Beckett and Proudfoot • NORTHERN SAW-WHET OWL MIGRATION 
529 
TABLE 2. Adult and juvenile proportions by year (1999-2008), and one-way analysis of mean adult banding latitude 
versus mean juvenile banding latitude of Northern Saw-whet Owls in eastern North America (Wilcoxon Rank-Sum tests). 
Yeir 
Adult n (%) 
Juvenile n (%) 
Mean - SE adull 
banding lat. N 
Mean t SE juvenile 
handing Int. N 
Wilcoxon Z 
p 
1999 
3.378 (32) 
7.298 (68) 
43.33 ± 0.05 
42.56 ± 0.04 
11.83 
<0.0001 
2000 
3,553 (60) 
2,320 (40) 
44.10 ± 0.04 
44.63 ± 0.05 
8.27 
<0.0001 
2001 
2.929 (41) 
4.067 (59) 
43.33 ± 0.05 
42.50 ± 0.05 
12.61 
<0.0001 
2002 
3,220 (53) 
2,894 (47) 
43.76 ± 0.04 
44.24 ± 0.05 
6.21 
<0.0001 
2003 
3.686 (40) 
5.590 (60) 
43.75 ± 0.09 
44.22 ± 0.04 
-6.21 
<0.0001 
2004 
4.743 (58) 
3,449 (42) 
43.36 ± 0.04 
43.73 ± 0.05 
5.56 
<0.0001 
2005 
4,268 (51) 
4,060 (49) 
44.19 ± 0.04 
44.18 ± 0.05 
0.26 
0.79 
2006 
2,645 (37) 
4.540 (63) 
44.69 ± 0.05 
45.47 ± 0.04 
-13.12 
<0.0001 
2007 
4.543 (32) 
9.659 (78) 
43.27 ± 0.04 
43.16 ± 0.03 
2.28 
0.02 
2008 
3.267 (75) 
1,076 (25) 
43.26 ± 0.05 
44.70 ± 0.09 
12.20 
<0.0001 
Totals 
36,231 (45) 
44,985 (55) 
43.67 ± 0.02 
43.69 ± 0.02 
-0.74 
0.46 
limitations in all interpolation modeling to predict 
values in regions with few or no sampling points. 
Banders may begin trapping in response to 
repons through Project Owlnet of banding success 
at more northern stations. Thus, the calendar of 
Northern Saw-whet Owl banding events in this 
study may conceivably be influenced by Project 
Owlnet, and this influence may bias the results. 
However, we observed a Gaussian distribution of 
banding events at each latitude bar (Fig. 2) 
demonstrating that a sufficient sampling of owl 
movement is achieved despite the potential timing 
biases associated with Project Owlnet communi¬ 
cation. We suspect most banders anxiously wait 
for migration to begin each season and open nets 
well before more-than-mcagcr numbers arrive, 
e -g-< 1-2 owls/night. If anything. Project Owlnet 
improves the accuracy of our seasonal timing 
assessment. 
The average speed (10.5 km/day. Fig. 4) was 
slower than the speed of individuals reported in 
other studies (14—32 km/day in Virginia, Brinker 
et al. 1997; 20-30 km/day in Wisconsin, Erdman 
el a l- 1997; 28.8 km/day in Indiana, Brittain el al. 
-009; 37 km/day in Alberta, Priestley et al. 2010). 
However, if we assume that each degree of 
latitude in our study is — 111 km. the progression 
of the peak banding window (Fig. 2) indicates the 
migration ‘front’ moves ~30 km/day. The 
discrepancy between these two rates may be 
because the migration ‘front* is a measure of fluid 
population movement, while the migration speed 
analysis (Fig. 4) represents individual movements 
including stopovers (Whalen and Watts 2002) not 
reflected in the measurement of overall population 
movement. The fastest records (Fig. 4) demon¬ 
strate that Northern Saw-whet Owls are capable of 
sustained movement even if normal migration 
behavior includes frequent stopovers. 
Migration Route Fidelity .—Catry et al. (2004) 
argue that migrant passerines rarely exhibit route 
and stopover-site fidelity because they are soli¬ 
tary, short-lived, and highly terrestrial (and 
therefore have more potential stopover sites than 
other types of birds). Their energetically costly 
flight style also hinders correction for wind drift. 
These qualities are true for Northern Saw-whet 
Owls as well, but our findings suggest this species 
may be generally faithful to migration routes. 
Seventy-two percent of owls recaptured > 1 year 
after banding ±0.5 latitude from banding 
location were recaptured <100 km from their 
banding locations (Table 1), suggesting that 
individuals follow similar migration routes among 
years. 
The Great Lakes may characterize geographic 
barriers that could constrict Northern Saw-whet 
Owl movement and cause a migratory bottleneck. 
°wls recaptured >500 km from handing location (mean = 172.5 ± 6.8"). D - distribution of 381 owls banded in the Great 
Lakes Basin and recaptured >100 km from banding location (mean = 184.9 ± 4.8 ). E = distribution of 46 owls banded in 
to Appalachian Mountains region and recaptured >100 km from banding location (mean - 209.2 ± 13.4 ). 
F = distribution of 80 owls banded along the Atlantic seaboard and recaptured >100 km from banding location (mean - 
'84.0 ± 4.6°). 
