Combined use of stable isotope and fatty acid analyses reveal almost total
segregation by salinity habitats in a coastal population of European eel.
1,2
2,3
1
1
Chris Harrod Jonathon Grey , T. Kieran McCarthy , and Michelle Morrissey .
1) National University of Ireland, Galway, Ireland 2) Max Planck Institute of Limnology, Plön, Germany & 3) Queen Mary, University of London, UK.
harrod@mpil-ploen.mpg.de
Elver
Leptocephalus
MARINE
COMPONENT
FRESHWATER
COMPONENT
MIXOHALINE
COMPONENT
Egg
12
FW
BW
MW
SW
11
-20
-18
-16
Mean (± 95% CI) lipid-treated δ13C (‰)
6
3
2
B
Wilk’s lambda = 0.075, P < 0.0001
Jackknifed classification success = 70 %
1
0
-1
Yellow eel
Silver eel
GROWTH HABITAT
Fig. 1: Schematic representation of variation in the level of catadromy typically expressed in
the life cycle of anguillid eels. Traditionally, anguillid eels were considered to undergo obligate
catadromous migrations. Recent studies have repeatedly demonstrated that yellow eel
populations2-3 may include components that recruit to coastal habitats, and that remain in
marine water throughout their life cycle, components that move into freshwater as juveniles
and remain until they mature and return to the spawning grounds, and a further component
where individuals follow mixed strategies and either inhabit brackish waters or move
between freshwater and marine habitats.
Using stable isotope analysis (SIA) of European eels
inhabiting a coastal catchment in Western Ireland (Fig. 2),
we recently demonstrated1 strong evidence for
population sub-structuring along a short, and abrupt
salinity gradient (Fig. 2: sites FW, BW & MW, salinity range =
0-25). We subsequently expanded our focus to include
fully marine habitats (Fig. 2: SW, salinity = 34). Carbon and
nitrogen stable isotope data (δ13C & δ15N) recorded from
eels collected from all four salinity habitats indicated
significant variation according to salinity habitat (Fig. 3A).
CO
N
N
MA
NE
RA
FW (< 0.1)
Tidal limit
-3
-10
-5
0
5
DF Axis I (96 %)
5 km
MW (25)
-2
FW
BW
MW
SW
-4
-6
-6
2
-2
DF Axis I (48%)
6
Fig. 3: A) Isotopic biplot showing variation in mean δ13C and δ15N values from eels captured in
different salinity habitats from the Screebe catchment. B) Scatterplot showing variation in
discriminant function scores based on δ13C, δ15N & C:N data. DFA revealed significant
differences between eels collected from each of the four salinity zones, and on average 70%
of the eels could be assigned to salinity zone using δ13C, δ15N & C:N data.
Fig. 4: Scatterplot showing variation in discriminant function scores based on individual fatty
acid data. MANOVA revealed significant differences between eels collected from each of the
four salinity zones, and on average 77% of the eels could be assigned to salinity zone using
fatty acid data alone.
2. METHODS:
» Yellow-phase eels were collected by fyke net from the
lower Screebe catchment (Fig. 2).
» FAA + SIA:
When SIA and FAA data were combined in a single DFA,
the results were striking. Classification success increased
to 93 %, with much reduced overlap, and almost total
separation of eels according to salinity habitat (Fig. 5:
Wilk’s lambda = 0.004, P < 0.00001; classification success
= 93%).
» Eels were captured from four salinity habitats:
— freshwater (FW, salinity <1, n = 30)
— brackish (BW, salinity ~12, n = 16)
— marine-dominated (MW, salinity ~25, n = 31)
— full seawater (SW, salinity ~34, n = 29).
10
» Dorsal muscle tissues were removed for SIA & FAA
analyses.
» SIA: δ13C (lipid-treated) & δ15N measured via continuous
flow isotope ratio mass spectrometry.
» Statistics: Data were analysed through DFA in SYSTAT
11.01. We compared the relative utility of SIA, FAA and
combined SIA/FAA data to correctly assign eels to salinity
habitat using jackknifed classification success.
SW (34)
0
10
» FAA: A total of 32 fatty acids were extracted, esterified
and analysed via gas chromatography. Individual fatty acid
compositions were quantified using standards of
odd-chained fatty acid methyl esters (FAME). FAME
compositions in eel samples were recorded as % of total
fatty acids by mass, and were arcsine-transformed prior to
analysis.
BW (12)
2
FW
BW
MW
SW
-2
SPAWNING GROUND
Wilk’s lambda = 0.021, P < 0.0001
Jackknifed classification success = 77 %
4
DF Axis II (33%)
Glass eel
13
DF Axis II (15%)
FRESHWATER
3. RESULTS:
» FAA:
DFA demonstrated clear differences in the fatty acid
composition (Table 1) of eels from each of the four
salinity habitats (Fig. 4: Wilk’s lambda = 0.021, P < 0.0001).
Discrimination of eels from FW and BW habitats was
improved relative to SIA data (see overlap in Fig. 3B), but
overall classification success was similar (75%), due to
some overlap in the fatty acid composition of eels from
BW and MW habitats.
A
-22
DF Axis II (4 %)
MIXOHALINE
MARINE
Mean (± 95% CI) δ15N (‰)
1: INTRODUCTION.
Scientific understanding of the ecology of many anguillid
eels, including that of the European eel (Anguilla anguilla)
has recently undergone considerable re-evaluation
(Fig.1). Previously, it was considered that anguillid eels
underwent obligate catadronous migration into
freshwater. However, recent studies have demonstrated
that many populations of yellow-phase eels include
sub-populations that display distinct migratory
behaviours (see Fig.1). Many current management and
conservation measures only relate to eels following
‘typical’ migratory behaviour and there is a pressing need
for means to assess how yellow/phase eel populations
structure themselves with regard to salinity.
Wilk’s lambda = 0.004, P < 0.00001
Jackknifed classification success = 93 %
5
0
FW
BW
MW
SW
-5
-10
-10
-5
5
0
DF Axis I (77%)
10
to Atlantic Ocean
1km
Table 1: Fatty acid composition of European eels collected from the lower Screebe system
NB: asterisks indicate significant differences (bootstrapped ANOVA)* P < 0.05, ** P <0.01, ***P< 0.001
Fig. 5: Combination of SIA and FAA data into a single DFA shows almost total classification
success, with 93% of eels correctly assigned to salinity habitat.
Mean (±95% CI) FAME compositions in eel samples as % of total fatty acids by mass
Fig. 2: The lower sections of the Screebe system. This Atlantic coastal catchment consists of
a series of contiguous basins that due to variation in tidal influence have extremely different
salinity regimes (e.g. salinity varies between 0-34 over a linear distance < 4 km). There is very
little evidence of cultural eutrophication, and eel stocks throughout the system are free
from exploitation. See Harrod et al1 for a description of the catchment. Figures in
parentheses relate to median benthic salinity values.
Discriminant function analysis (DFA) demonstrated that
isotope ratios differed in eels collected along the salinity
gradient (Fig. 3B: Wilk’s lambda = 0.075, P < 0.0001), but
that the ability to classify eels according to salinity habitat
declined relative to our earlier study (e.g. classification
success was reduced from 85 % to 70 %). In an attempt to
improve classification success, here we apply a method
used to distinguish between freshwater and marine
foraging behaviour in seals4, and combined SIA with
complementary fatty acid analysis (FAA). The utility of this
technique is that consumer fatty acid compositions, like
stable isotope ratios, reflect that of their diet, and that
different food webs (e.g. marine and freshwater) are
characterised by contributions of certain fatty acids5.
FATTY ACID
FW
BW
MW
SW
C12:0
C14:0
C14:1 (n-9)*
C16:0
C16:1 (n-9)***
C16:2***
C16:3***
C18:0***
C18:1 (n-9)
C18:1 (n-7)*
C18:2 (n-6)***
C18:3 (n-6)***
C18:3 (n-3)
C18:4 (n-3)
C20:0**
C20:1 (n-9)
C20:1 (n-7)***
C20:2 (n-6)
C20:3 (n-6)***
C20:4 (n-6)***
C20:3 (n-3)***
C20:4 (n-3)**
C20:5 (n-3)***
C22:0
C22:1 (n-11)*
C22:1 (n-9)
C22:2 (n-6)
C22:4 (n-6)*
C22:5 (n-3)***
C22:6 (n-3)
C24: 1 (n-9)**
0.4 (± 0.19)
2.7 (± 0.5)
0 (± 0.03)
33.4 (± 1.87)
4 (± 1.02)
1.4 (± 0.38)
0.1 (± 0.08)
11.6 (± 1.43)
16.1 (± 2.2)
5 (± 0.54)
2.5 (± 0.48)
0.2 (± 0.1)
0.2 (± 0.08)
0 (± 0.02)
1.1 (± 0.39)
0.7 (± 0.3)
0.2 (± 0.1)
0.6 (± 0.12)
0.5 (± 0.06)
7.9 (± 1.63)
0.4 (± 0.10)
0.8 (± 0.37)
2.2 (± 0.62)
0.2 (± 0.06)
0.1 (± 0.12)
0 (± 0.02)
0 (± 0.02)
0.6 (± 0.16)
2.9 (± 0.37)
3.7 (± 0.76)
0.6 (± 0.18)
0.2 (± 0.14)
4.7 (± 2.78)
0 (± 0.01)
34.8 (± 6.91)
3.7 (± 0.9)
0.5 (± 0.5)
0.1 (± 0.08)
9.5 (± 2.06)
17.9 (± 3.11)
4.1 (± 0.57)
3.4 (± 0.98)
0.1 (± 0.06)
0.1 (± 0.05)
0.1 (± 0.06)
1 (± 0.54)
2.1 (± 1.37)
0.2 (± 0.11)
0.8 (± 0.33)
0.3 (± 0.12)
4.6 (± 1.62)
0.4 (± 0.58)
1.1 (± 1.41)
1.7 (± 0.54)
0.1 (± 0.07)
0.6 (± 0.85)
0.1 (± 0.07)
0 (± 0.01)
0.3 (± 0.16)
1.5 (± 0.51)
5.5 (± 2.3)
0.4 (± 0.22)
0.4 (± 0.09)
3.3 (± 0.48)
0.1 (± 0.03)
38.7 (± 3.26)
6.9 (± 1.05)
1 (± 0.27)
0.2 (± 0.06)
8.1 (± 1.04)
16.4 (± 2.25)
4.6 (± 0.52)
3.7 (± 2.2)
0.3 (± 0.04)
0.1 (± 0.05)
0.1 (± 0.03)
0.4 (± 0.08)
0.5 (± 0.08)
0.3 (± 0.09)
0.4 (± 0.18)
0.4 (± 0.05)
3.9 (± 0.58)
0.1 (± 0.04)
0.8 (± 0.12)
2.1 (± 0.28)
0.1 (± 0.03)
0 (± 0.02)
0 (± 0.01)
0 (± 0.01)
0.4 (± 0.05)
2 (± 0.24)
4 (± 0.64)
0.7 (± 0.11)
0.5 (± 0.16)
3 (± 0.46)
0.1 (± 0.07)
34.5 (± 3.44)
5.8 (± 1.06)
1 (± 0.18)
0.1 (± 0.07)
8.5 (± 1.29)
17.7 (± 3.08)
5 (± 0.66)
1.5 (± 0.42)
0.3 (± 0.07)
0.1 (± 0.04)
0.1 (± 0.03)
1.1 (± 0.5)
0.9 (± 0.17)
0.9 (± 0.17)
0.6 (± 0.09)
0.3 (± 0.05)
5.2 (± 0.9)
0.1 (± 0.05)
0.3 (± 0.07)
3.1 (± 0.49)
0.2 (± 0.05)
0 (± 0.02)
0 (± 0.02)
0 (± 0.02)
0.3 (± 0.07)
3.5 (± 0.41)
4.7 (± 0.85)
0.7 (± 0.17)
Σn-3
Σn-6*
Σn-3:Σn-6***
ΣSFA
ΣPUFA***
ΣMUFA
1.4 (± 0.7)
1.6 (± 0.68)
0.9 (± 0.11)
49.3 (± 2.56)
23.9 (± 2.68)
26.9 (± 3.41)
1.9 (± 1.73)
1.2 (± 0.67)
1.2 (± 0.51)
50.3 (± 6.83)
20.5 (± 5.3)
29.2 (± 4.91)
0.7 (± 0.23)
0.7 (± 0.33)
1.2 (± 0.14)
51.1 (± 3.32)
19.5 (± 2.37)
29.4 (± 3.37)
1.6 (± 0.96)
1.1 (± 0.49)
1.5 (± 0.13)
47.7 (± 3.17)
21.2 (± 2.38)
31.1 (± 4.23)
30
16
31
29
n eels
REFERENCES:
1Harrod et al (2005) Oecologia 144, 673-683; 2Tsukamoto et al (1998) Nature 396 635-636, ;
3Tzeng et al (2000) Mar Biol 137 93-98; 4Smith et al. (1996) CJFAS 53, 272-279; 5Henderson &
Tocher (1987) Prog Lipid Res 26, 281-347.
4. CONCLUSIONS:
» Significant varition in eel C & N stable isotope ratios
and fatty acid profiles associated with different
salinity zones.
» DFA classification success similar using SIA (70 %)
and FAA (77%) data collected from same individual.
» Combined analyses of SIA & FAA data allows almost
complete classification success (93%) and provides
further evidence for population sub-structuring in
coastal eel populations.
» This study demonstrates the potential for
combined use of SIA and FAA techniques to identify
individual variation in resource use within animal
populations.
ACKNOWLEDGEMENTS:
Funding: Max Planck Society (Germany) & HEA (Ireland). Laboratory: Heinke Buhtz & Anita
Möller (MPIL). Field: Eoin Macloughlin, Conor Graham & Jennie Mallela. This work forms a
contribution to the HEA PRTI-3 funded project: Population biology of eels in Irish marine and
mixohaline waters. The work could not have gone ahead without permissions kindly granted
by the Western Regional Fisheries Board and the Screebe Fishery, Ireland.