIW = 
n not ; P < 0 001) ’ and da,c <^13.809 = 2.45, P = 
U.U03) on number of northern breeders counted. A 
effect approached significance (F H 8U9 = 
•90, P - 0.056). More northern breeding birds 
were counted in shoreline habitat than at any other 
distance inland (Tukey P < 0.001; Fig. 3i. 
Northern breeding bird abundance was associated 
with midge abundance, after controlling for dale 
(t k = 2.38, n = 420, P = 0.02). 
Fall. — Fewer long-distance migrants (mean 
birds per point) were detected in fall (1.07 ± 
0.07) relative to spring (2.38 ± 0.07) (/ = 108, 
dt = 2,541. P < 0.001). Year did not appear lo 
alfect number of long-distance migrants counted 
l^i.ii47 — 3.24, P = 0.072) although distance 
from the shoreline (F 4JI47 = 3 90. P = 0.004) 
transect (F 8 . n47 = 3.17, P = 0.001), and date 
(7*'i5.ii47 = 6.14, P < 0.001) influenced number 
ol birds observed. More birds were detected at the 
shoreline than at other distances during fall 
(Fig. 2B). However, significant differences were 
only evident when comparing birds detected at the 
immediate shoreline to 0.4 km (Tukey P = 0.030) 
and 0.4 km to 3.2 km (Tukey P = 0.003; Fig. 2B>. 
Fall long-distance migrant abundance was not 
associated with midge abundance after controlling 
for date (/ k = -0.20, n = 420, P = 0.85). 
Abundance of fall short-distance migrants was 
not affected by year (F u042 = 0.72, P = 0.40) 
although it was influenced by distance from 
shoreline (F 4 , 1042 = 3.34, P = 0.010; Fig. 2B), 
transect (^.,^2 = 1.98, P = 0.046), and date 
(T^is.iors — 169, P = 0.047). There were more 
short-distance migrants counted at the immediate 
shoreline than at other distances (Tukey P < 
0.040). except 0.8 km (Tukey P = 0.066) and 
3.2 km (Tukey P = 0.071). There were no other 
differences in number of birds observed relative to 
distance from the shoreline (Tukey P > 0.070). 
Fall short-distance migrant abundance was not 
associated with midge abundance after controlling 
for date (/ k = 0.36, n = 420. P = 0.72). 
There was no effect of year (F,., l47 = 1-02, P 
= 0.31) nor distance from the shoreline (Fj.iM" ' 
2.22, P ~ 0.064) for species that breed north of 
the study area during fall migration. However, 
date (F 1S , JI47 = 1.95, P = 0.016) and transect 
(7^8,1147 = 2.55, P = 0.009) influenced number of 
transients counted during fall migration (Fig. 31. 
Abundance of northern breeders was associated 
with midge abundance after controlling for date 
(f k = 3.27. n = 420, P = 0.0014). 
Vegetation .—Forest composition was similar 
between shoreline and inland plots. The dominant 
tree in the study area, northern white cedar, had 
the same rank order in terms of importance value 
and density at both shoreline and inland sites 
(Table 1). Northern white cedar density was lour 
