Snon i a la el at. • MIGRANT STOPOVER ECOLOGY ON ISLANDS 
219 
Appledore Island, increasing recapture likelihood 
and compensating for the increased chance of 
recapture due to greater netting effort on Apple¬ 
dore Island. 
We analyzed spring and fall separately. We 
pooled data within a season from 1999 and 2000 
for mass gain/loss analysis to better represent 
typical conditions (Dunn 2000); data from Morris 
et al. (1996) also showed similarities in mass gain 
on Appledore Island between years. We used the 
same pooling of data for stopover time and fat 
class analyses so the results could be compared to 
the mass change results. We compared differences 
in fat class between islands using a Chi-square test 
and analyzed only those species with 20 or more 
individuals captured on each island. 
We compared the number, stopover length, and 
mass change of recaptured birds for those 
migrants that stayed for a minimum of one night, 
and were recaptured at least I day after their 
initial capture. Migrants recaptured only on the 
day of initial capture were not included in the 
analyses of recaptured birds. We compared the 
number of individuals recaptured between islands 
using Fisher's exact test (Zar 1999).We calculated 
the minimum length of stopover by subtracting 
the init ial date of capture from the date of the last 
recapture for each individual (Cherry 1982, 
Moore and Kerlinger 1987, Morris el al. 1994). 
and compared differences in mean minimum 
stopover length between islands using a two 
sample /-test. We analyzed only those individual 
species with 20 or more individuals banded on 
each island and five or more recaptured on at least 
one island for both tests. 
We calculated mass change during stopover for 
each recaptured individual by subtracting mass at 
first capture (initial mass) from mass at last 
capture (final mass) (Morris et al. 1994, 1996). 
We used a Chi-square test to compare the number 
of birds gaining and losing mass during stopover 
between islands and analyzed only those species 
with 10 or more recaptured individuals on each 
island. We calculated percent mass change during 
stopover as (final mass-initial mass)/initial mass 
x 100 (Bairlein 1985; Morris el al. 1994, 1996). 
We used a two sample /-test to compare mean 
percent mass change for all recaptures between 
the islands and analyzed only those species with 
five or more recaptured individuals on each 
island. We did not adjust the mass of recaptured 
individuals to a standard time of day following 
Bairlein (1985) and Winker et al. (1992). We also 
analyzed mass gain for the same species on each 
island individually using paired /-tests to compare 
initial mass and final mass. 
We analyzed hourly mass change by analyzing 
mass al first capture of all birds captured using a 
linear regression of mass on time of day (Winker 
et al. 1992, Morris et al. 1996, Dunn 2000). We 
conducted a multiple regression analysis of mass 
on lime of capture with w ing length included as an 
independent variable (Dunn 2000). Time of 
capture was measured in hours since sunrise. 
Wing length outliers were removed if outside the 
range for that species in Pyle (1997). We used 
ANCOVA to compare regressions from each 
island to examine if there w-as a significant 
difference in hourly rate of mass gain or loss 
between islands. Only those species with a 
minimum total of 20 captured (1999 and 2000 
combined) were analyzed. 
RESULTS 
Birds. —We banded 1,572 migrants during 
spring and 2.261 during fall on Star Island and 
3,240 migrants during spring and 3,364 during fall 
on Appledore Island. Scientific naines of species 
and numbers captured are in the Appendix. The 
majority of birds captured during fall W'ere young 
(hatching year; Star Island, 87.2%; Appledore 
Island, 93.0%). 
Fat Class of Birds Captured. —Appledore Is¬ 
land had consistently higher percentages of birds 
with some fat for all species combined; this dif¬ 
ference was significant for both spring and fall 
(spring: x’ = 29.141, df = 1, P < 0.001; fall: f 
— 5.803, df = 1. B = 0.016). Two species had a 
greater percentage of individuals with some fat on 
Star Island, both during fall; 13 species had a 
greater percentage with fat on Appledore Island, 
six during spring and seven during fail (Table 1). 
Recapture Rates. —The average recapture rate 
in spring was 5.7% (91 individuals) on Star Island 
and 3.6% (116 individuals) on Appledore Island. 
Star Island had a higher recapture rate than 
Appledore Island in both spring sampling periods 
but the difference was only significant in 1999 (P 
— 0.001; 2000, P = 0.160, Fisher's exact test). 
The average recapture rate in fall was 9.7% (215 
individuals) on Star Island and 12.3% (411 
individuals) on Appledore Island. Appledore 
Island had a higher recapture rate than Star Island 
in both fall sampling periods but the difference 
was only significant in 2000 (1999. P = 0.325: 
