Content uploaded by Kimmo K Kahilainen
Author content
All content in this area was uploaded by Kimmo K Kahilainen
Content may be subject to copyright.
ORIGINAL PAPER
The role of gill raker number variability in adaptive
radiation of coregonid fish
Kimmo K. Kahilainen
•
Anna Siwertsson
•
Karl Ø. Gjelland
•
Rune Knudsen
•
Thomas Bøhn
•
Per-Arne Amundsen
Received: 24 February 2010 / Accepted: 13 July 2010 / Published online: 27 July 2010
Ó Springer Science+Business Media B.V. 2010
Abstract Gill raker divergence is a general pattern in adaptive radiations of postglacial
fish, but few studies have addressed the adaptive significance of this morphological trait in
foraging and eco-evolutionary interactions among predator and prey. Here, a set of sub-
arctic lakes along a diversifying gradient of coregonids was used as the natural setting to
explore correlations between gill raker numbers and planktivory as well as the impact of
coregonid radiation on zooplankton communities. Results from 19 populations covering
most of the total gill raker number gradient of the genus Coregonus, confirm that the
number of gill rakers has a central role in determining the foraging ability towards
zooplankton prey. Both at the individual and population levels, gill raker number was
correlated with pelagic niche use and the size of utilized zooplankton prey. Furthermore,
the average body size and the abundance and diversity of the zooplankton community
decreased with the increasing diversity of coregonids. We argue that zooplankton feeding
leads to an eco-evolutionary feedback loop that may further shape the gill raker mor-
phology since natural selection intensifies under resource competition for depleted prey
communities. Eco-evolutionary interactions may thus have a central role creating and
maintaining the divergence of coregonid morphs in postglacial lakes.
Keywords Ecological speciation Foraging trait Polymorphism Vendace
Whitefish morphs
K. K. Kahilainen
Department of Environmental Sciences, University of Helsinki, P.O. Box 65, 00014 Helsinki, Finland
K. K. Kahilainen (&)
Kilpisja
¨
rvi Biological Station, University of Helsinki, Ka
¨
sivarrentie 14622, 99490 Kilpisja
¨
rvi, Finland
e-mail: kimmo.kahilainen@helsinki.fi
A. Siwertsson K. Ø. Gjelland R. Knudsen T. Bøhn P.-A. Amundsen
Department of Arctic and Marine Biology, Faculty of Biosciences, Fisheries and Economics,
University of Tromsø, 9037 Tromsø, Norway
T. Bøhn
GenØk – Centre for Biosafety, The Science Park, P.O. Box 6418, 9294 Tromsø, Norway
123
Evol Ecol (2011) 25:573–588
DOI 10.1007/s10682-010-9411-4
Introduction
In adaptive radiation, a common ancestor is diverged into two or more species via eco-
logical processes and morphological adaptations to utilize different niches (Schluter 2000;
Grant and Grant 2008). Foraging trait evolution in relation to adaptive radiations has been
intensively studied in simplified and isolated ecosystems such as distant islands or their
continental counterparts, newly formed lakes (Dieckman et al. 2004; Losos and Ricklefs
2009). A classic text book example is the adaptive radiation of the beak size and shape
of Geospiza spp., where a common ancestor has diversified into a variety of species
specialized to feed on specific types of plant seeds within a wide range of seed sizes and
hardnesses (Grant and Grant 2008). In fishes, the adaptive radiation of East African
cichlids represents an excellent example of distinct morphological adaptations of head and
jaws correlated with specific foraging niches (Clabaut et al. 2007; Salzburger 2009).
However, adaptive radiations also occur in much less diverse environments such as in
many fish lineages in postglacial lakes (Schluter 1996). The general pattern is a divergence
along the pelagic-benthic resource axis, where morphological adaptations in body and head
shape seem to be important in the radiation process (Schluter and McPhail 1993; Robinson
and Parsons 2002; Amundsen et al. 2004a). We focus on one of these traits, the gill raker
number, as surprisingly few large scale studies have been made to reveal the adaptive
significance of this trait even though it is an important trophic trait in variety of fish species
(see e.g., Janssen 1980; Gibson 1988; Friedland et al. 2006).
Coregonid fishes have a circumpolar distribution with frequent co-occurrence of
multiple ecologically and morphologically distinct morphs (Sva
¨
rdson 1979; Bernatchez
et al. 1999; Amundsen et al. 2004b). Both ecological and genetic evidence suggests that
adaptive radiation is the most likely explanation for the observed patterns (Bernatchez 2004;
Østbye et al. 2006; Hudson et al. 2007). Different morphs of coregonids have traditionally
been identified from the number of gill rakers (Sva
¨
rdson 1952; Lindsey 1981; Bernatchez
2004) which is a heritable and ecologically important trait (Sva
¨
rdson 1979; Rogers and
Bernatchez 2007). The European whitefish (Coregonus lavaretus (L.)) is the most diverse
coregonid species, and has repeatedly and independently radiated from a common ancestor
into multiple morphs in a large number of postglacial lakes (Østbye et al. 2005). Genetic
results indicated similar divergence of pelagic and littoral morphs in replicate lakes sug-
gesting parallel evolution within each lake (Østbye et al. 2006). This previous study using
microsatellite data indicated that sympatric pelagic and littoral whitefish morphs are
genetically different (Østbye et al. 2006) and the vicinity of current study area with similar
morphs suggests that genetic differences are likely to exist. Due to highly similar radiation
patterns of morphs in different lakes, we clustered whitefish as three different groups
according to their specific ecomorphology. Here, whitefish exhibit distinct morphs for all
three principal lake habitats (i.e., the littoral, profundal and pelagic), in which each has
specific prey resources (Kahilainen et al. 2003, 2005; Jensen et al. 2008). The littoral is
structurally complex with diverse benthic resources, comprising a sharp contrast to the low
light conditions and scanty sediment-buried benthic resources in the profundal habitat (i.e.,
the deep benthic zone). The pelagic zone is a structurally homogenous habitat providing
zooplankton resources for fish. These principal lacustrine habitats can be considered as
peaks in an adaptive landscape that requires morphological adaptations to enhance utili-
zation of their specific diet resources. Accordingly, one should expect morphs from these
principal habitats to differ in important foraging related traits such as the gill raker
apparatus (Schluter and McPhail 1993; Robinson and Parsons
2002; Amundsen et al.
2004a).
574 Evol Ecol (2011) 25:573–588
123
The role of the gill raker apparatus is related to prey retention efficiency, where the gill
rakers function as a cross-flow filter (Sanderson et al. 2001; Smith and Sanderson 2008).
An increasing number of gill rakers enhance crossflow filtering and the closely spaced gill
rakers also limit the escape possibilities of small prey. However, a dense gillraker appa-
ratus is more likely to be clogged by sediments than more sparse gillrakers, and foraging in
the muddy bottom of the profundal most likely require other gillraker adaptations.
Accordingly, a high number of long gill rakers is common in planktivorous fish species and
morphs, whereas benthic species and morphs usually display a lower number of shorter gill
rakers (Janssen 1980; Schluter and McPhail 1992; Robinson and Parsons 2002). Corego-
nids have a wider gillraker range than other polymorphic fish lineages and thus represent
an excellent candidate taxon to evaluate the significance of such phenotype-environment
associations. Furthermore, the principal prey resource associated with this trait (i.e.,
zooplankton) can be examined in detail qualitatively and quantitatively both in the envi-
ronment and the predator diet. Such comparisons in natural settings are ideal to explore the
adaptive significance of the predator’s functional morphology. In their seminal paper,
Brooks and Dodson (1965) revealed that size selective predation of planktivorous fish
alters the species composition and reduces the body size of prey communities. This has
lead to a wide consensus that planktivorous fish regulates zooplankton communities (Zaret
1980; Lampert and Sommer 2007). When a proportion of fish population is adapting to
a zooplankton resource, the zooplankton community response by decreased body sizes
provides a feedback loop that further strengthen the selection pressure towards high for-
aging efficiency on small prey items. Such eco-evolutionary interactions have rarely been
addressed in relation to adaptive radiations of postglacial fish.
Here, we used a set of subarctic lakes that comprises a diversity gradient of coregonid
assemblages with increasing range and numbers of gill rakers, including (1) monomorphic
whitefish with ca. 20–30 gill rakers, (2) polymorphic whitefish populations with ca. 15–40
rakers, and (3) polymorphic whitefish and vendace, Coregonus albula (L.) with ca. 15–50
rakers). This range constitutes a natural setting to explore the role of increasing gill raker
numbers in zooplankton foraging, including the impact of coregonid radiation on zoo-
plankton prey communities. We assumed that foraging efficiency is associated with the
ability to utilize small prey, and predicted that zooplankton prey utilization is correlated to
the gill raker number. Furthermore, we predicted that zooplankton size, density and
community structure would change along the gradient from monomorphic to polymorphic
and finally to polymorphic whitefish and vendace lakes. Such a pattern would provide an
eco-evolutionary feedback mechanism where the prey community over evolutionary time
(e.g., under adaptive radiation) could shape the morphology of the predator.
Materials and methods
Study area and fish populations
We examined a set of eight northern Fennoscandian postglacial lakes situated in the
large subarctic Paatsjoki/Pasvik watercourse, including five Finnish (Lakes Aksuja
¨
rvi,
Vuontisja
¨
rvi, Vastusja
¨
rvi, Muddusja
¨
rvi and Paadar) and three Norwegian (Lakes Ellentj-
ern, Tjærebukta and Skrukkebukta) lakes (Fig. 1). This set of lakes represents a wide
gradient of coregonid populations. The Finnish headwater lakes discharge into the large
Lake Inarija
¨
rvi (hereafter L. Inari), whereas Norwegian lakes are situated in the lower
reaches of the watercourse (Fig. 1b). The study lakes are all oligotrophic (tot P 3–9 lgl
-1
,
Evol Ecol (2011) 25:573–588 575
123
tot N 145–240 lgl
-1
), well-oxygenated with neutral pH-values (6.8–7.2). Surface areas
range from 1 to 48 km
2
and maximum depths from 7 to 73 m (Table 1). The ice-free
season generally lasts from May–June to October–November. Coregonids, represented by
three different whitefish morphs and vendace, are the main zooplankton predators and the
dominant fish species (70–91% of numerical catches) in all lakes, except L. Ellentjern, but
the composition of the coregonid assemblage differs among the lakes.
The lakes were classified according to an increasing diversity of coregonids. Lakes
Aksuja
¨
rvi (hereafter L. Aksu), Ellentjern and Vuontisja
¨
rvi (L. Vuontis) were classified as
type 1, with only one whitefish morph present, the large sparsely rakered (LSR) morph.
LSR whitefish is identified and named according to body size and number of gill rakers,
which usually ranges from approx. 20–30 (Fig. 2). In lake type 2, LSR whitefish co-exist
with a densely rakered (DR) whitefish morph with approx. 30–40 gill rakers (L. Vastus) or
with DR whitefish and a small sparsely rakered (SSR) whitefish morph with approx. 15–20
gill rakers (Lakes Muddus and Paadar). In the most complex lake type 3 (Lakes Tjærebukta
and Skrukkebukta), polymorphic whitefish (LSR, DR and SSR whitefish morphs) co-exist
with vendace, which has the highest number of gill rakers (approx. 40–50) (Fig. 2).
Vendace is a pelagic zooplankton specialist (Helland et al. 2008) and does not occur
naturally in the Paatsjoki/Pasvik watercourse (Amundsen et al. 1999). Vendace was
introduced to L. Inari in the 1950–1960s and formed a very dense population during the
1980s leading to an invasion and colonization of the lower Paatsjoki lakes around 1990
(Amundsen et al. 1999). As a superior planktivore competitor over DR whitefish, vendace
has become the dominant fish species in the pelagic food web in many lakes in the lower
parts of the Paatsjoki/Pasvik watercourse (Bøhn and Amundsen 2001; Gjelland et al. 2007;
Bøhn et al. 2008).
Fish sampling
Sampling was conducted during September (years 2000–2007) in all the lakes. A long
sampling period was needed to include several lakes and coregonid populations from both
countries. This should not have any significant influence on the main patterns, since gill
Fig. 1 Map of (a) the northern Fennoscandia and (b) Paatsjoki/Pasvik watercourse. Study lakes with lake type
definition indicated in the parenthesis. 1 monomorphic whitefish, 2 polymorphic whitefish, 3 polymorphic
whitefish and vendace
576 Evol Ecol (2011) 25:573–588
123
Table 1 Background data on location, morphometry, water chemistry and fish fauna of the study lakes
Parameter Lake Aksu Lake Ellentjern Lake Vuontis Lake Vastus Lake Muddus Lake Paadar Lake Skrukkebukta Lake Tjærebukta
Lake type 1 1 1 2 2 2 3 3
Latitude (°N) 69°14
0
69°12
0
69°01
0
69°03
0
69°00
0
68°52
0
69°33
0
69°13
0
Longitude (°E) 26°53
0
29°06
0
27°04
0
27°07
0
26°50
0
26°35
0
30°07
0
29°11
0
Surface area (km
2
) 4 1 11 4 48 21 7 15
Altitude (m.a.s.l.) 206 71 151 146 146 144 21 52
Max depth (m) 10 7 31 15 73 56 38 30
Mean depth (m) 3.5 2.5 6.5 2.7 8.5* 11.7 14 4
Secchi depth (m) 2.5 4.5 8 2 3 6* 4–5.5 3–4.5
pH – 6.9 7.2* 7.0 7.2* 7.1* 6.9 6.8
Tot P (lgl
-1
) – 3 7* 7 5* 6* 7 9
Tot N (lgl
-1
) – 165 170* 240 160* 160* 156 145
Coregonid
proportion (%)
81 39 90 70 86 91 85 78
Species/morphs
present
b, f, g, i, j,
k, l, m
b, g, i,
j, k, l
b, f, g, h, i,
j, k, l, m
a, b, f, g, h,
i, j, k, l, m
a, b, c, e, f,
g, h, i, j, k, l, m
a, b, c, f, g,
h, i, j, k, l, m
a, b, c, d, f,
g, i, j, k, l, m
a, b, c, d, f,
g, i, j, k, l, m
Coregonids and other fish species present in the study lakes are indicated with abbreviations. Lake type refers to the diversity of coregonid fish communities (1 monomorphic
whitefish, 2 polymorphic whitefish, 3 polymorphic whitefish and vendace)
* Data from Lapland Regional Environment Centre; a DR whitefish, b LSR whitefish, c SSR whitefish, d vendace, e Arctic charr, f grayling (Thymallus thymallus (L.),
g minnow (Phoxinus phoxinus (L.)), h three-spined stickleback, i nine-spined stickleback Pungitius pungitius (L.)), j perch (Perca fluviatilis L.), k pike (Esox lucius L.),
l burbot (Lota lota (L.)), m brown trout (Salmo trutta L.)
Evol Ecol (2011) 25:573–588 577
123
raker traits, habitat and diet selection of the studied coregonid populations are highly stable
among different years (Amundsen et al. 2004a, b; Kahilainen et al. 2004, 2007, 2009).
Coregonids were sampled from the three main habitats (littoral, pelagic and profundal)
using a combination of gill net series and pelagic trawling. The Finnish lakes were sampled
using a gill net set with eight nets, each having a length of 30 m and a height of 1.8 m, with
mesh sizes 12, 15, 20, 25, 30, 35, 45, and 60 mm from knot to knot. In addition, we used a
small pair trawl (5 m high, 8 m wide and cod-end mesh size 3 mm) in the pelagic zone of
these headwater lakes (see Kahilainen et al. 2004 for details). In the Norwegian lakes
coregonids were caught in the littoral and profundal habitats using benthic gill nets series
(length 40 m and height 1.5 m) with the mesh sizes of 10, 12.5, 15, 18.5, 22, 26, 35, and
45 mm, and in the pelagic using floating gill net series (length 40 m and height 6 m) with
mesh sizes of 8, 10, 12.5, 15, 18.5, 22, 26, and 35 mm.
Coregonids were field-identified to morph/species according to their overall habitus, head
shape and gill rakers (Amundsen et al. 2004b; Kahilainen and Østbye 2006). Minor overlap of
gill raker counts exist between the whitefish morphs, but these individuals can be classified
using combined information from body, head and gill rakers morphology. Uncertain SSR
whitefish can be defined from LSR whitefish due to its very peculiar habitus with large eye,
robust head, pronounced subterminal mouth and short bend gill rakers (Kahilainen and
Østbye 2006; Harrod et al. 2010). Uncertain DR whitefish was classified according to longer
gill rakers, terminal mouth and pointed head shape (Amundsen et al. 2004b; Harrod et al.
2010). Vendace can be separated from DR whitefish accurately as it has a characteristic
up-pointing protruding lower jaw, very pointed head, and very long and slender gill rakers.
The number of gill rakers was counted from the first left gill arch under a preparation
microscope. Stomachs were removed and prey items were identified as accurately as
possible. The relative contribution of each prey category was estimated (Amundsen et al.
1996). The coregonids diet consisted of pelagic zooplankton (mainly Bosmina spp.,
Daphnia spp., Holopedium gibberum, Calanoid and Cyclopoid copepods) and benthic
Fig. 2 Combined gill raker
distributions of whitefish morphs
(SSR small sparsely rakered,
LSR large sparsely rakered,
DR densely rakered) and vendace
in study lakes. Line illustrations
present the first left gill arch and
gill raker morphology of different
whitefish morphs
578 Evol Ecol (2011) 25:573–588
123
invertebrates (mainly molluscs, insect larvae and some benthic crustaceans). In the present
study, we focused on zooplankton prey, which is the only diet category considered here-
after. Body length of up to 30 individuals of undigested zooplankton was measured from
each stomach, when possible. In copepods, we measured the length from rostrum to furca
and in cladocerans from head to base of the tail spine (Kahilainen et al. 2005).
Zooplankton sampling
Zooplankton was sampled in September from the whole water column using two replicate
samples from each lake. In the Finnish lakes, samples were taken with a Limnos-tube (1 m,
volume 7.1 l) and zooplankton net (diameter 25 cm, mesh size 50 lm). Tube samples were
sieved through 50 lm zooplankton net and all samples were stored in 5% formalin solu-
tion. In the Norwegian lakes, samples were taken with a 30 l Schindler–Patalas trap or with
a vertically hauled zooplankton net (diameter 26 cm, mesh size 90 lm) and stored in 4%
formalin solution. Differences in sampling gears between countries may have effect on
zooplankton community results. The somewhat larger mesh size of zooplankton net and
higher volume of Schindler–Patalas trap used in Norwegian lakes could be more effective
to capture larger individuals as well as higher density and diversity of zooplankton com-
munity than the smaller gear used in Finnish side (Kalff 2002). We recognize this potential
bias in the interpretation of results and take this into account in prey size comparisons
among coregonids. However, the smallest zooplankton found in coregonid diet (0.30 mm,
Bøhn and Amundsen 1998) is substantially larger than the largest zooplankton net mesh
sizes used in this study (0.09 mm), ensuring that both sampling methods have captured
zooplankton sizes available to fish. Zooplankton samples were counted and measured in
the laboratory, excluding nauplii since they were not observed in the fish diet. The body
length of 30–50 randomly selected individuals from each zooplankton taxa (Bosmina,
Daphnia, Calanoid and Cyclopoid copepods) was measured. The average zooplankton
density per litre in the whole water column and the relative proportion of the main taxa
were calculated.
Statistical analyses
At the population level, the average number of gill rakers was compared with the proportion
of pelagic habitat use and diet as well as zooplankton prey size in the stomachs using
Spearman correlations. The same approach was used to explore potential correlations
between gill raker number and the proportion of pelagic diet and zooplankton prey size at the
individual level. In individual diet data, we calibrated datasets according to the lowest
samples sizes per morph/lake and then used random re-sampling for other morph/lake
combinations. Differences in zooplankton prey length data among morphs/species types
were harmonized by random re-sampling of 25 samples from a morphs/species within a lake
when available. This approach enabled separation of effects at the individual and population
levels. Differences in the zooplankton prey size among whitefish morphs and vendace were
tested with analysis of covariance (ANCOVA) using the zooplankton average length in each
lake as a covariate. The effect of individual gill raker number on median zooplankton prey
length in the stomach was tested with a general linear model (GLM) using gill raker number,
mean zooplankton length in the environment, and species/morph as predictor variables.
The effect of individual gill raker number on median zooplankton prey length was finally
tested with regressions within each species/morph on the full dataset, using prey length as the
response and gill raker number as the predictor variable.
Evol Ecol (2011) 25:573–588 579
123
A GLM was used to test for zooplankton size differences among lake types using
different zooplankton taxa and lake type as categorical variables. Analysis of variance
(ANOVA) was used to test for differences in the average zooplankton abundance (log
(x ? 1) transformed data) among lake types. The number of samples used for zooplankton
length measurements in coregonid stomachs and in the environment was different among
the lakes and we used bootstrapping for calibration of sample sizes before performing
ANCOVA, GLM and ANOVA. Subsequent pairwise comparisons in these analyses were
made with Tukey’s HSD tests.
Results
Strong positive correlations were found at the population level between the number of gill
rakers and both the pelagic habitat use (Fig. 3a, n = 19, Spearman correlation; r
s
= 0.83,
P \ 0.01) and the proportion of zooplankton in the diet (Fig. 3b, n = 19, r
s
= 0.84,
P \ 0.01). Polymorphic SSR and LSR whitefish populations with an average number of
gill rakers from 16–21 and 22–25, respectively, mainly used the benthic niche. Mono-
morphic LSR whitefish populations had an average number of gill rakers from 24 to 28 and
used both pelagic and benthic prey and habitat. DR whitefish had in average 33–35 gill
rakers and was the most pelagic and planktivorous whitefish morph. Vendace had on
average 43 gill rakers and was consistently a pelagic planktivore. In accordance with the
niche utilization, a significant negative correlation was observed between the average
number of gill rakers and zooplankton prey size (n = 19, r
s
=-0.83, P \ 0.01) (Fig. 3c).
The ANCOVA indicated that the average length of ingested zooplankton was dependent on
coregonid morph/species (F
3,187
= 91, P \ 0.001), but not on the observed average zoo-
plankton length in the pelagic environment (F
1,187
= 2, P = 0.13). The average length of
zooplankton prey gradually decreased from 1.90 mm in SSR whitefish, 0.95 mm in LSR,
0.61 mm in DR to 0.57 mm in vendace. These size differences in ingested prey were
different between all coregonid taxa (Tukey’s HSD tests, P \ 0.001), except between DR
whitefish and vendace (P = 0.978). At the individual level, the general patterns observed
at the population level were supported by a positive correlation between the number of gill
rakers and the proportion of zooplankton in diet (Fig. 4a, n = 48, r
s
= 0.68, P \ 0.01) and
Fig. 3 Correlations between gill raker number and (a) the proportion of pelagic habitat use, (b) the
proportion of zooplankton in the diet and (c) the zooplankton prey size at the population level. Population
types are marked with different labels: SSR whitefish (white circle), LSR whitefish (grey triangle), DR
whitefish (white square) and vendace (black diamond)
580 Evol Ecol (2011) 25:573–588
123
a negative correlation between number of gill rakers and zooplankton prey size (Fig. 4b,
n = 245, r
s
=-0.69, P \ 0.01). The GLM-analysis (adj. r
2
= 0.59) confirmed a strong
effect of morph/species on median zooplankton prey size (P \ 0.05 for all morphs/spe-
cies), and also indicated a negative effect of individual gill raker number within the morph/
species (P = 0.082). The regressions within each morph/species revealed that the negative
effect of individual gill raker number on zooplankton prey size was significant only within
the LSR and SSR whitefish (Table 2).
Fig. 4 Individual level
correlations between gill raker
number and (a) the proportion of
zooplankton in the diet and b the
average zooplankton prey size.
Population types are marked
with different labels:
SSR whitefish (white circle),
LSR whitefish (grey triangle),
DR whitefish (white square)
and vendace (black diamond)
Table 2 Results from the regression model prey length (in mm) = constant (a) ? gill raker number (Grn)
Species/morph aP(a) Grn P(Grn) Adj. r
2
P(overall) n
SSR 2.9 \0.001 -0.052 0.002 0.1 0.002 88
LSR 2.3 \0.002 -0.043 0.007 0.03 0.007 212
DR 0.23 0.46 0.011 0.24 0.001 0.24 331
Vendace 0.8 0.017 -0.007 0.22 0.18 0.22 6
Level of significance (P) is included for constant, gill raker number and overall model. Zooplankton length
in the environment was initially included in the models, but removed from all as it had no significant
contribution
Fig. 5 Zooplankton (a) body length, (b) density and (c) community composition along an increased diversity
gradient of coregonids (lake types: 1 monomorphic whitefish, 2 polymorphic whitefish, 3 polymorphic
whitefish and vendace). Zooplankton taxa indicated in bars are Bosmina spp. (white), Daphnia spp. (grey),
Holopedium gibberum (black), Calanoid (vertical hatching) and Cyclopoid copepods (diamond hatching).
In lake type 3, Daphnia spp. refers mainly to Daphnia cristata
Evol Ecol (2011) 25:573–588 581
123
There were distinct trends in the zooplankton community structure along the increasing
coregonid diversity gradient (Fig. 5). The average size of zooplankton differed among lake
types (GLM, F
2,1025
= 67, P \ 0.01) and gradually decreased from 0.65 mm in lake type
1, 0.60 mm in type 2 to 0.54 mm in lake type 3 (Tukey’s HSD tests, P \ 0.05) (Fig. 5a).
A sign of decreasing zooplankton abundance along the coregonid diversity gradient was
observed (Fig. 5b), but this was not statistically significant (ANOVA, F
2, 13
= 0.41,
P = 0.67). The average density in lake type 1 (8.2 ind l
-1
) tended to be higher than in lake
type 2 and 3 (5.4 and 5.8 ind l
-1
, respectively). Zooplankton community composition
changed from equal proportions of copepods and cladocerans in lake type 1 to a clear
dominance of cladocerans in lake type 3 (Fig. 5c). This was mainly due to a decrease in
the proportion of cyclopoid copepods and an increase in the proportion of Daphnia spp.,
in particular the small-sized and transparent species D. cristata, in lake type 3.
Discussion
We documented a strong relationship between gill raker numbers and the degree of
planktivory; a pattern that appears to be common in polymorphic fish populations in the
northern hemisphere (Schluter and McPhail 1992; Sku
´
lason et al. 1999; Amundsen et al.
2004a). The current study extended this common pattern to a much larger scale by includ-
ing all principal habitat types and a very wide range of gill raker number utilizing 19
different populations. There were strong positive correlations between predator trophic
morphology (gill rakers) and pelagic niche utilization (habitat and diet) as well as an
adaptive significance of increasing number of gill rakers facilitating the utilization of
smaller prey. The study furthermore extends the link between gill raker traits and niche
utilization from the commonly occurring littoral-pelagic morph pairs of various fish spe-
cies in the northern hemisphere (e.g., Schluter and McPhail 1993; Robinson and Parsons
2002), to also include the far less explored profundal niche. The fish with the lowest gill
raker numbers (\20) were almost exclusively associated with the profundal habitat, the
intermediate numbers (20–30) mainly with the littoral, whereas the highest numbers
(whitefish: 30–40; vendace [ 40) were associated with the pelagic habitat and a typical
zooplanktivore niche. These phenotype-environment correlations proved to be strong both
at the individual and population levels, suggesting that gill raker trait divergence is central
in adaptive radiation of whitefish between these three principal habitats of subarctic lakes.
The number of gill rakers is a single heritable trait in coregonid fishes (Sva
¨
rdson 1979;
Rogers and Bernatchez 2007), but apparently it also effectively captures much of other
morphological traits. Several trophic traits (i.e., head and body morphology) are associated
to fish feeding niche utilization along traditional pelagic-littoral resource axis in many
postglacial fish morphs (Schluter 1996; Robinson and Parsons 2002), but very little is
known about profundal adaptations. The SSR whitefish typically residing in the profundal
habitat, has the lowest gill raker counts among the explored whitefish morphs and a body
and head morphology that likely have an adaptive value in profundal foraging (Kahilainen
and Østbye 2006). Similar adaptations in trophic related traits were shown to be heritable
in a profundal Arctic charr (Salvelinus alpinus (L.)) morph specializing on soft bottom
benthos (Klemetsen et al. 2002; Knudsen et al. 2006). Foraging on prey items buried in soft
bottom profundal sediments requires some suction of mud (Kahilainen et al. 2003). A low
number of short, widely spaced gill rakers is probably sufficient to retain typical profundal
prey types (i.e., Pisidium bivalves and chironomid larvae) while allowing the mud to be
disposed through the gillraker slits (Kahilainen and Østbye 2006). A dense gillraker
582 Evol Ecol (2011) 25:573–588
123
apparatus would in contrast likely be clogged by mud (Amundsen et al. 2004b). The SSR
whitefish mainly consumed relatively large-sized prey, suggesting a limited foraging
efficiency on zooplankton. The LSR whitefish morph has intermediate numbers, length and
spacing of gill rakers and subterminal mouth which likely facilitate benthic foraging
(Kahilainen and Østbye 2006; Harrod et al. 2010). LSR whitefish is apparently less effi-
cient in predation of small-sized zooplankton than the specialized planktivore DR whitefish
morph that has large number of long and dense gill rakers and a terminal mouth and slender
body shape (Kahilainen and Østbye 2006). These morphological traits of DR whitefish are
well suited for pelagic planktivores (Webb 1984) and are likely to have evolved in the
absence of resource competitors like ciscoes/vendace (Bernatchez 2004; Bøhn et al. 2008).
The differences between coregonids in gill raker apparatus can be compared to the
divergence of beak shape in birds, jaw shape in amphibians, mandible shape in bats or baleen
plates in whales which all facilitate the use of different dietary niches (Werth 2004; Pfennig
et al. 2006; Price 2008; Nogueira et al. 2009). In fish, there is a common trend of increasing
number of gill rakers from piscivores to benthivores and finally to planktivores (Gibson
1988; Langeland and Nøst 1995). Our results on Coregonus demonstrate a similar intra-
genus benthivore-planktivore trend in gill raker numbers. Our field data furthermore show a
negative correlation between the number of gill rakers and zooplankton prey size both at the
population and individual levels. Zooplankton prey is available in all principal lake habitats
(littoral, profundal and pelagic zones), providing an opportunity for planktivory for all
whitefish morphs. The gill raker apparatus functions as a crossflow filter that directs prey
particles towards the oesophagus (Sanderson et al. 2001), and explains why increasing
number of gill rakers facilitates the retention of smaller prey sizes. Previous studies failing to
find similar correlations between gill raker traits and prey size in salmonids (Seghers 1975;
Sandlund et al. 1987; Budy et al. 2005), may not have captured the essential range of trait
variation that is demonstrated among the coregonids in the present study.
The observed strong correlation between gill raker number and prey utilization at the
individual level suggests a significant role of gill rakers in individual foraging efficiency
that may promote disruptive selection. Adaptive evolution and divergence of trophic traits
are generally linked to unequal utilization efficiency of prey resources between individuals
(Knudsen et al. 2007; Arau
´
jo et al. 2008), which may ultimately lead to differences in
fitness and promote disruptive selection that may act in the formation of new morphs
(Rueffler et al. 2006). In a monomorphic three-spined stickleback (Gasterosteus aculeatus
L.) population, Bolnick and Lau (2008) found evidence for disruptive selection via
intraspecific competition, as individuals with high or low gill raker counts had higher
growth rates than individuals with intermediate gill raker numbers. In addition, if mating is
assortative between phenotypically and ecologically similar individuals, the disruptive
selection provides a pathway to population divergence into morphs (Snowberg and Bolnick
2008) and subsequently to speciation (Dieckmann and Doebeli 1999; Schluter 2000).
Monomorphic LSR whitefish with intermediate number of gill rakers is the most common
population type in northern Fennoscandia and probably represents the ancestral morpho-
type (Østbye et al. 2006), since allopatric SSR or DR whitefish populations have not been
found in the region (Lehtonen and Niemela
¨
1998; Amundsen et al. 2004b). During the
early colonization of these postglacial lakes, ecological opportunities have presumably
been high for specialization to each of the principal habitat types and their associated prey
communities. These three principal trophic niches may promote disruptive selection on gill
raker traits by constituting peaks in an adaptive landscape, where each whitefish morph has
adapted morphologically to utilize one of these peaks. Monomorphic LSR whitefish with
intermediate gill raker number use all the principal lake habitats foraging both on
Evol Ecol (2011) 25:573–588 583
123
zooplankton and benthic macroinvertebrates (Amundsen et al. 2004b; Kahilainen et al.
2007). In sympatry with other morphs (i.e., in polymorphic lakes) the LSR whitefish
prefers littoral macroinvertebrates, whereas the SSR whitefish utilizes profundal benthos
and DR whitefish zooplankton (Harrod et al. 2010). Interestingly, the effect of gill raker
number on zooplankton prey size was strongest in the SSR whitefish, weaker but still
significant in the LSR whitefish, and with no significant effect in the DR whitefish and
vendace. This suggests a directional selection towards increasing gillraker numbers for
SSR and LSR whitefish individuals that utilize a planktivorous niche, whereas there seems
to be little support for directional selection on increasing gill raker number in DR whitefish
or vendace in these lakes. Taken collectively, our results support a scenario where LSR
whitefish has diverged into SSR and DR whitefish morphs via disruptive selection
primarily acting on gill raker morphology and foraging abilities (Østbye et al. 2006).
The differences in zooplankton community structure among the three lake types suggest a
general importance of gill raker numbers in relation to planktivore predation. Although the
sampling was performed only in September, the previous seasonal open water datasets of
zooplankton and niche utilization of whitefish morphs and vendace support the observed
patterns in this study (Bøhn and Amundsen 1998, 2001; Kahilainen et al. 2004, 2005;
Gjelland et al. 2009). However, there is a need for winter sampling during ice cover when
zooplankton community is certainly different due to lack of cladocerans (Tolonen 1998) and
niche utilization of coregonids may also differ (Jurvelius and Marjoma
¨
ki 2008). In this study,
we found that zooplankton body size and density decreased with increasing coregonid
diversity, a pattern commonly observed in zooplankton communities when the number of
specialized planktivorous fish species increases (Nilsson and Pejler 1973; Post et al. 2008;
Amundsen et al. 2009). However, this pattern has previously not been connected to adaptive
radiation in postglacial fish. These zooplankton community patterns could have been even
stronger, if sampling gear had been identical. The vendace-whitefish lakes (lake type 3) were
sampled using larger zooplankton gear that may have increased the average length of zoo-
plankton. In lakes with only LSR whitefish present the competition for zooplankton resources
in the pelagic habitat is expected to be weak. Accordingly, we observed large body size, high
density and wide diversity of zooplankton in these lakes. In lakes including DR whitefish,
however, increased competition for zooplankton was indicated by reduced body size, density
and availability of zooplankton. Under such conditions, the frequency of planktivory in the
LSR whitefish is low (Amundsen et al. 2004b; Kahilainen et al. 2007). This trend was even
more pronounced in polymorphic lakes with vendace present as both the LSR and even the
DR whitefish morphs were forced to utilize the benthic food resources (Bøhn and Amundsen
2001; Bøhn et al. 2008). Hence, in each step of increased coregonid diversity, predation
efficiency for zooplankton increases and accordingly modifies the zooplankton community.
Subsequently, this reduces the opportunities of SSR and LSR whitefish morphs to utilize the
zooplanktivore dietary niche. We argue that this represents an eco-evolutionary process with
a feedback loop that reduces the formation of intermediate phenotypes (and hybrids), and
increases resource segregation among morphs. Similar feedback loops between predator
morphology and resources have been found in zooplanktivore alewife Alosa pseudoharengus
populations (Palkovacs and Post 2008, 2009) and in seed-feeding Geospiza finches (Grant
and Grant 2008). This process is able both to create and maintain polymorphism in various
ecosystems, and may over time lead to the formation of new species. Our data represent
empirical support for the early stages of this process in pristine and relatively young fish
communities. In a broader perspective, including the well known adaptive radiation in much
older systems (like e.g., the speciation of cichlids), a profound link between ecological and
evolutionary timescales is strongly indicated (see also Hairston et al. 2005).
584 Evol Ecol (2011) 25:573–588
123
In conclusion, our study demonstrates the adaptive significance of gill rakers in for-
aging: an increasing number of gill rakers facilitates the utilization of smaller prey and is
advantageous to planktivory, but at the same time disadvantageous to benthivory, in
particular to feeding in the profundal sediments (Fig. 6). Apparently, the three principal
lacustrine habitats represent adaptive peaks, promoting disruptive selection leading to gill
raker divergence and polymorphism. The phenotype-environment correlations between
gill raker number and pelagic niche utilization proved to be strong both at the indi-
vidual and population levels. Evidently, the coregonid gill raker divergence influ-
ences the zooplankton community structure and likely creates an eco-evolutionary
feedback loop maintaining and possibly strengthening the segregation of pelagic and
benthic morphs.
Acknowledgments The authors thank the Ministry of Agriculture and Forestry, Municipality of Inari,
Finnish Cultural Foundation, Ella and Georg Ehrnrooth Foundation, Otto A. Malm Foundation, Emil
Aaltonen Foundation, European Regional Developmental Fund (project A30205), The Norwegian Research
Council (NFR 186320/V40 and 183984/S30), Norwegian Directorate for Nature Management, The County
Governor of Finnmark and Pasvik Kraft AS for funding. We also acknowledge the field and laboratory work
by Aikio O., Antti-Poika P., Dalsbø L., Eloranta A., Helminen M., Johannesen K.S., Johansson K.,
Ja
¨
a
¨
skela
¨
inen P., Kervinen J., Lien C., Marttila J., Ma
¨
enpa
¨
a
¨
K., Niemisto
¨
J., Pennanen, M., Pohtila J.,
Salonen, M., Sa
´
ren, J., Solberg K.G., Tuomaala, A. and Vatanen S. Muddusja
¨
rvi Research Station kindly
provided facilities during the field sampling. We like to thank White E. and Antti-Poika P. for line
illustrations and Malinen T. for comments on manuscript.
Fig. 6 Ecomorphological gradient of studied coregonid populations. Whitefish morphs (SSR small sparsely
rakered, LSR large sparsely rakered, DR densely rakered) and vendace body shapes are illustrated with line
drawings. Normal distributions illustrate niche widths and accompanying text indicates main habitat, diet
and ecological classification of different coregonids. Lowest arrow indicates increasing morphological
specialization towards zooplanktivory
Evol Ecol (2011) 25:573–588 585
123
References
Amundsen P-A, Gabler HM, Staldvik FJ (1996) A new approach to graphical analysis of feeding strategy
from stomach contents data—modification of Costello (1990) method. J Fish Biol 48:607–614
Amundsen P-A, Staldvik FJ, Reshetnikov YS, Kashulin N, Lukin A, Bøhn T, Sandlund OT, Popova OA
(1999) Invasion of vendace Coregonus albula in a subarctic watercourse. Biol Conserv 88:405–413
Amundsen P-A, Bøhn T, Va
˚
ga GH (2004a) Gill raker morphology and feeding ecology of two sympatric
morphs of European whitefish (Coregonus lavaretus). Ann Zool Fenn 41:291–300
Amundsen P-A, Knudsen R, Klemetsen A, Kristoffersen R (2004b) Resource competition and interactive
segregation between sympatric whitefish morphs. Ann Zool Fenn 41:301–307
Amundsen P-A, Siwertsson A, Primicerio R, Bøhn T (2009) Long-term responses of zooplankton to
invasion by a planktivorous fish in a subarctic watercourse. Freshw Biol 54:24–34
Arau
´
jo MS, Guimara
˜
es PR, Svanba
¨
ck R, Pinheiro A, Guimara
˜
es P, dos Reis SF, Bolnick DI (2008) Network
analysis reveals contrasting effects of intraspecific competition on individual vs. population diets.
Ecology 89:1981–1993
Bernatchez L (2004) Ecological theory of adaptive radiation: empirical assessment from Coregonine fishes
(Salmoniformes). In: Hendry AP, Stearns SC (eds) Evolution illuminated: salmon and their relatives.
Oxford University Press, Oxford, pp 175–207
Bernatchez L, Chouinard A, Lu G (1999) Integrating molecular genetics and ecology in studies of adaptive
radiation: whitefish, Coregonus sp., as a case study. Biol J Linn Soc 68:173–194
Bøhn T, Amundsen P-A (1998) Effects of invading vendace (Coregonus albula L.) on species composition
and body size in two zooplankton communities of the Pasvik River System, northern Norway.
J Plankton Res 20:243–256
Bøhn T, Amundsen P-A (2001) The competitive edge of an invading specialist. Ecology 82:2150–2163
Bøhn T, Amundsen P-A, Sparrow A (2008) Competitive exclusion after invasion? Biol Inv 10:359–368
Bolnick DI, Lau OL (2008) Predictable patterns of disruptive selection in stickleback in postglacial lakes.
Am Nat 172:1–11
Brooks JL, Dodson SI (1965) Predation, body size, and composition of plankton. Science 150:28–35
Budy P, Haddix T, Scheidervin R (2005) Zooplankton size selection relative to gill raker spacing in rainbow
trout. Trans Am Fish Soc 134:1228–1235
Clabaut C, Bunje PME, Salzburger W, Meyer A (2007) Geometric morphometric analyses provide evidence
for the adaptive character of the Tanganyikan cichlid fish radiations. Evolution 61:560–578
Dieckmann U, Doebeli M (1999) On the origin of species by sympatric speciation. Nature 400:354–357
Dieckman U, Doebeli M, Metz JAJ, Tautz D (eds) (2004) Adaptive speciation. Cambridge University Press,
Cambridge
Friedland KD, Ahrenholz DW, Smith JW, Manning M, Ryan J (2006) Sieving functional morphology of the
gill raker feeding apparatus of Atlantic menhaden. J Exp Zool 305A:974–985
Gibson RN (1988) Development, morphometry and particle retention capability of gill rakers in the herring,
Clupea harengus L. J Fish Biol 32:949–962
Gjelland KØ, Bøhn T, Amundsen P-A (2007) Is coexistence mediated by microhabitat segregation?—an in-
depth exploration of a fish invasion. J Fish Biol 71(Suppl. D):196–209
Gjelland KØ, Bøhn T, Horne JK, Jensvoll I, Knudsen FR, Amundsen P-A (2009) Planktivore vertical
migration and shoaling under a subarctic light regime. Can J Fish Aquat Sci 66:525–539
Grant PR, Grant BR (2008) How and why species multiply? The radiation of Darwin’s Finches. Princeton
University Press, New Jersey
Hairston NG Jr, Ellner SP, Geber MA, Yoshida T, Fox JA (2005) Rapid evolution and the convergence of
ecological and evolutionary time. Ecol Lett 8:1114–1127
Harrod C, Mallela J, Kahilainen KK (2010) Phenotype-environment correlations and character displacement
in a putative fish radiation. J Anim Ecol (in press). doi:10.1111/j.1365-2656.2010.01702.x
Helland IP, Harrod C, Freyhof J, Mehner T (2008) Co-existence of a pair of pelagic planktivorous coregonid
fishes. Evol Ecol Res 10:373–390
Hudson AG, Vonlanthen P, Mu
¨
ller R, Seehausen O (2007) Review: the geography of speciation and
adaptive radiation of coregonines. Arch Hydrob Spec Iss Adv Limnol 60:111–146
Janssen J (1980) Alewives (Alosa pseudoharengus) and ciscoes (Coregonus artedii) as selective and non-
selective planktivores. In: Kerfoot WC (ed) Evolution and ecology of zooplankton communities.
University Press of New England, New Hampshire, pp 580–586
Jensen H, Kahilainen KK, Amundsen P-A, Gjelland KØ, Tuomaala A, Malinen T, Bøhn T (2008) Predation
by brown trout (Salmo trutta) along a diversifying prey community gradient. Can J Fish Aquat Sci
65:1831–1841
586 Evol Ecol (2011) 25:573–588
123
Jurvelius J, Marjoma
¨
ki TJ (2008) Night, day, sunrise, sunset: do fish under snow and ice recognize the
difference? Freshw Biol 53:2287–2294
Kahilainen K, Østbye K (2006) Morphological differentiation and resource polymorphism in three sym-
patric whitefish Coregonus lavaretus (L.) forms in a subarctic lake. J Fish Biol 68:63–79
Kahilainen K, Lehtonen H, Ko
¨
no
¨
nen K (2003) Consequence of habitat segregation to growth rate of two
sparsely rakered whitefish (Coregonus lavaretus (L.)) forms in a subarctic lake. Ecol Freshw Fish
12:275–285
Kahilainen K, Malinen T, Tuomaala A, Lehtonen H (2004) Diel and seasonal habitat and food segregation
of three sympatric (Coregonus lavaretus (L.)) forms in a subarctic lake. J Fish Biol 64:418–434
Kahilainen K, Alaja
¨
rvi E, Lehtonen H (2005) Planktivory and diet-overlap of densely rakered whitefish
(Coregonus lavaretus (L.)) in a subarctic lake. Ecol Freshw Fish 14:50–58
Kahilainen KK, Malinen T, Tuomaala A, Alaja
¨
rvi E, Tolonen A, Lehtonen H (2007) Empirical evaluation of
phenotype-environment correlation and trait utility with allopatric and sympatric whitefish (Coregonus
lavaretus (L.)) populations in subarctic lakes. Biol J Linn Soc 92:561–572
Kahilainen K, Malinen T, Lehtonen H (2009) Polar light regime and piscivory govern diel vertical
migrations of planktivorous fish and zooplankton in a subarctic lake. Ecol Freshw Fish 18:481–490
Kalff J (2002) Limnology. Prentice-Hall, New Jersey
Klemetsen A, Elliott JM, Knudsen R, Sørensen P (2002) Evidence for genetic differences in the offspring of
two sympatric morphs of Arctic charr. J Fish Biol 60:933–950
Knudsen R, Klemetsen A, Amundsen P-A, Hermansen B (2006) Incipient speciation through niche
expansion: an example from the Arctic charr in a subarctic lake. Proc R Soc B 273:2291–2298
Knudsen R, Amundsen P-A, Primicerio R, Klemetsen A, Sørensen P (2007) Contrasting niche-based var-
iation in trophic morphology within Arctic charr populations. Evol Ecol Res 9:1005–1021
Lampert W, Sommer U (2007) Limnoecology, 2nd edn. Oxford University Press, Oxford
Langeland A, Nøst T (1995) Gill raker structure and selective predation on zooplankton by particulate
feeding fish. J Fish Biol 47:719–732
Lehtonen H, Niemela
¨
E (1998) Growth and population structure of whitefish (Coregonus lavaretus (L.)) in
mountain lakes of northern Finland. Arch Hydrobiol Spec Iss Adv Limnol 49:81–95
Lindsey CC (1981) Stock’s are chameleons: plasticity in gill rakers of Coregonid fishes. Can J Fish Aquat
Sci 38:1497–1506
Losos JB, Ricklefs RE (2009) Adaptation and diversification on islands. Nature 457:830–836
Nilsson N-A, Pejler B (1973) On the relation between fish fauna and zooplankton composition in north
Swedish lakes. Rep Inst Freshw Res Drott 53:51–77
Nogueira MR, Peracchi AL, Monteiro LR (2009) Morphological correlates of bite force and diet in the skull
and mandible of phyllostomid bats. Func Ecol 23:715–723
Østbye K, Bernatchez L, Næsje TF, Himberg M, Hindar K (2005) The evolutionary history of European
whitefish (Coregonus lavaretus L.) as inferred from mtDNA phylogeography and gillraker numbers.
Mol Ecol 14:4371–4387
Østbye K, Amundsen P-A, Bernatchez L, Klemetsen A, Knudsen R, Kristoffersen R, Næsje TF, Hindar K
(2006) Parallel evolution of ecomorphological traits in the European whitefish Coregonus lavaretus
(L.) species complex during postglacial times. Mol Ecol 15:3983–4001
Palkovacs EP, Post DM (2008) Eco-evolutionary interactions between predators and prey: can predator-
induced changes to prey communities feed back to shape predator foraging traits. Evol Ecol Res 10:
699–720
Palkovacs EP, Post DM (2009) Experimental evidence that phenotypic divergence in predators drives
community divergence in prey. Ecology 90:300–305
Pfennig DW, Rice AM, Martin RA (2006) Ecological opportunity and phenotypic plasticity interact to
promote character displacement and species coexistence. Ecology 87:769–779
Post DM, Palkovacs EP, Schielke EG, Dodson SI (2008) Intraspecific variation in a predator affects
community structure and cascading trophic interactions. Ecology 89:2019–2032
Price T (2008) Speciation in birds. Roberts and Company Publishers, Colorado
Robinson BW, Parsons KJ (2002) Changing times, spaces, and faces: tests and implications of adaptive
morphological plasticity in the fishes of northern postglacial lakes. Can J Fish Aquat Sci 59:1819–1833
Rogers SM, Bernatchez L (2007) The genetic architecture of ecological speciation and the association with
signatures of selection in natural lake whitefish (Coregonus sp. Salmonidae) species pairs. Mol Biol
Evol 24:1423–1438
Rueffler C, Van Dooren TJM, Leimar O, Abrams PA (2006) Disruptive selection and then what? Trends
Ecol Evol 21:238–245
Salzburger W (2009) The interaction of sexually and naturally selected traits in the adaptive radiations of
cichlid fishes. Mol Ecol 18:169–185
Evol Ecol (2011) 25:573–588 587
123
Sanderson SL, Cheer AY, Goodrich JS, Graziano JD, Callan WT (2001) Crossflow filtration in suspension-
feeding fishes. Nature 412:439–441
Sandlund OT, Næsje TF, Kjellberg G (1987) The size selection of Bosmina longispina and Daphnia galeata
by co-occuring cisco (Coregonus albula), whitefish (C. lavaretus) and smelt (Osmerus eperlanus).
Arch Hydrobiol 110:357–363
Schluter D (1996) Ecological speciation in postglacial fishes. Phil Trans R Soc Lond B 351:807–814
Schluter D (2000) The ecology of adaptive radiation. Oxford University Press, Oxford
Schluter D, McPhail JD (1992) Ecological character displacement and speciation in sticklebacks. Am Nat
140:85–108
Schluter D, McPhail JD (1993) Character displacement and replicate adaptive radiation. Trends Ecol Evol
8:197–200
Seghers BH (1975) Role of gill rakers in size-selective predation by lake whitefish, Coregonus clupeaformis
(Mitchill). Verh Int Verein Limnol 19:2401–2405
Sku
´
lason S, Snorrason S, Jo
´
nsson B (1999) Sympatric morphs, populations and speciation in freshwater fish
with emphasis on arctic charr. In: Magurran AE, May RM (eds) Evolution of biological diversity.
Oxford University Press, Oxford, pp 70–92
Smith JC, Sanderson SL (2008) Intra-oral flow patterns and speeds in a suspension-feeding fish with gill
rakers removed versus intact. Biol Bull 215:309–318
Snowberg LK, Bolnick DI (2008) Assortative mating by diet in a phenotypically unimodal but ecologically
variable population of stickleback. Am Nat 172:733–739
Sva
¨
rdson G (1952) The coregonid problem. IV. The significance of scales and gillrakers. Rep Inst Freshw
Res Drott 33:141–166
Sva
¨
rdson G (1979) Speciation of Scandinavian Coregonus. Rep Inst Freshw Res Drott 57:1–95
Tolonen A (1998) Application of a bioenergetics model for analysis of growth and food consumption of
subarctic whitefish Coregonus lavaretus (L.) in Lake Kilpisja
¨
rvi, Finnish Lapland. Hydrobiologia
390:153–169
Webb PW (1984) Form and function in fish swimming. Sci Am 251:58–68
Werth AJ (2004) Models of hydrodynamic flow in the bowhead whale filtering apparatus. J Exp Biol
207:3569–3580
Zaret TM (1980) Predation and freshwater communities. Yale University Press, London
588 Evol Ecol (2011) 25:573–588
123