Beckett and Proudfoot • NORTHERN SAW-WHET OWL MIGRATION 
523 
FIG. I. Northern Saw-whel Owl banding stations in eastern North America. 1999-2008. Bars are 01 latitude in width 
and used to group banding events for timing and age-differentiated analyses. Labels show the number of banding events 
within each bar. Dashed lines define the Great Lakes Basin, Appalachian Mountains, and Atlantic seaboard regions 
compared in directionality und route fidelity analyses. 
for Northern Saw-whet Owls captured twice in the 
same season. We used the slope of the line-of- 
best-fit to approximate average speed. 
Route Fidelity .—We predicted a species with 
high migration route fidelity would demonstrate 
* ow longitudinal deviation with respect to season¬ 
al latitudinal position, i.e.. the longitude at which 
an owl crosses a given latitude during migration 
would be similar among years. The longitudinal 
distance between the two locations would repre¬ 
sent route deviation. Thus, if migrating owls 
maintain high route fidelity, they should not be 
recaptured long distances east or west within a 
single latitude bar from banding location. Longi¬ 
tudinal deviation and recapture records have been 
used for assessing migration route fidelity in other 
species (Rimmer and Darmstadt 1996, Alerstam et 
al. 2006). Thus, we chose to examine migration 
route fidelity by isolating from the data base all 
owls recaptured within 0.5 latitude of their 
banding location at least 1 year after banding. 
We consider a 0.5 latitude window sufficiently 
conservative to accurately represent route devia¬ 
tion while controlling for spurious influence of 
latitudinal position, i.e., we assume longitudinal 
route position may change with latitudinal 
