Beckett and Proudfoot • NORTHERN SAW-WHET OWL MIGRATION 
525 
FIG. 2. Seasonal timing by latitude of Northern Saw-whet Owl migration across eastern North America 1999-2008 
based on 81.584 banding events grouped into 01 c latitude bars. Diamonds show mean ± SD banding day for each latitude 
bar. Reference Julian days: 1 October = 274; I November = 305. 
banded and recaptured <33 nights apart because 
there were insufficient same-season recapture data 
>33 days to accurately represent migration speed 
at longer time intervals. 
Migration Route Fidelity. —Seventy-two per¬ 
cent of 512 Northern Saw-whet Owls whose route 
deviation could be assessed (recaptured >1 year 
after banding within 0.5° latitude of banding 
location) were recaptured within 100 km east- 
west of their banding location. Thirty-four percent, 
were recaptured within 20 km. The individual 
deviating farthest from its previous migration 
route was recaptured 981 km from its banding 
location. Forty-one percent of the data was 
represented by individuals banded and recaptured 
at two proximate stations west of Lake Michigan. 
We removed individuals encountered at these 
stations to isolate their effect on the overall 
analysis. The difference in results was statistically 
significant (x% = 30.34. P < 0.001); however, 
removal of these stations resulted in an increase in 
route fidelity (Table 1). 
There was no significant difference in route 
fidelity measures among owls banded in the Great 
Lakes Basin, Appalachian Mountains, and Atlan¬ 
tic seaboard regions (x : g = 9.08, P = 0.33, 
Table 1). There were no significant pair-wise 
differences between all owls and regional group¬ 
ings (x 2 4 < 8.16. P > 0.09). 
Migration Direction .—Mean ± SD vector 
azimuth of 688 Northern Saw-whet Owls banded 
and recaptured at different locations during the 
same fall migration was 191.5 ± 3.8° and 
statistically similar (F t = 0.16, P = 0.689) to 
owls banded and recaptured >100 km apart 
(Fig. 5A, B). Significantly different {F i = 13.74, 
P < 0.001) south-southeastern movement was 
found in owls banded and recaptured >500 km 
apart (Fig. 5C). Eight percent (8.2%) of owls 
banded and recaptured at different locations were 
recaptured north of where they were banded; 4.4% 
of northbound owls were recaptured >100 km 
from banding location, and none was recaptured 
>500 km distant. 
There was a significant difference in direction¬ 
ality among all owls recaptured >100 km from 
banding site and all three regional groupings (Fi > 
4.26, P < 0.04, Fig. 5B. D-F). There was a 
significant difference in mean directionality among 
owls banded in the Great Lakes Basin versus those 
banded in the Appalachian Mountains (Fi = 11.54, 
p < 0.001 Fig. 5D, E) and the Atlantic seaboard 
(F, = 25.00. P < 0.001, Fig. 5D, F). There was no 
difference in mean directionality among owls 
banded in the Appalachian Mountains region and 
the Atlantic seaboard region. 
Age-differentiated Migration. —Proportions of 
adults versus juveniles differed significantly 
among years (x 2 9 = 5,071.01, P < 0.0001). 
Years 1999. 2001, 2003, 2006, and 2007 were 
significant deviants below the overall mean 
proportion. All other years were significant 
deviants above the overall mean proportion 
(Table 2). 
