156 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123, No. 1, March 2011 
TABLE 1. Timing of autumnal migration of Sharp-shinned Hawks at Hawk Ridge Bird Observatory, Duluth 
Minnesota, 1974-2009. Julian dates for 1974 indicate when the specified proportion of the total number of migrating hawks 
was attained; Julian dates for other years are rounded approximations of percentile attainment based on mean values (in 
parentheses ± SE) of shift in days relative to Julian dates in 1974. Calendar dates are provided for descriptive reference. 
Percentile 
1974 
1975-1993 
1994-2009 
1975-2009 
25 
50 
75 
99.5 
254; 11 Sep 
258; 15 Sep 
269; 26 Sep 
289; 16 Oct 
253 (-1.37 ± 0.85) 
261 (3.16 ± 0.77) 
272 (2.63 ± 0.75) 
293 (3.47 ± 0.87) 
257 (2.5 ± 0.96) 
266 (7.69 ± 1.0) 
278 (8.87 ± 0.83) 
298 (9.3 ± 0.99) 
254 (0.4 ± 0.71) 
263 (5.23 ± 0.72) 
275 (5.49 ± 0.76) 
295 (6.14 ± 0.82) 
the last 16 years (1994—2009) because a greater 
shift occurred since 1994. We calculated proba¬ 
bility values using SYSTAT (Wilkinson 1992). 
Statistical significance was accepted at a < 0.05. 
RESULTS 
There was a statistically significant increase in 
Julian dates for the 50th percentile of the seasonal 
totals of migrating Sharp-shinned Hawks across 
36 years at HRBO (Fig. 1); greater proportions of 
hawks were migrating later in autumn. The extent 
of the shift was on average ~3 days later in the 
second through fourth percentiles during the first 
19 years since 1974 (1975-1993), but increased to 
a consistent average of ~8 or 9 days later 
compared to 1974 in the second through fourth 
percentiles during 1994-2009 (Table 1). Twenty- 
two (29%) of the 76 total Julian dates registered 
an earlier day of percentile attainment during 
1975-1993 versus 1974 (9 of the 22 earlier dates 
occurred in percentiles >25%). Only 5% ( n = 3) 
of 64 Julian dates were earlier during 1994-2009 
versus 1974 (all earlier dates occurred in the first 
percentile; i.e„ the first month of migration). 
here was a consistent later shift in the last 
1.5 months of migration during the last 16 years 
of the study. The shift in migration was on 
average 4.31 (±0.42) days later for all percentiles 
combined, and consistently ~5 days later on 
average in the last three percentiles across 35 years 
since 1974 (Table 1). 
DISCUSSION 
This study is to our knowledge the first to show a 
shift m timing of the autumnal migration of a raptor 
We demonstrated that Sharp-shinned Hawks since 
m?nrn a !T migrated on average -4 days later at 
HRBO during 1975-2009. We speculate on 
possible factors influencing this phenomenon. 
There is a sequence in movements of age 
cohorts of Sharp-shinned Hawks at HRBO; 
juvenile birds (<1 yr) precede adults (>2 yrs) by 
~2 weeks and, within age groups, females precede 
males by ~1 week (Rosenfield and Evans 1980). 
This sequence in the migration of cohorts, as 
indexed by trapping data obtained during the same 
days and months in each year that counts are 
conducted at HRBO, has not changed during our 
study years (DLE and RNR, unpubl. data). It 
appears that juveniles still migrate principally 
during the first month of migration, and adults 
predominate in the last 1.5 months (DLE and RNR, 
unpubl. data). Precisely which cohort movements 
may have changed temporally in the overall 
migration of Sharp-shinned Hawks at HRBO is 
not known because counters cannot identify age and 
gender of the majority of migrating individuals. 
About 13,300 Sharp-shinned Hawks were 
counted at HRBO annually during 1974-2009. 
Farmer et al. (2008) reported a low, non-signi¬ 
ficant, average percent change per year (~0.7) 
for Sharp-shinned Hawks observed at HRBO 
across most of our study years (1974-2004) based 
on standardized count effort. Sharp-shinned 
Hawks moving through HRBO originate from 
northern Minnesota and a large part of interior 
and, possibly, western Canada (Evans and Rosen¬ 
field 1985, Goodrich and Smith 2008). The long¬ 
term duration of our study ensures that we have 
cross-generational data for Sharp-shinned Hawks 
(Bildstein and Meyer 2000). The long-term shift 
in timing of the migration we documented is not 
likely due to variation in inter-year counts of 
migrating birds (cf. Miller-Rushing et al. 2008), or 
to behavioral plasticity of an age cohort ( cf. 
Miller-Rushing et al. 2008), nor to some local 
geographical effect (Lyon et al. 2008). 
Climate change may have an influence on the 
availability of food for higher trophic species such 
as birds as a result of advanced phenology of 
lower trophic organisms and a prolonged summer 
season (Penuelas and Filella 2001). Several 
