Smith and Pawn • SONGBIRDS SHIFT AUTUMN MIGRATION 
563 
320 
310 
300 
Dark-eyed Junco 
Slope = 0.07; P= 0.22; r> = 0.05 
Year 
FIG. 2. Relationship between mean Julian capture date and year for six short-distance migrant : 
Kingston Wildlife Research Station. Rhode Island between 1960 and 2007. Trend lines are P 
significant linear trends in mean capture date over time. 
distance migrants (n = 6) significantly delayed 
migration based on 45 years of banding records. 
These species delayed migration chronology by an 
average of 3.0 days per decade; long-distance 
migrants delayed migration by an average of 
3.3 days per decade, while short-distance migrants 
delayed migration by 2.8 days per decade. These 
shifts are similar to those reported by Mills (2005) 
at Long Point Bird Observatory. Canada, where 
four short-distance migrants and one long-distance 
migrant exhibited significant delays averaging 
3.0 days per decade. However, Mills (2005) also 
reported mean advancement of migration by an 
average of 3.7 days per decade in two long-distance 
migrants, while we did not detect evidence of any 
species arriving earlier at the stopover site we 
monitored. We found no evidence of long-term 
shifts in migration chronology during autumn for 
seven of 19 species, which concurs with MacMy- 
nowski and Root (2007). These authors found no 
evidence of timing shifts during autumn migration 
based on birds killed at a building in Chicago from 
1979 to 2002. Van Buskirk et al. (2009) also 
reported no overall changes in timing of autumn 
migration over a period of 46 years in western 
Pennsylvania. 
The' pattern of delayed migration that we 
documented agrees with at least one European 
