African cichlid fish: A model system in adaptive radiation research

Abstract

The African cichlid fish radiations are the most diverse extant animal radiations and provide a unique system to test predictions of speciation and adaptive radiation theory. The past few years have seen major advances in the phylogenetics, evolutionary biogeography and ecology of cichlid fish. Most of this work has concentrated on the most diverse radiations. Unfortunately, a large number of small radiations and 'non-radiations' have been overlooked, potentially limiting the contribution of the cichlid system to our understanding of speciation and adaptive radiation. I have reviewed the literature to identify 33 intralacustrine radiations and 76 failed radiations. For as many as possible I collected information on lake size, age and phylogenetic relationships. I use these data to address two questions: (i) whether the rate of speciation and the resulting species richness are related to temporal and spatial variation in ecological opportunity and (ii) whether the likelihood of undergoing adaptive radiation is similar for different African cichlid lineages. The former is a key prediction of the ecological theory of adaptive radiation that has been presumed true but remains untested for cichlid radiations. The second is based on the hypothesis that the propensity of cichlids to radiate is due to a key evolutionary innovation shared by all African cichlids. The evidence suggests that speciation rate declines through time as niches get filled up during adaptive radiation: young radiations and early stages of old radiations are characterized by high rates of speciation, whereas at least 0.5 Myr into a radiation speciation becomes a lot less frequent. The number of species in cichlid radiations increases with lake size, supporting the prediction that species diversity increases with habitat heterogeneity, but also with opportunity for isolation by distance. Finally, the data suggest that the propensity to radiate within lakes is a derived property that evolved during the evolutionary history of some African cichlids, and the appearance of which does not coincide with the appearance of proposed key innovations in morphology and life history.

Full text

Review
African cichlid fish: a model system
in adaptive radiation research
Ole Seehausen
1,2,
*
1
Department of Aquatic Ecology and Evolution, Institute of Zoology, University of Bern,
Baltzerstrasse 6, 3012 Bern, Switzerland
2
EAWAG Ecology Research Centre, Seestrasse 79, 6047 Kastanienbaum, Switzerland
The African cichlid fish radiations are the most diverse extant animal radiations and provide a unique
system to test predictions of speciation and adaptive radiation theory. The past few years have seen major
advances in the phylogenetics, evolutionary biogeography and ecology of cichlid fish. Most of this work
has concentrated on the most diverse radiations. Unfortunately, a large number of small radiations and
‘non-radiations’ have been overlooked, potentially limiting the contribution of the cichlid system to our
understanding of speciation and adaptive radiation. I have reviewed the literature to identify 33
intralacustrine radiations and 76 failed radiations. For as many as possible I collected information on lake
size, age and phylogenetic relationships. I use these data to address two questions: (i) whether the rate of
speciation and the resulting species richness are related to temporal and spatial variation in ecological
opportunity and (ii) whether the likelihood of undergoing adaptive radiation is similar for different
African cichlid lineages. The former is a key prediction of the ecological theory of adaptive radiation that
has been presumed true but remains untested for cichlid radiations. The second is based on the
hypothesis that the propensity of cichlids to radiate is due to a key evolutionary innovation shared by all
African cichlids. The evidence suggests that speciation rate declines through time as niches get filled up
during adaptive radiation: young radiations and early stages of old radiations are characterized by high
rates of speciation, whereas at least 0.5 Myr into a radiation speciation becomes a lot less frequent. The
number of species in cichlid radiations increases with lake size, supporting the prediction that species
diversity increases with habitat heterogeneity, but also with opportunity for isolation by distance. Finally,
the data suggest that the propensity to radiate within lakes is a derived property that evolved during the
evolutionary history of some African cichlids, and the appearance of which does not coincide with the
appearance of proposed key innovations in morphology and life history.
Keywords: adaptive radiation; cichlids; ecological speciation; sexual selection; key innovations;
species–area relationship
1. INTRODUCTION
One hundred and fifty years after the publication of
Darwin’s (1859) book ‘The origin of species’, and after
several decades of intensified empirical speciation research
(Schluter 2000; Turelli et al. 2001; Coyne & Orr 2004) the
origin of species remains poorly understood. The
importance of divergent selection is now firmly established
(Coyne & Orr 2004; Gavrilets 2004), and recent work
indicates that speciation driven by divergent selection can
be remarkably rapid (Ungerer et al. 1998; Hendry et al.
2000; Higgie et al. 2000; Feder et al. 2003). Theory
predicts that such ecological speciation is most likely when
populations invade new adaptive zones with underutilized
niches and may occur in bursts that characterize adaptive
radiation (Simpson 1953; Schluter 2000; Dieckmann et al.
2004; Gavrilets & Vose 2005). Numerous empirical
studies suggest such conditions may indeed be a common
feature of adaptive radiation (Schluter 2000).
Two problems associated with the empirical evidence
for rapid ecological speciation as an important source of
biological diversity are: (i) quantitative comparative tests
of the main theoretical predictions are rare and (ii) most
comparative studies ignore the cases where radiation did
not take place (but see Bernatchez 2003; Vamosi 2003).
Clearly, there are numerous situations where populations
failed to radiate despite inhabiting environments appar-
ently conducive to adaptive radiation. Understanding, and
being able to predict, when adaptive radiation does and
does not occur will perhaps provide the strongest test of
theories of ecological speciation and adaptive radiation
during the next decade.
The most serious constraint to empirically testing
theory is the lack of appropriate model systems. Next to
experimental microbial evolution (Rainey et al. 2000), the
cichlid fish in African lakes are emerging as one of the
potentially most powerful model systems in speciation and
adaptive radiation research (Kocher 2004). Cichlid fish
have radiated into endemic species assemblages in more
Proc. R. Soc. B (2006) 273, 1987–1998
doi:10.1098/rspb.2006.3539
Published online 9 May 2006
The electronic supplementary material is available at http://dx.doi.
org/10.1098/rspb.2006.3539 or via http://www.journals.royalsoc.ac.
uk.
*ole.seehausen@aqua.unibe.ch
1987 q 2006 The Royal Society

than 30 African lakes (table 1 of electronic supplementary
material). Between 1000 and 2000 speciation events
occurred in the past 5 Myr alone. The large number of
independent replicate radiations, their phenotypic diver-
sity, their wide range of ages (10 000 to over 10 Myr;
figure 1) and the presence of many more instances where
cichlids failed to radiate (table 2 of electronic supplemen-
tary material), make this a uniquely powerful model
system for empirically testing speciation and adaptive
radiation theory.
Inspired by the classic monographs of Fryer & Iles
(1972), and Greenwood (1974), effort toward character-
izing the nature of cichlid radiations has progressed on
multiple fronts, as evidenced by several recent reviews
(Stiassny & Meyer 1999; Kornfield & Smith 2000;
Seehausen 2000; Danley & Kocher 2001; Markert et al.
2001; Turner et al. 2001; Kocher 2004; Salzburger &
Meyer 2004; Van Alphen et al. 2004; Genner & Turner
2005). Simultaneously, speciation research has undergone
a transition from verbal models and descriptive empirical
work to a quantitative predictive science with strong
foundations in ecology (Schluter 2000; Dieckmann et al.
2004) and mathematical biology (Gavrilets 2004).
Given these advances, an attempt to integrate cichlid
research into the emerging quantitative framework of
speciation and adaptive radiation research is timely. To
this end, I here explore the environmental and phyloge-
netic correlates of variation in cichlid speciation propen-
sity. I collected published information on 33 lacustrine
radiations with minimum species numbers between 2 and
451 species, and on 76 colonization events without
intralacustrine speciation. I use these data to ask why
cichlid speciation is sometimes stunningly fast but most of
the time not. Even though this is a review, I had to resort to
some meta-analysis in the §§3a,b and 4.
2. THEORETICAL PREDICTIONS
Two assumptions are pervasive in the cichlid literature.
The first is that variation in cichlid speciation rates is
explained by ecological opportunity. The second is that a
difference between cichlids and other fishes in their
propensity to radiate is explained by an evolutionary key
innovation (Nitecki 2000). Ecological opportunity is
thought important since most radiations occur following
the colonization of lakes. However, that speciation is rapid
upon invasion of a new environment does not prove that
radiation is driven by ecological opportunity. In organisms
with limited dispersal capabilities (like cichlids), coloniza-
tion of large lakes leads to increased genetic population
subdivision and may allow for ecologically neutral
population divergence and speciation by drift and sexual
selection (Dominey 1984). Indeed several authors have
suggested that drift and ecologically neutral divergent
sexual selection between isolated populations may drive
rapid speciation in cichlids (Ribbink et al. 1983; McKaye
1991; Knight & Turner 2004). Adaptive radiation theory
(Schluter 2000) and models (Gavrilets & Vose 2005),
as well as the extended logic of ecological speciation
models (Rosenzweig 1978; Dieckmann & Doebeli 1999;
Kondrashov & Kondrashov 1999), however, make the
prediction that after an initial burst speciation rates slow
down as niche space fills up. In contrast, speciation driven
by niche-independent mechanisms such as non-ecological
sexual selection, does not predict such a temporal trend of
declining speciation rates in the course of a radiation
(although increased extinction rates can generate a
superficially similar pattern). Further, consideration of
evolutionary and ecological species–area relationships
predicts that larger lakes generate more species and
allow coexistence of a larger number of species (Losos &
Schluter 2000). In the first part of this review, I use
published data from 33 African lacustrine cichlid radiations
to test these predictions. I arrange the evidence by the nature
of the data: distribution of species richness, phylogenetics,
population genetics, contemporary evolution.
In the second part, I will compare 33 successful
radiations with 76 cases of non-radiation and ask whether
they differ in external conditions or phylogenetic history.
The evolutionary key innovation hypothesis posits that the
invasion of new environments, required for ecology-driven
speciation, may be aided by the origin of evolutionary
innovations that permit utilization of previously inaccess-
ible niches (Simpson 1953). Liem (1973) and Galis &
Drucker (1996) suggested that the functional decoupling
of the upper and the lower pharyngeal jaws in cichlid fish is
0
2
4
6
8
10
12
14
16
0.25 0.58 1 1.5 2.2 3 4 5.3 6.9 9 11.6
Hawaiian
honeycreepers
Darwin’s
finches
postglacial
salmonids,
sticklebacks
Carribean
anoline lizards
number of radiations
a
pp
roximate a
g
e (million
y
ears)
Figure 1. Frequency distribution of the age of African lacustrine cichlid fish radiations contrasted with approximate ages of other
well-studied vertebrate adaptive radiations.
1988 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)

such a key innovation that permitted survival on novel
food types through increased behavioural flexibility, and
subsequent genetic adaptation to such new resources
(West-Eberhard 1989; Galis & Metz 1998). Because for,
as far as known, all African cichlids share the decoupled
pharyngeal jaw anatomy, if the hypothesis is correct, the
high propensity to radiate should be a property of all
African cichlid lineages. Salzburger et al. (2005) proposed
that the parental care behaviour and sexually dimorphic
anal fin in one group of cichlids are two key innovations
that together allowed for faster divergence by sexual
selection. If this was correct, the propensity to radiate
should increase concomittently with the origin of female
mouthbrooding and male egg spots.
3. THE DYNAMICS OF SPECIATION RATES DURING
ADAPTIVE RADIATION
(a) Number games: larger lakes generate more
species and speciation is fastest when lakes are
young
Most cichlid species of the African lakes are endemic to
single lakes. Mitochondrial DNA (mt DNA) genealogies
suggest mono- or paraphyly for most lake faunas (e.g. the
Lake Malawi and Lake Victoria radiations; Meyer 1993;
Nagl et al. 2000; Seehausen et al. 2003; Verheyen et al.
2003) or for large clades within a lake fauna (Lake
Tanganyika; Salzburger et al. 2002b). It is reasonable in
such cases to use the lake’s geological age as the maximum
possible age for its endemic cichlid radiations, unless
molecular clock estimates suggest more recent beginnings
for a radiation. Speciation intervals or time for speciation
(TFS; McCune 1997) have been calculated that way for
all of the larger African cichlid radiations and for several
other animal radiations using the number of known extant
species (McCune 1997; Seehausen 1999, 2000, 2002,
2005; Turner 1999; Coyne & Orr 2004). One robust result
has emerged from these analyses: the shortest average
speciation intervals are observed in the youngest radi-
ations (Seehausen 2002).
Here I generalize this result to all known cichlid
radiations: I found information on the maximum age
(‘maximum time in lake’, defined as age of lake, or
molecular clock estimate of the age of a radiation if less
than lake age) for 24 of the 33 radiations. The data are in
table 1 of electronic supplementary material. Time for
speciation is highly positively correlated with radiation age
( yZ0.32xK12 434, R
2
Z0.90, p!0.0001; figure 2a).
I found information on lake surface area for 30 radiations
(table 1 of electronic supplementary material). It emerges
as a highly significant predictor of species richness across
–100
0
100
02468
100
1000
10000
100000
1000000
10000000
1000 10000 100000 1000000 10000000
1
10
100
1000
0.1 1 10 100 1000 10000 100000
0.1 1 10 100000
(a)(b)
(d)(c)
time for speciation (log model)
100
1000
10000
100 000
1000000
10000000
time for speciation (log model)
radiation age (years)
lake surface area (km
2
)
100001000100
number of species
lake surface area (km
2
)
radiation a
g
e (million
y
ears)
residual number of species
400
300
200
–200
7531
Figure 2. (a) The relationship between time for speciation (TFS) and radiation age calculated using 20 African cichlid radiations
for which TFS was available (table 1 of electronic supplementary material). (b) The evolutionary species–area relationship for
African cichlid radiations, calculated using 29 African cichlid radiations for which lake surface area and number of species were
available (table 1 of electronic supplementary material). (c) The relationship between TFS and lake size calculated using 20
African cichlid radiations for which TFS and lake surface area were available (table 1 of electronic supplementary material). (d )
The relationship between residual number of species (after the effect of lake size is taken into account) and radiation age,
calculated using 20 African cichlid radiations for which both time in lake and lake surface area were available.
African cichlid fish O. Seehausen 1989
Proc. R. Soc. B (2006)

these radiations ( yZ4.45e
6
!10
K5x
, R
2
Z0.47, p!0.0001;
figure 2b), whereas it is unrelated to TFS (figure 2c). The
shape of the evolutionary species–area relationship is
similar to that observed for the island radiations of anoline
lizards (Losos & Schluter 2000)withanimportant
difference: in situ speciation of cichlids occurs even in
the smallest lakes. The number of species is unrelated to
lake size among small and medium sized lakes with
between 2 and 11 species, whereas it steeply increases with
lake size above 1000 km
2
surface area. Once the effects of
lake size are taken out, residual species richness is
significantly negatively related to radiation age: younger
radiations contain more species than predicted by
lake size ( yZK2!10
K5x
C29.94, R
2
Z0.29, pZ0.009;
figure 2d ). This is consistent with the prediction from
theory that speciation is fastest early in radiations. It is also
consistent with recent empirical (Gillespie 2004) and
simulation (Gavrilets & Vose 2005) studies that have
documented an ‘overshooting effect’ during adaptive
radiation. However, the test cannot discriminate between
diminishing speciation and increasing extinction rates. It
assumes that the number of extant species is a valid
approximation of the number of species that were
generated in the course of a radiation. If extinction rates
were high, many more species may have been generated,
and speciation could be a lot faster. The magnitude of the
underestimation of speciation rates would be positively
correlated with radiation age, generating a pattern of
apparently diminishing speciation rates with increasing
radiation age.
Hence, the observed pattern is consistent with dimin-
ishing rates of gross speciation as predicted if speciation
rates are ecological opportunity-dependent, but also with
unchanged speciation rates against a background of
increasing extinction rates. Either way, the relationships
do imply that both ecological opportunity (net speciation
rates decline as an adaptive zone fills up with species) and
area size are two important determinants of species
richness in the African cichlid fish radiations. The causes
of the area effect are likely to be twofold: (i) a positive
correlation of area with environmental heterogeneity and
hence diversity of niches, (ii) increased opportunity for
isolation by distance. To partition the variance contributed
by the two components, data on the number of sympatric
and parapatric species (alpha diversity) within each lake are
required. Such data are currently not available for many
lakes. The limited data that are available (e.g. Genner et al.
2004), suggest (i) that alpha diversity in large lakes is
considerably lower than total species richness (approx.
30% of the latter), implying a large contribution of isolation
by distance and (ii) that the shape of the alpha diversity–
lake size relationship is similar to that of the species–lake
size relationship, implying increased niche diversity contri-
butes to increased species richness in large lakes too.
(b) Tree shape analyses: early speciation bursts
and diminishing rates as radiations mature
Changes in net speciation rates through time can be
estimated from the lengths of internal branches that
separate successive speciation events in a phylogenetic
tree. Lineage-through-time plots (Barraclough & Nee
2001) are a way of visualizing such trends. Molecular
phylogenies for several African cichlid fish radiations are
available for such analysis. If cichlid speciation is driven
and constrained by the availability of vacant niches, we
should expect to see the rate of lineage multiplication
peak early during radiations and decline as adaptive
radiation proceeds.
(i) Young radiations
Absence of mtDNA haplotype lineage sorting, consistent
with multiple speciation events in very short succession,
characterize every cichlid radiation for which geological or
molecular clock age estimates implicated origins within
the past 300 000 years: Lake Victoria (Nagl et al. 1998,
2000; Verheyen et al. 2003), Palaeolake Makgadikgadi
( Joyce et al. 2005), Lake Natron (Seegers et al. 1999) and
Lake Ejagham (Schliewen et al. 2001). Since speciation is
too recent for lineage sorting to have been completed in
these radiations, testing for trends in speciation rates
within this early phase of radiation is beyond the temporal
resolution of mitochondrial gene trees. Species trees can
then only be built using large numbers of genomic loci.
Clocklike trees built from microsatellite frequency data or
Amplified Fragment Length Polymorphisms (AFLPs) are
not currently available. However, ultrametric trees for the
Lake Victoria radiation (Seehausen et al. 2003) confirm
that the radiation began as a starburst with very short or no
branches separating any two speciation events. However,
it is impossible to confirm with existing data whether
speciation rates have declined during the course of the
radiation.
(ii) Radiations of middle-age
Lake Malawi
Trees based on mtDNA (Meyer 1993; Moran & Kornfield
1993; Kocher et al. 1995) suggest the Lake Malawi flock is
structured into six well-defined lineages that emerged as
an unresolved starburst early in the radiation. The
divergence time for these lineages is not well estimated,
the lake is 2–5 Myr (ago) old but was largely dry between
1.6 and 1 Myr (ago; Delvaux 1995). It is likely that the six
divergent lineages survived this drought. Four of them
have radiated further into between 10 and 250 species
each, which is thought to have occurred upon lake refilling
(Sturmbauer et al. 2001). Whereas no haplotype sharing
has been observed between any two species from these
different lineages (Kocher et al. 1995; Parker & Kornfield
1997; Shaw et al. 2000; Turner et al . 2004), all of the
secondary radiations are characterized by lack of haplo-
type sorting just as the radiations in young lakes (Moran &
Kornfield 1993; Parker & Kornfield 1997; Turner et al.
2004). Hence, an early radiation burst in Lake Malawi
appears to have been followed by a period of stasis and
most likely lineage extinction, until renewed radiation
bursts in four surviving lineages gave rise to the modern
species diversity. Like in the young lakes, these speciation
bursts are too recent to detect possible levelling off of
speciation rates with lineage-through-time plots using
mtDNA gene trees, and ultrametric species-level phylo-
genies from multilocus data are either not yet available,
or cover only a small subclade (Albertson et al. 1999;
Allender et al. 2003).
Lake Barombi Mbo
The lineage-through-time plot for the 0.5–1 Myr old
radiation in Lake Barombi Mbo, generated using an
ultrametric tree based on genomic AFLP loci (Schliewen &
1990 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)

Klee 2004), suggests an early and a late speciation burst,
with a period of stasis following each (figure 3). Trans-
specific mtDNA haplotype-sharing is restricted to a single
species pair in the second radiation burst, providing
additional evidence that speciation rates have recently
been low (Schliewen & Klee 2004).
(iii) Old radiations
Where time-intervals between successive speciation events
were shorter than required for lineage sorting, the
signature of ancient rapid speciation will be preserved in
form of polytomies in, and conflict between gene trees. All
the major clades in the Lake Tanganyika tree emerged
from two such polytomies inferred from mitochondrial
genealogies (Salzburger et al. 2002) and incongruent short
interspersed element (SINE) insertion patterns (Ta ka ha s h i
et al. 2001a,b;nodes2and3infigure 4). The simultaneous
origin of all Lake Tanganyika lineages in two major
speciation bursts is inconsistent with the hypothesis that
the origin of these lineages predates the origin of the lake
and they independently colonized the lake from rivers
(Nishida 1991; Salzburger et al. 2002), but implies instead
that the oldest lineages are survivors of an early radiation
burst within Lake Tanganyika or a precursor lake basin.
Similar to the situation in Lake Malawi, each of these
lineages has subsequently radiated into between 6 and 73
extant species, but these radiations are between 2 and
5 Myr old.
I used published linearized trees on five of these
radiations (Koblmueller et al. 2004, 2005; Duftner et al.
2005) for plotting lineages-through-time (figure 3a). All
five show evidence of speciation bursts around the same
time followed by a period of stasis to the present, although
the Cyprichromini show signs of a recent renewed
increase in speciation rate (figure 3a). The ancient
Bathybatini lineage shows evidence for a much earlier
burst as well, and there may have been an earlier burst in
the Ectodini. A sixth lineage, the Tropheini, also show
evidence of an early starburst followed by declining
speciation rates (Sturmbauer et al. 2003), even though
no ultrametric tree is published that would have allowed a
plot. Incomplete lineage sorting seems uncommon among
Lake Tanganyika cichlids and when observed among
closely related species has been interpreted, based on
distribution data, as evidence of hybridization upon
secondary contact (Ruber et al. 2001; Salzburger et al.
2002). Hence, all evidence suggests that speciation rates
have slowed down from the past to the present in Lake
Tanganyika.
Short interspersed element analysis allowed the dis-
covery of the oldest burst of speciation in the African
cichlid tree as a third area of incongruent SINE insertion
patterns (node 1 in figure 4a,b). In this ancient radiation,
the precursors of the Lake Tanganyika cichlids, several
endemic Congolese lineages, and several now pan-African
lineages appear to have emerged simultaneously or in very
close succession about 14 Myr (ago), followed by a period
of lower speciation (or high extinction) rates (Terai et al.
2003). Ultrametric trees that would allow lineage-
through-time plots for this part of the African cichlid
phylogeny are not currently available. The environmental
settings in which this first of the African cichlid radiations
occurred have not been investigated. It is tempting to
speculate that the radiation happened in an ancient lake
that no longer exists (cf. Joyce et al. 2005). Time and
geographical location make Palaeolake Congo, a large
inland sea that existed in the central Congo basin in the
Neogene (Cahen 1954), a candidate lake.
(c) Divergent population genetics suggest
sustained high speciation rates in sections of
the Lake Malawi radiation
Divergent population genetics (Machado et al. 2002) hold
much promise for studying cichlid speciation. Recent
work on some Lake Malawi cichlids suggests that high
speciation rates have been maintained in at least one of the
four secondary radiations even 0.5–1 Myr after the onset
of radiation. Wo n et al. (2005) studied sequence
divergence among closely related populations and species
of the ‘Mbuna’ genus Tropheops using variation in the
number of short-tandem repeats (microsatellites) and
linked flanking region sequences (HapSTRs). Estimating
divergence times with parameter-rich maximum-
likelihood models, they arrived at times since speciation
of between 1000 and 17 000 years for the three species
studied. Taking a mean speciation interval of 10 000 years
as representative for the rock-dwelling Mbuna, the
number of 230 extant species in the clade implies that
one new Mbuna species arises every 46 years in Lake
Malawi (Wo n et al. 2005), a TFS similar to that of the
much younger Lake Victoria radiation. These data suggest
the speciation rate has remained at ‘young radiation levels’
0.5–1 Myr into the Mbuna radiation.
If confirmed, this would implicate speciation mechan-
isms less closely tied to ecological opportunity than in the
tilapiine cichlids of Lake Barombi Mbo and the precursors
of the modern haplochromines in the Lake Tanganyika
radiation. Mbuna, like Lake Victoria haplochromines, are
extremely sexually dimorphic in coloration and very
diverse in male colour patterns. It is conceivable that the
interaction of drift and sexual selection in subdivided
populations is responsible for much of this sustained rapid
speciation (Dominey 1984). It is also possible that an
interaction between sexual and ecological selection allows
a faster and more fine scale response of mating systems to
divergent ecological selection (Van Doorn & Weissing
2001; Gavrilets 2004). Won et al.’s result requires a high
10
20
30
–15 –10 –5
0
Bathybatini
Cyprichromini
Limnochromini
Ectodini
Perissodini
(a)(b)
10
20
–2
0
number of species
se
q
uence diver
g
ence (m
y
)
Figure 3. Species number through time plots for African
cichlid radiations. (a) Several lineages of Lake Tanganyika
cichlids (Limnochromini, Perissodini, Cyprichromini from
Duftner et al. 2005; Bathybatini from Koblmueller et al. 2004,
Ectodini from Koblmueller et al. 2005). (b) The radiation in
Lake Barombi Mbo (Schliewen & Klee 2004).
African cichlid fish O. Seehausen 1991
Proc. R. Soc. B (2006)

rate of species turnover in Mbuna. If the radiation of
Mbuna began 1 000 000 years ago with the estimated
TFS, and has proceeded with steady state rates of
speciation and extinction, there may have been 23 000
different species over the years, which is two orders of
magnitude more species than we observe in the lake today!
How their TFS estimate for Tropheops (the most species
rich genus of Mbuna) compares to other genera remains to
be seen. Many Lake Malawi Mbuna are local endemics
with narrow distribution ranges (Ribbink et al. 1983;
Genner et al. 2004), which is consistent with high-
stochastic extinction rates required by the extant richness
and TFS estimate.
(d) Adaptive diversification versus reproductive
isolation
The widespread combination of (i) a monophyletic
mitochondrial genealogy, (ii) starburst-like gene and
species trees and (iii) lack of mtDNA lineage sorting can
be explained by two scenarios, both of which require some
Figure 4. (Caption opposite.)
1992 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)

non-trivial assumption. In one scenario speciation is
concentrated in bursts early during a radiation but begins
only after the geographical and demographic expansion of
the founding population, in the wake of which the
sequence polymorphisms must have arisen that are now
still shared between species. This scenario implies a lag
between colonization of a lake or newly available habitat
(e.g. after a lake level rise) and the beginning of speciation.
In this scenario speciation can hence not be simply a by-
product of increasing opportunity for isolation by
distance. The alternative scenario, that speciation begins
during the population expansion (as suggested by
Sturmbauer et al. 2001), and before the current sequence
polymorphisms can have arisen, requires that interspecific
hybridization continues for a sufficiently long time after
speciation to allow the spread of newly arising mutants
across the species, consistent with some empirical
observations (Seehausen et al. 1997a; Ruber et al. 2001;
Salzburger et al. 2002; Smith et al. 2003).
These alternative scenarios have very different impli-
cations for the dynamics of the speciation process.
Whereas both result in starburst gene trees because all
but the divergently selected genes keep coalescing through
the entire population, speciation (in the sense of establish-
ment of reproductive isolation) would be rapid in the
former, but not in the latter. In the latter only incipient
speciation would be fast, but the completion of reproduc-
tive isolation would require more time. The hybridization
scenario would imply that functional radiation happened
in a syngameon situation, as inferred to have occurred in
radiations of Darwin’s finches (Freeland & Boag 1999),
crossbills (Parchman et al. in press) corals (Van Oppen
et al. 2002) and oak trees (Petit et al. 2002). This scenario
also implies that the divergent natural selection required
for adaptive radiation may itself be insufficient to cause
completion of reproductive isolation. The latter is
consistent with results of a simulation study of adaptive
radiation that found that genomes of species remained
‘porous’ throughout the radiation (Gavrilets & Vose 2005)
(e) Contemporary evolution: rapid change
is possible
At least two observations of contemporary evolution in
African cichlid fishes suggest rapid incipient speciation
may be facilitated by interspecific hybridization associated
with environmental perturbation. Lake Victoria experi-
enced major environmental stress beginning in the late
1970s. Loss of water clarity due to nutrient pollution
(Verschuren et al. 2002) lead to reduced female mate
selectivity (Seehausen et al. 1997a). Population growth of
an introduced top predator (Nile perch, Lates spp.)
simultaneously caused a crash of cichlid populations
(Witte et al. 1992). Strongly diminished abundances and
impaired vision are probably responsible for the break-
down of reproductive isolation between species. Simul-
taneously, the ecological resource base and habitat
structure underwent major changes too due to eutrophica-
tion. After most populations in sublittoral and pelagic
habitats had collapsed in the late 1980s (Witte et al. 1992),
some recovered rapidly in the early 1990s (Seehausen et al.
1997b; Witte et al. 2000). Among these were populations
composed of novel morphological and ecological pheno-
types that despite intensive sampling had not been
reported prior to the 1980s crash (Seehausen et al.
1997b). Some of these phenotypes appear to have been
relatively stable since, and may represent incipient species
which have arisen within 20 years. The possibility that
hybridization contributed to the emergence of these new
phenotypes is presently under investigation.
Following translocation from its natural range in
northern Lake Malawi to Tumbi West Island in the
south in the 1970s, a population of Cynotilapia afra
Figure 4. (Opposite.)(a) The ratios of number of successful radiations over number of ‘failed’ radiations mapped on a
mitochondrial gene tree (ND2) for the African cichlid fish (tree from Klett & Meyer 2002). ‘Node 1’ corresponds to a hard
polytomy deep in the African cichlid radiation, that has been interpreted as the signature of an adaptive radiation burst 14 Myr
(ago) (Terai et al. 2003; figure 4b). ‘Node 2’ is a second major hard polytomy corresponding with the origin of the four deep
lineages of Lake Tanganyika cichlids (Takahashi et al. 2001b). ‘Node 3’ is a third major hard polytomy corresponding to the
radiation burst of the modern Lake Tanganyika lineage (Takahashi et al. 2001b; figure 4c). Note that several weakly supported
nodes in this gene tree collapse to the same polytomy in multilocus trees (figure 4b,c). The different shading indicates sections of
the tree that were compared for the ratio of successful to failed radiations. The light grey branch is the ‘East African’ lineage that
emerged in the Tanganyika primary radiation, then radiated again in Lake Tanganyika, and much later seeded Lakes Malawi,
Victoria, Makgadikgadi and many others (also figure 4c). Genera that were not sampled for this gene tree, but whose
phylogenetic position is known from other mtDNA sequence data, are indicated in the ratio column. Note that the exact position
of these does not affect the analysis because all of them are unambiguously assigned to one of the shaded branches. The origins of
two proposed key evolutionary innovations, decoupled pharyngeal jaws (at the base of the tree) and egg dummies, are indicated.
The third one, female mouthbrooding is not indicated because it evolved multiple times and is scattered throughout the tree.
Abbreviations stand for lakes: NaZNatron; MyZManyara; JiZJipe; ChaZChala; TZTanganyika; VZVictoria; RZRukwa;
PMZPaleo–Makgadikgadi; MZMalawi; BZBangweulu; MwZMweru; KZKivu; EZEdward; AZAlbert; CZChad; KiZ
Kinneret; TuZTurkana; TaZTana; SZStephanie; BaZBarombi Mbo; EjZEjagham; UZUpemba; FZFwa; SaZSaka; NsZ
Nshere; LuZLutoto; CiZChilwa; NbZNabugabo; TmZTumba; GZGuinas sink hole; BeZBemin; KoZBarombi ba Kotto;
NdZMayi Ndombe. (b) A phylogenetic tree based on short interspersed element (SINE) insertion data. Each boxed number is
a SINE locus, and arrowheads indicate their origins. The MVhL lineage is resolved in figure 4c. The grey circle at node 1
indicates the period when retention of ancestral polymorphisms (presence or absence of a SINE) was assumed to have occurred
at loci 223, 260, 304, 316, 1223 and 1544. Reproduced with permission from Terai et al. 2003.(c) A phylogenetic tree for cichlid
species in the 12 tribes of Lake Tanganyika, based on SINE insertion data. Arrowheads indicate internodes deduced from
insertion of a SINE unit at each of 24 loci analysed. The three clades were supported by the patterns of insertion of a SINE unit
at loci 213, 214, 245, 247, 254, 314, 455, 1569, 1666 and 1715 (the MVhL clade); at loci 328, 330, 343, 1221, 1238, 1262,
1269, 1277 and 1654 (the MVHT clade); and at loci 1233, 1265, 1281, 1291 and 1528 (the MVH clade). The grey portion of
the tree indicates the period during which extensive putative incomplete lineage sorting of ancestral polymorphisms occurred
(reproduced with permission from Takahashi et al. 2001b).
African cichlid fish O. Seehausen 1993
Proc. R. Soc. B (2006)

hybridized with resident Maylandia zebra, a Mbuna cichlid
with different tooth shape but similar male coloration
(Streelman et al. 2004). Hybridization was first docu-
mented in the early 1990s (Stauffer et al. 1996); in the
following years the hybrid population appears to have
expanded its range around the island and within 30 years
after the introduction populations from the north and
south sides of the island became phenotypically differ-
entiated (Streelman et al. 2004). However, it remains to be
shown that C. afra were not introduced from more than
one source population (K. Young 2005, personal communi-
cation). Both observations indicate that rapid phenotypic
divergence, reminiscent of incipient speciation, is possible
following interspecific hybridization.
4. PHYLOGENETIC HISTORY DETERMINES THE
PROBABILITY OF ADAPTIVE RADIATION
It is rarely appreciated that the cases in which cichlids
radiated are only a minority among all cases in which
cichlids successfully colonized lakes in Africa. I use the
data on 33 successful and 76 ‘failed’ radiations to ask
whether anything is different between the circumstances in
which diversification did or did not take place. The answer
is there is no difference in either time in lake (tZ0.56,
d.f.Z50; pZ0.58) or lake size (tZK0.34, d.f.Z80;
pZ0.74). This may be surprising given that the same
variables are strong predictors of speciation rate and
species numbers in adaptive radiations. Yet, they do not
predict whether or not a colonizing population undergoes
adaptive radiation.
It is difficult to rigorously test for phylogenetic inertia in
cichlid speciation rates because the number of species
generated in individual cichlid radiations is often
unknown, there are large researcher biases in estimates
(Genner et al. 2004), all larger phylogenies are highly
incomplete, and the level of completeness is generally
unbalanced between branches. The phylogenetic distri-
bution of presence and absence of any intralacustrine
speciation may under such circumstances be the most
robust measure of intrinsic speciation propensity. To test
for phylogenetic inertia I mapped lake colonizations with
and without intralacustrine speciation onto the only
published gene tree of African cichlids that sampled
species broadly and without bias towards the large
radiations (Klett & Meyer 2002; figure 4a). Although
other gene trees may deviate in detail from this one, the
overall topology is robust between published African
cichlid genealogies.
The occurrence of intralacustrine speciation is not
randomly distributed on the tree. All 33 African
intralacustrine radiations have occurred within the mono-
phyletic tilapiine–haplochromine superlineage (figure 4a).
Although other lineages have Pan-African distributions
and colonized many lakes (most prominently the genera
Hemichromis and Tylochromis), they do not appear to have
undergone intralacustrine speciation anywhere. Further,
Tilapia, Oreochromis, Sarotherodon and the ancient Lake
Tanganyika lineages have all given rise to only small
intralacustrine radiations, whereas all the large radiations
([10 species) are in the ‘East African’ lineage.
I used likelihood ratio tests to ask whether the ratios of
successful to ‘failed’ radiations differed between branches
of the African cichlid tree. The test detected significant
increases in the ratio from below the polytomy referred to
as node 1 in figure 4 to above this and below the second
polytomy referred to as node 2 (black versus hatched in
figure 4a), from below to above polytomy 2, and from
below to above polytomy 3. Within the (light grey) East
African branch there was no difference between ratios in
the section above versus below the origin of egg spots
(figure 4a). Ratios did not differ between the hatched
branch (between polytomies 1 and 2) and the grey
branches (above polytomy 2), nor between below and
above polytomy 2 when the ‘East African’ branch was
excluded, or between the Oreochromis/Sarotherodon branch
and anything below it. Hence, intralacustrine radiation
was significantly more common above than below the
deepest major polytomy, independent of an effect of the
‘East African’ branch, and also significantly more
common above than below the second Lake Tanganyika
polytomy (polytomy 3).
Hence, the propensity to radiate in response to
ecological opportunity is a derived property that accumu-
lated or increased sequentially within the evolutionary
history of one lineage among the African cichlids. The
decoupled cichlid pharyngeal jaw, although very likely
required, was not the key innovation that triggered
adaptive radiations in African lakes. Whether its posses-
sion affects rates of diversification should be tested by
comparing teleost families with and without a decoupled
pharyngeal anatomy. Neither maternal mouthbrooding,
nor the egg spots on the anal fin of males were required for
rapid radiation either. Maternal mouthbrooding is com-
mon in many African cichlid lineages, including those that
did not radiate in any lake (e.g. Tylochromis) but radiation is
not confined to mouthbrooding lineages. Egg dummies are
confined to the modern haplochromines, a lineage that is
nested within a group that possessed high propensity to
radiate, even before the origin of egg dummies.
The propensity for intralacustrine speciation increases
along one branch in the phylogeny, concomittent with
multiple sequentially nested lacustrine radiations. The
precursor lineages to all lineages that gave rise to
radiations in extant lakes emerged in the wake of an
ancient speciation burst 14 Myr (ago) (Terai et al. 2003;
node 1 in figure 4a,b). Further, the lineage that later gave
rise to all the large modern species flocks emerged during a
subsequent speciation burst 10 Myr (ago) in one of the
lineages that had emerged from the first burst (Takahashi
et al. 2001a,b;node2infigure 4a,b). Finally, the
haplochromine lineage that eventually radiated into the
unparalleled diversity in Lakes Victoria and Malawi was
born within the modern Lake Tanganyika radiation
(Takahashi et al. 2001b; Salzburger et al. 2002, 2005;
node 3 in figure 4a,c).
5. CONCLUSIONS
There is evidence for early bursts of speciation in every
lacustrine cichlid radiation in Africa for which data exist.
Lineage-through-time plots, and the rareness of incom-
plete lineage sorting among any two species in the older
Lakes Barombi (Schliewen & Klee 2004), and Tanga-
nyika (Sturmbauer et al. 2003; Koblmueller et al. 2004,
2005) and between the lineages of the old primary
radiation in Malawi (Parker & Kornfield 1997; Shaw et al.
2000) suggest that the frequencies of speciation
1994 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)

diminished in the course of all of the older radiations,
even though there appears to have been more than one
cycle of burst and stasis in Tanganyika and Malawi,
possibly associated with extinction events due to lake
level fluctuations. Both lakes shrunk to small fractions of
their current size several times in their history (Delvaux
1995; Cohen et al. 1997). Importantly, whereas lineage-
through-time plots or TFS can only estimate changes in
net species accumulation but cannot distinguish between
effects of changing speciation and extinction rates, the
distribution of lineage sorting in cichlid radiations
suggests speciation rates have declined as radiations
matured. Consistent with recent numerical simulations
of adaptive radiation (Gavrilets & Vose 2005), this
suggests the mechanism that makes species number
plateau is the diminution of speciation rate rather than
increasing extinction rates against a background of
consistently high speciation rates. The implication is
that speciation itself (not just coexistence) is driven by
ecological opportunity, inconsistent with predictions of
speciation through drift and ecologically neutral sexual
selection. However, it cannot be ruled out that similar
patterns might emerge if sexual niche space was
genetically (rather than ecologically) constrained such
that it will get filled up with time, limiting the
opportunity for new species to emerge through sexual
selection.
Interestingly, ecological opportunity (the availability of
an unoccupied adaptive zone), though explaining rates of
diversification in radiating lineages, is alone not sufficient
to predict whether a radiation occurs. The available data
suggest that the propensity to undergo adaptive radiation
in lakes evolved sequentially along one branch in the
phylogenetic tree of African cichlids, but is completely
absent in other lineages. Instead of attributing the
propensity for intralacustrine speciation to morphological
or behavioural innovations, it is tempting to speculate that
the propensity is explained by genomic properties that
reflect a history of repeated episodes of lacustrine radiation:
the propensity to radiate was significantly higher in
lineages whose precursors emerged from more ancient
adaptive radiations than in other lineages.
If the rapid part of adaptive radiations in cichlid fish
typically takes place in a syngameon phase in which
selection-driven incipient speciation and hybridization
interact (Seehausen et al. 1997a; Ruber et al. 2001;
Salzburger et al. 2002a,b; Smith et al. 2003; Streelman
et al. 2004) populations may become enriched in adaptive
variation at a large number of quantitative trait loci and get
rid of genetic constraints. It is likely that only a small
fraction of these loci would become fixed upon completion
of any one speciation event. Most species would retain
much adaptive genetic potential. Such enriched popu-
lations may possess an increased propensity to undergo
rapid diversification if opportunity arises again.
Further work must aim at identifying differences in
genome structure between radiating and conservative
lineages to test this hypothesis. Further work is also
required to identify the ecological and population genetic
causes of variation in speciation rates within and among
the radiating lineages. Together, such work will go a long
way towards explaining how ecological opportunity and
evolutionary history interact to regulate the occurrence
and character of adaptive radiation.
Thanks to Kyle Young for many good discussions and
valuable comments on the manuscript.
REFERENCES
Albertson, R. C., Markert, J. A., Danley, P. D. & Kocher,
T. D. 1999 Phylogeny of a rapidly evolving clade: the
cichlid fishes of Lake Malawi, East Africa. Proc. Natl Acad.
Sci. USA 96, 5107–5110. (doi:10.1073/pnas.96.9.5107)
Allender, C. J., Seehausen, O., Knight, M. E., Turner, G. F.
& Maclean, N. 2003 Divergent selection during speciation
of Lake Malawi cichlid fishes inferred from parallel
radiations in nuptial coloration. Proc. Natl Acad. Sci. USA
100,14074–14079.(doi:10.1073/pnas.2332665100)
Barraclough, T. G. & Nee, S. 2001 Phylogenetics and
speciation. Trends Ecol. Evol. 16, 391–399. (doi:10.1016/
S0169-5347(01)02161-9)
Bernatchez, L. 2003 Ecological theory of adaptive radiation:
an empirical assessment from coregonine fishes (Salmo-
niformes). In Salmonid perspectives on evolution (ed. A. P.
Hendry & S. C. Stearns), pp. 175–207. Oxford, UK:
Oxford University Press.
Cahen, L. 1954 Geologie du Congo Belge., p. 577. Liege,
Belgium: Vaillant-Carmanne.
Cohen, A. S., Lezzar, K. E., Tiercelin, J. J. & Soregan, M.
1997 New palaeogeographic and lake-level reconstruc-
tions of Lake Tanganyika: implications for tectonic,
climatic and biological evolution in a rift lake. Basin Res.
9, 107–132. (doi:10.1046/j.1365-2117.1997.00038.x)
Coyne, J. A. & Orr, H. A. 2004 Speciation. Sunderland, MA:
Sinauer Associates.
Danley, P. D. & Kocher, T. D. 2001 Speciation in rapidly
diverging systems: lessons from Lake Malawi. Mol. Ecol. 10,
1075–1086. (doi:10.1046/j.1365-294X.2001.01283.x)
Darwin, C. 1859 On the origin of species by means of natural
selection. London, UK: John Murray.
Delvaux, D. 1995 Age of Lake Malawi (Nyasa) and water
level fluctuations. Mus. R. Afr. Centre Tervuren (Belg.)
Dept. Geol. Min. Rapp. Ann. 1995–1996, 99–108.
Dieckmann, U. & Doebeli, M. 1999 On the origin of species
by sympatric speciation. Nature 400, 354–357. (doi:10.
1038/22521)
Dieckmann, U., Metz, J. A. J., Doebeli, M. & Tautz, D. (eds)
2004 Adaptive speciation. Cambridge, UK: Cambridge
University Press.
Dominey, W. J. 1984 Effects of sexual selection and life
history on speciation: species flocks in African cichlids and
Hawiian Drosophila.InEvolution of fish species flocks (ed.
A. A. Echelle & I. Kornfield), pp. 231–249. Orono, ME:
University of Maine Press.
Duftner, N., Koblmueller, S. & Sturmbauer, C. 2005
Evolutionary relationships of the Limnochromini, a tribe
of benthic deepwater cichlid fish endemic to lake
Tanganyika, East Africa. J. Mol. Evol. 60, 277–289.
(doi:10.1007/s00239-004-0017-8)
Feder, J. L., Berlocher, S. H., Roethele, J. B., Dambroski, H.,
Smith, J. J., Perry, W. L., Gavrilovic, V., Filchak, K. E. &
Rull, J. 2003 Allopatric genetic origins for sympatric host-
plant shifts and race formation in Rhagoletis. Proc. Natl
Acad. Sci. USA 100, 10 314–10 319. (doi:10.1073/pnas.
1730757100)
Freeland, J. R. & Boag, P. T. 1999 The mitochondrial and
nuclear genetic homogeneity of the phenotypically diverse
Darwin’s ground finches. Evolution 53, 1553–1563.
Fryer, G. & Iles, T. D. 1972 The cichlid fishes of the great lakes of
Africa: their biology and evolution. London, UK: Oliver and
Boyd.
African cichlid fish O. Seehausen 1995
Proc. R. Soc. B (2006)

Galis, F. & Drucker, E. G. 1996 Pharyngeal biting mechanics
in centrarchids and cichlids: insights into a key evolution-
ary innovation. J. Evol. Biol. 9, 641–670. (doi:10.1046/
j.1420-9101.1996.9050641.x)
Galis, F. & Metz, J. A. J. 1998 Why are there so many cichlid
species? Trends Ecol. Evol. 13, 1–2. (doi:10.1016/S0169-
5347(97)01239-1)
Gavrilets, S. 2004 Fitness landscapes and the origin of species.
Princeton, NJ: Princeton University Press.
Gavrilets, S. & Vose, A. 2005 Dynamic patterns of adaptive
radiation. Proc. Natl Acad. Sci. USA 102, 18 040–18 045.
(doi:10.1073/pnas.0506330102)
Genner, M. J. & Turner, G. F. 2005 The mbuna cichlids of
Lake Malawi: a model for rapid speciation and adaptive
radiation. Fish Fish. 6, 1–34.
Genner, M. J., Seehausen, O., Cleary, D. F. R., Knight,
M. E., Michel, E. & Turner, G. F. 2004 How does the
taxonomic status of allopatric populations influence
species richness within African cichlid fish assemblages?
J. Biogeogr. 31, 93–102.
Gillespie, R. 2004 Community assembly through adaptive
radiation in Hawaiian spiders. Science 303, 356–359.
(doi:10.1126/science.1091875)
Greenwood, P. H. 1974 Cichlid fishes of Lake Victoria, East
Africa: the biology and evolution of a species flock. Bull.
Br. Mus. Nat. Hist. Zool. Suppl. 6, 1–134.
Hendry, A. P., Wenburg, J. K., Bentzen, P., Volk, E. C. &
Quinn, T. P. 2000 Rapid evolution of reproductive isolation
in the wild: evidence from introduced salmon. Science 290,
516–518. (doi:10.1126/science.290.5491.516)
Higgie, M., Chenoweth, S. & Blows, M. W. 2000 Natural
selection and the reinforcement of mate recognition.
Science 290, 519–521. (doi:10.1126/science.290.5491.
519)
Joyce, D. A., Lunt, D. H., Bills, R., Turner, G. F., Katongo,
C., Duftner, N., Sturmbauer, C. & Seehausen, O. 2005
An extant cichlid fish radiation emerged in an extinct
Pleistocene lake. Nature 435, 90–95. (doi:10.1038/
nature03489)
Klett, V. & Meyer, A. 2002 What, if anything, is a
Tilapia?—mitochondrial ND2 phylogeny of tilapiines
and the evolution of parental care systems in the African
cichlid fishes. Mol. Biol. Evol. 19, 865–883.
Knight, M. E. & Turner, G. F. 2004 Laboratory mating trials
indicate incipient speciation by sexual selection among
populations of the cichild fish Pseudotropheus zebra from
Lake Malawi. Proc. R. Soc. B 271, 675–680. (doi:10.1098/
rspb.2003.2639)
Koblmu
¨
ller, S., Salzburger, W. & Sturmbauer, C. 2004
Evolutionary relationships in the sanddwelling cichlid
lineage of lake Tanganyika suggest multiple colonization of
rocky habitats and convergent origin of biparental
mouthbrooding. J. Mol. Evol. 58, 79–96.
Koblmu
¨
ller, S., Duftner, N., Katongo, C., Phiri, H. &
Sturmbauer, C. 2005 Ancient divergence in bathypelagic
Lake Tanganyika deepwater cichlids: mitochondrial phy-
logeny of the tribe Bathybatini. J. Mol. Evol. 60, 297–314.
Kocher, T. D. 2004 “Adaptive evolution and explosive
speciation: the cichlid fish model.”. Nat. Rev. Genet. 5,
288–298. (doi:10.1038/nrg1316)
Kocher, T. D., Conroy, J. A., McKaye, K. R., Stauffer, J. R. &
Lockwood, S. F. 1995 Evolution of NADH dehydrogen-
ase subunit 2 in East African cichlid fish. Mol. Phylogenet.
Evol. 4, 420–432. (doi:10.1006/mpev.1995.1039)
Kondrashov, A. S. & Kondrashov, F. A. 1999 Interactions
among quantitative traits in the course of sympatric
speciation. Nature 400, 351–354. (doi:10.1038/22514)
Kornfield, I. & Smith, P. F. 2000 African cichlid fishes: model
systems for evolutionary biology. Annu. Rev. Ecol. Syst. 31,
163–196. (doi:10.1146/annurev.ecolsys.31.1.163)
Liem, K. F. 1973 Evolutionary strategies and morphological
innovations: cichlid pharyngeal jaws. Syst. Zool. 22,
425–441.
Losos, J. B. & Schluter, D. 2000 Analysis of an evolutionary
species–area relationship. Nature 408, 847–850. (doi:10.
1038/35048558)
Machado, C. A., Kliman, R. M., Markert, J. A. & Hey, J.
2002 Inferring the history of speciation from multilocus
DNA sequence data: the case of Drosophila pseudoobscura
and close relatives. Mol. Biol. Evol. 19, 472–488.
Markert, J. A., Danley, D. D. & Arnegard, M. E. 2001 New
markers for new species: microsatellite loci and the East
African cichlids. Trends Ecol. Evol. 16, 100–107. (doi:10.
1016/S0169-5347(00)02064-4)
McCune, A. 1997 How fast is speciation? Molecular,
geological, and phylogenetic evidence from adaptive
radiations of fishes. In Molecular evolution and adaptive
radiation (ed. T. J. Givnish & K. J. Sytsma), pp. 585–610.
Cambridge, UK: Cambridge University Press.
McKaye, K. R. 1991 Sexual selection and the evolution of the
cichlid fishes of Lake Malawi, Africa. In Cichlid fishes (ed.
M. H. A. Keenleyside) Behaviour, ecology and evolution,
pp. 241–257. London, UK: Chapman & Hall.
Meyer, A. 1993 Phylogenetic relationships and evolutionary
processes in East African cichlid fishes. Trends Ecol. Evol.
8, 279–284. (doi:10.1016/0169-5347(93)90255-N)
Moran, P. & Kornfield, I. 1993 Retention of an ancestral
polymorphism in the Mbuna species flock (Teleostei:
Cichlidae) of Lake Malawi. Mol. Biol. Evol. 10,
1015–1029.
Nagl, S., Tichy, H., Mayer, W. E., Takahata, N. & Klein, J.
1998 Persistence of neutral polymorphisms in Lake
Victoria cichlid fish. Proc. Natl Acad. Sci. USA 95,
14 238–14 243. (doi:10.1073/pnas.95.24.14238)
Nagl, S., Tichy, H., Mayer, W. E., Takezaki, N., Takahata, N.
& Klein, J. 2000 The origin and age of haplochromine
fishes in Lake Victoria, East Africa. Proc. R. Soc. B 267,
519–553. (doi:10.1098/rspb.2000.1109)
Nishida, M. 1991 Lake Tanganyika as an evolutionary
reservoir of oil lineages of East African cichlid fishes:
inferences from allozyme data. Experientia 47, 974–979.
(doi:10.1007/BF01929896)
Nitecki, M. H. (ed.) 2000 Evolutionary innovations. Chicago,
IL: The University of Chicago Press.
Parchman, T. L., Benkman, C. W. & Britch, S. C. In press.
Patterns of genetic variation in the adaptive radiation of
new world crossbills (Aves: Loxia). Mol. Ecol.
Parker, A. & Kornfield, I. 1997 Evolution of the mitochon-
drial DNA control region in the mbuna (Cichlidae)
species flock of Lake Malawi, East Africa. J. Mol. Evol.
45, 70–83.
Petit, R. J. et al. 2002 Chloroplast DNA variation in European
white oaks phylogeography and patterns of diversity based
on data from over 2600 populations. Forest Ecol. Manage.
156, 5–26. (doi:10.1016/S0378-1127(01)00645-4)
Rainey, P. B., Buckling, A., Kassen, R. & Travisano, M. 2000
The emergence and maintenance of diversity: insights
from experimental bacterial populations. Trends Ecol. Evol.
15, 243–247. (doi:10.1016/S0169-5347(00)01871-1)
Ribbink, A. J., Marsh, B. A., Marsh, A. C., Ribbink, A. C. &
Sharp, B. J. 1983 A preliminary survey of the cichlid fishes
of rocky habitats in Lake Malawi. S. Afr. J. Zool. 18,
149–309.
Rosenzweig, M. L. 1978 Competitive speciation. Biol.
J. Linn. Soc. 10, 275–289.
Ruber, L., Meyer, A., Sturmbauer, C. & Verheyen, E. 2001
Population structure in two sympatric species of the Lake
Tanganyika cichlid tribe Eretmodini: evidence for intro-
gression. Mol. Ecol. 10, 1207–1225. (doi:10.1046/j.1365-
294X.2001.01259.x)
1996 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)

Salzburger, W. & Meyer, A. 2004 The species flocks of East
African cichlid fishes: recent advances in molecular
phylogenetics and population genetics. Naturwissenschaf-
ten 91, 277–290. (doi:10.1007/s00114-004-0528-6)
Salzburger, W., Baric, S. & Sturmbauer, C. 2002a Speciation
via introgressive hybridization in East African cichlids?
Mol. Ecol. 11, 619–625. (doi:10.1046/j.0962-1083.2001.
01438.x)
Salzburger, W., Meyer, A., Baric, S., Verheyen, E. &
Sturmbauer, C. 2002b Phylogeny of the Lake Tanganyika
cichlid species flock and its relationship to the central and
East African haplochromine cichlid fish faunas. Syst. Biol.
51, 113–135. (doi:10.1080/106351502753475907)
Salzburger, W., Mack, T., Verheyen, E. & Meyer, A. 2005
Out of Tanganyika: genesis, explosive speciation, key-
innovations and phylogeography of the haplochromine
cichlid fishes. BMC Evol. Biol. 5, 17. (doi:10.1186/1471-
2148-5-17)
Schliewen, U. & Klee, B. 2004 Reticulate sympatric
speciation in Cameroonian crater lake cichlids. Frontiers
Zool. 1, 1–12.
Schliewen, U., Rassmann, K., Markmann, M., Markert, J.,
Kocher, T. & Tautz, D. 2001 Genetic and ecological
divergence of a monophyletic cichlid species pair under
fully sympatric conditions in Lake Ejagham, Cameroon.
Mol. Ecol. 10, 1471–1488. (doi:10.1046/j.1365-294X.
2001.01276.x)
Schluter, D. 2000 The ecology of adaptive radiation. Oxford,
UK: Oxford University Press.
Seegers, L., Sonnenberg, R. & Yamamoto, R. 1999
Molecular analysis of the Alcolapia flock from lakes Natron
and Magadi, Tanzania and Kenya (Teleostei: Cichlidae),
and implications for their systematics and evolution.
Ichthyol. Explor. Freshw. 10, 175–199.
Seehausen, O. 1999 Speciation and species richness in African
cichlids. Leiden, The Netherlands: University of Leiden.
Seehausen, O. 2000 Explosive speciation rates and unusual
species richness in haplochromine cichlid fishes: effects of
sexual selection. Adv. Ecol. Res. 31, 237–274.
Seehausen, O. 2002 Patterns in fish radiation are compatible
with Pleistocene desiccation of Lake Victoria and 14600
year history for its cichlid species flock. Proc. R. Soc. B 269,
491–497. (doi:10.1098/rspb.2001.1906)
Seehausen, O. 2005 The cichlid fish radiations of East Africa:
a model for understanding origin and loss of biodiversity.
In Freshwater ecoregions of Africa and Madagascar—a
conservation assessment (ed. M. L. Thieme et al.),
pp. 48–51. Washington, DC: WWF and Island Press.
Seehausen, O., van Alphen, J. J. M. & Witte, F. 1997a Cichlid
fish diversity threatened by eutrophication that curbs
sexual selection. Science 277, 1808–1811. (doi:10.1126/
science.277.5333.1808)
Seehausen, O., Witte, F., Katunzi, E. F. B., Smits, J. &
Bouton, N. 1997b Patterns of the remnant cichlid fauna in
southern Lake Victoria. Conserv. Biol. 11, 890–905.
(doi:10.1046/j.1523-1739.1997.95346.x)
Seehausen, O., Koetsier, E., Schneider, M. V., Chapman,
L. J., Chapman, C. A., Knight, M. E., Turner, G. F., van
Alphen, J. J. M. & Bills, R. 2003 Nuclear markers reveal
unexpected genetic variation and a Congolese–Nilotic
origin of the Lake Victoria cichlid species flock. Proc. R.
Soc. B 270, 129–137. (doi:10.1098/rspb.2002.2153)
Shaw, P. W., Turner, G. F. T., Idid, M. R., Robinson, R. L. &
Carvalho, G. R. 2000 Genetic population structure
indicates sympatric speciation of Lake Malawi pelagic
cichlids. Proc. R. Soc. B 267, 2273–2280. (doi:10.1098/
rspb.2000.1040)
Simpson, G. G. 1953 The major features of evolution. New
York, NY: Columbia University Press.
Smith, P. F., Konings, A. & Kornfield, I. 2003 Hybrid origin
of a cichlid population in Lake Malawi: implications for
genetic variation and species diversity. Mol. Ecol. 12,
2497–2504. (doi:10.1046/j.1365-294X.2003.01905.x)
Stauffer Jr, J. R., Bowers, N. J., Kocher, T. D. & McKaye,
K. R. 1996 Evidence of hybridization between Cynotilapia
afra and Pseudotropheus zebra (Teleostei: Cichlidae)
following an intralacustrine translocation in Lake Malawi.
Copeia 1996, 203–208.
Stiassny, M. L. J. & Meyer, A. 1999 Cichlids of the rift lakes.
Sci. Am. 280, 64–69.
Streelman, J. T., Gmyrek, S. L., Kidd, M. R., Kidd, C.,
Robinson, R. L., Hert, E., Ambali, A. J. & Kocher, T. D.
2004 Hybridization and contemporary evolution in an
introduced cichlid fish from Lake Malawi National Park.
Mol. Ecol. 13, 2471–2479. (doi:10.1111/j.1365-294X.
2004.02240.x)
Sturmbauer, C., Baric, S., Salzburger, W., Rueber, L. &
Verheyen, E. 2001 Lake level fluctuations synchronize
genetic divergence of cichlid fishes in African lakes. Mol.
Biol. Evol. 18, 144–154.
Sturmbauer, C., Hainz, U., Baric, S., Verheyen, E. &
Salzburger, W. 2003 Evolution of the tribe Tropheini
from Lake Tanganyika: synchronized explosive speciation
producing multiple evolutionary parallelism. Hydrobiologia
500, 51–64. (doi:10.1023/A:1024680201436)
Takahashi, K., Nishida, M., Yuma, M. & Okada, N. 2001a
Retroposition of the AFC family of SINEs (short
interspersed repetitive elements) before and during the
adaptive radiation of cichlid fishes in Lake Malawi and
related inferences about phylogeny. J. Mol. Evol. 53,
496–507. (doi:10.1007/s002390010240)
Takahashi, K., Terai, Y., Nishida, M. & Okada, N. 2001b
Phylogenetic relationships and ancient incomplete lineage
sorting among cichlid fishes in Lake Tanganyika as
revealed by analysis of the insertion of retroposons. Mol.
Biol. Evol. 18, 2057–2066.
Terai, Y., Takahashi, K., Nishida, M., Sato, T. & Okada, N.
2003 Using SINEs to probe ancient explosive speciation:
“Hidden” radiation of African cichlids? Mol. Biol. Evol.
20, 924–930. (doi:10.1093/molbev/msg104)
Turelli, M., Barton, N. H. & Coyne, J. A. 2001 Theory and
speciation. Trends Ecol. Evol. 16, 330–343. (doi:10.1016/
S0169-5347(01)02177-2)
Turner, G. F. 1999 Explosive speciation of African cichlid
fishes. In The evolution of biological diversity (ed. A. E.
Magurran & R. M. May), pp. 217–229. Oxford, UK:
Oxford University Press.
Turner, G. F., Seehausen, O., Knight, M. E., Allender, C. J.
& Robinson, R. L. 2001 How many species of cichlid
fishes are there in African lakes? Mol. Ecol. 10, 793–806.
(doi:10.1046/j.1365-294x.2001.01200.x)
Turner, G. F., Robinson, R. L., Shaw, P. W. & Carvalho,
G. R. 2004 Identification and biology of Diplotaxodon,
Rhamphochromis and Pallidochromis.InThe cichlid diversity
of Lake Malawi/Nyasa: identification, distribution and
taxonomy (ed. J. Snoeks), pp. 198–251. El Paso, TX:
Cichlid Press.
Ungerer, M. C., Baird, S. J. E., Pan, J. & Rieseberg, L. R.
1998 Rapid hybrid speciation in wild sunflowers. Proc.
Natl Acad. Sci. USA 95, 11 757–11 762. (doi:10.1073/
pnas.95.20.11757)
Vamosi, S. M. 2003 The presence of other fish specues affects
speciation in threespine sticklebacks. Evol. Ecol. Res. 5,
717–730.
Van Alphen, J. J. M., Seehausen, O. & Galis, F. 2004
Speciation and radiation in African haplochromine cichlids.
In Adaptive speciation (ed. U. Dieckmann, M. Doebeli,J. A. J.
Metz & D. Tautz) Cambridge studies in adaptive dynamics.
Cambridge, UK: Cambridge University Press.
African cichlid fish O. Seehausen 1997
Proc. R. Soc. B (2006)

Van Doorn, G. S. & Weissing, F. J. 2001 Ecological versus
sexual selection models of sympatric speciation: a
synthesis. Selection 2, 17–40. (doi:10.1556/Select.2.2001.
1-2.3)
Van Oppen, M. J. H., Willis, B. L., Van Rheede, T. & Miller,
D. J. 2002 Spawning times, reproductive compatibilities
and genetic structuring in the Acropora aspera group:
evidence for natural hybridization and semi-permeable
species boundaries in corals. Mol. Ecol. 11, 1363–1376.
(doi:10.1046/j.1365-294X.2002.01527.x)
Verheyen, E., Salzburger, W., Snoeks, J. & Meyer, A. 2003
Origin of the superflock of cichlid fishes from Lake
Victoria, East Africa. Science 300, 325–329. (doi:10.
1126/science.1080699)
Verschuren, D., Johnson, T. C., Kling, H. J., Edgington,
D. N., Leavitt, P. R., Brown, E., Talbot, M. R. & Hecky,
R. E. 2002 History and timing of human impact on Lake
Victoria, East Africa. Proc. R. Soc. B 269, 289–294.
(doi:10.1098/rspb.2001.1850)
West-Eberhard, M. J. 1989 Phenotypic plasticity and the
origins of diversity. Annu. Rev. Ecol. Syst. 20, 249–278.
(doi:10.1146/annurev.es.20.110189.001341)
Witte, F., Goldschmidt, T., Wanink, J., van Oijen, M.,
Goudswaard, K., Witte-Maas, E. & Bouton, N. 1992 The
destruction of an endemic species flock: quantitative data
on the decline of the haplochromine cichlids of Lake
Victoria. Environ. Biol. Fish. 34, 1–28. (doi:10.1007/
BF00004782)
Witte, F., Msuku, B. S., Wanink, J. H., Seehausen, O.,
Katunzi, E. F. B., Goudswaard, P. C. & Goldschmidt, T.
2000 Recovery of cichlid species in Lake Victoria:
an examination of factors leading to differential extinction.
Rev. Fish Biol. Fish. 10,233–241.(doi:10.1023/
A:1016677515930)
Won, Y.-J., Sivasundar, A., Wang, Y. & Hey, J. 2005 On the
origin of Lake Malawi cichlid species: a population genetic
analysis of divergence. Proc. Natl Acad. Sci. USA 102
(Suppl. 1), 6581–6586. (doi:10.1073/pnas.0502127102)
1998 O. Seehausen African cichlid fish
Proc. R. Soc. B (2006)