La Sala et al. • OLROG’S GULL BREEDING BIOLOGY 
245 
TABLE 1. Incubation period (days) (mean ± SD) for Olrog's Gulls in the 2005 breeding season at Isla del Puerto, 
Argentina breeding colony. Data stratified by laying order and clutch size. 
Clinch size 
Laying order One-egg Two-eggs Three-eggs 
A-eggs 28.00 ± 2.83 (26-30, n = 2) 27.26 ±1.10 (25-29, n = 19) 27.64 ± 1.21 (25-29, n = 11) 
B-eggs NA 25.70 ± 1.03 (24-27, n = 20) 26.00 ± 2.00 (24-29, n = 29) 
C-eggs NA NA 25.75 ± 0.96 (25-27, n = 4) 
NA = Nor applicable. 
models (Crawley 2007) using the same approach 
described above, except that NEST ID was the 
only variable evaluated (as random effect) and no 
independent variables were included in the 
models. Repeatability of each egg measurement 
(i.e., the proportion of variance in a character that 
occurs among rather than within individuals; 
Boag and van Noorwijk J987) was estimated 
front the variance component models. 
The size of difference in volume among A-. B-. 
and C-eggs from three-egg clutches in each year 
was calculated using Cohen's d value as effect size 
statistic (Nakagawa 2004). Effect size was ex¬ 
pressed as percentage using the Common Language 
Effect Size Statistic (CL) (McGraw and Wong 
1992) for easier interpretation. For example, a CL 
of 64.4 between B- and C-eggs indicates (he 
probability that a randomly selected B-egg will be 
larger than a randomly selected C-cgg is 64.4%. 
Differences in hatching success among years 
and differences in hatching success among 
different clutch sizes (only in 2005) within 
three-egg clutches were assessed using Chi-square 
tests for trend. Differences in total clutch volume 
in three-egg clutches between 2006 and 2007, and 
in length of incubation between A- and B-eggs 
regardless of clutch size (only in 2005), were 
assessed using unpaired (-tests. Level of signi¬ 
ficance was defined as P < 0.05. Descriptive 
statistics are presented as means ± SD, and 
medians and ranges are reported where appropri¬ 
ate. Residual and normal probability plots were 
used to check for validity of model assumptions. 
RESULTS 
Laying and Incubation. —Egg laying in 2005 
began during the first week of September and 
followed an asynchronous pattern. Mean ± SD 
clutch size was 1.86 ± 0.73 eggs per nest (n = 
101 clutches) and modal clutch size was two eggs 
(range = 1-3). Data for incubation periods varied 
(Table I). The incubation length (all clutch sizes 
considered) for A-eggs in 2005 was 1.67 days 
longer than for B-eggs (27.44 ± 1.22 days vs. 
25.77 ± 1.36 days, respectively; unpaired (-test: 
df = 60, P < 0.001), and incubation for C-eggs 
was 25.75 ± 0.96 days (n = 4). 
Egg Dimensions and Size Variation. —We 
measured 253 eggs from 113 one-, two-, and 
three-egg clutches during the 2006 and 2007 
breeding seasons (Table 3). Large ranges in egg 
size were detected with the largest eggs being 
31.5% (length), 21.3% (breadth), and 66.5% 
(volume) larger than the smallest. Egg length 
TABLE 2. Generalized linear mixed model describing factors associated with i 
2006 and 2007 breeding seasons al Isla del Puerto. Argentina breeding colony. 
egg volume of Olrog's 
; Gulls during the 
Terms 
Coefficients 
SE 
p 
A AIC* 
Intercept 
79.4672 
1.1766 
<0.001 
NA 
Hatching order (Ref.: First) 
Second 
-0.1122 
0.8058 
0.89 
Third 
-3.5039 
0.8066 
<0.001 
NA 
Year (Ref.: 2006) 
2007 
4.0287 
1.6375 
0.015 
NA 
Hatching order (Second)*2007 
-3.2167 
1.1563 
0.006 
Hatching order (Third)*2007 
-4.2275 
1.1604 
<0.001 
8.0 
" AIC value increment if the single term is dropped. 
