Page 7 - Jakubas_et_al._2014_Storm_Petrels

76 D. Jakubas et al.

130 25 30
males

128 females
126 unsexed

124
Mean wing length [mm]
AT
SST
20 25

122 15 20

120 10 15

118

116 -15 -10 -5 0 5 10 15 5 AT Jul–Aug 10
-20 SST Jul–Aug
0 5
Longitude 117 118 119 120 121 122 123 124
Mean wing length (mm)
Fig. 3. Relationship between mean wing length of the
European Storm Petrel and longitude at the species scale. Least Fig. 5. Relationship between wing length in European Storm
squares regression lines are shown for significant relationships Petrels and sea surface temperature (SST) and air temperature
(solid black line — males, solid grey line — females, dotted (AT) in July–August at the subspecies H. p. pelagicus scale. Least
black line — unsexed individuals). Abbreviations and source of squares regression lines are shown for significant relation-
data: see Table 1. ships.

periods was negatively correlated with body mass wing, tail or tarsus length and colony coordinates
(Table 2) indicating that smaller birds occurred in (all p > 0.20). There were also no significant rela-
the areas with stronger winds. At the subspecies tionships between wing, tarsus, tail and bill length
scale, we found significant negative correlations in males and females and the value of SSD
between wing length and sea surface temperature (Spearman rank correlation coefficient, all p > 0.20).
(SST) and air temperature (AT) in all periods indi- In both sexes, there was a significant positive rela-
cating that wing length increased with decreasing tionship between wing length and longitude of
AT and SST values (Table 2, Fig. 5). Body mass was the colony indicating that wing length generally
related significantly to WS in all periods increas- increased from west to east (females: rs = 0.97,
ing with the decrease of the wind speed. At the p = 0.005, n = 5; males: rs = 0.90, p = 0.04, n = 5)
regional North Atlantic scale, we found the same (Fig. 3). There was no such relationship for lati-
pattern for wing length and SST and AT as at the tude (p > 0.62). There was no significant relation-
subspecies scale (Table 2). In the case of body ship between tarsus and tail length in both sexes
mass, we found a tendency towards or significant and colony coordinates (all p > 0.20).
positive relationship with SST and AT in all peri-
ods (Table 2) indicating an increase in body mass Sexual dimorphism of birds captured in the
with increasing temperatures. Faeroes
Four traits differed significantly between the
Geographical variation in sexual size dimor- European Storm Petrel males and females cap-
phism tured in the Faeroes: wing, tail, head-bill length
Sexual size dimorphism (SSD) values varied
among the studied sites (Fig. 6). However, there 2 Wing Tail Head+bill
was no significant relationship between SSD in male-biased SSD
1 female-biased SSD
125 0
FAR WAL1 SHE1 AEG
123 -1
-2
121 -3
-4
119
-5
117 CAN

115
25 30 35 40 45 50 55 60 65 70

Latitude

Fig. 4. Relationship between mean wing length of the
European Storm Petrel and latitude at the subspecies H. p.
pelagicus scale. The least squares regression line is shown for
significant relationship.
Mean wing length [mm]
SSD

POR

Fig. 6. Sexual size dimorphism in different populations of the
European Storm Petrel. Locations are sorted longitudinally;
their abbreviations — see Table 1. POR — birds from the
Atlantic populations captured during migration.