Biological Journal of the Linnean Society, 2016, 117, 812–822. With 2 figures.
Introgressive hybridization and natural selection in
Darwin’s finches
PETER R. GRANT* and B. ROSEMARY GRANT
Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ, 08544-1003,
USA
Received 22 July 2015; revised 13 September 2015; accepted for publication 13 September 2015
Introgressive hybridization, i.e. hybridization with backcrossing, can lead to the fusion of two species, but it can
also lead to evolution of a new trajectory through an enhancement of genetic variation in a new or changed
ecological environment. On Daphne Major Island in the Gal
apagos archipelago, ~1–2% of Geospiza fortis finches
breed with the resident G. scandens and with the rare immigrant species G. fuliginosa in each breeding season.
Previous research has demonstrated morphological convergence of G. fortis and G. scandens over a 30-year
period as a result of bidirectional introgression. Here we examine the role of hybridization with G. fuliginosa in
the evolutionary trajectory of G. fortis. Geospiza fuliginosa (~12 g) is smaller and has a more pointed beak than
G. fortis (~17 g). Genetic variation of the G. fortis population was increased by receiving genes more frequently
from G. fuliginosa than from G. scandens (~21 g). A severe drought in 2003–2005 resulted in heavy and selective
mortality of G. fortis with large beaks, and they became almost indistinguishable morphologically from
G. fuliginosa. This was followed by continuing hybridization, a further decrease in beak size and a potential
morphological fusion of the two species under entirely natural conditions. © 2015 The Linnean Society of London,
Biological Journal of the Linnean Society, 2016, 117, 812–822.
ADDITIONAL KEYWORDS: adaptive radiation – hybrid zone – introgression – natural selection – speciation reversed.
INTRODUCTION
Many vertebrate species achieve post-zygotic reproductive isolation only after millions of years have
elapsed since their initial separation from a common
ancestor (Prager & Wilson, 1980; Price & Bouvier,
2002; Fitzpatrick, 2004). Before that point is reached
they may be prone to hybridize. The consequences of
hybridization vary according to how different the
species or diverging lineages have become prior to
hybridizing.
In the wild hybridization often takes the form of
introgression, i.e. F1 hybrids backcross to one or other
of the parental species (Arnold, 2006; Schwenk, Brede
& Streit, 2008; Abbott et al., 2013). At one extreme
speciation may collapse and the two species fuse into
one (Dobzhansky, 1941; Grant et al., 2004; Taylor
et al., 2006; Grant & Grant, 2008, 2014a; Behm, Ives
& Boughman, 2010; Webb, Marzluff & Omland, 2011;
*Corresponding author. E-mail: [email protected]
812
Garrick et al., 2014; Kleindorfer et al., 2014). At the
other extreme divergence continues, facilitated by
enhanced genetic variation and reduced covariation
among traits (Grant & Grant, 1994; Hedrick, 2013;
Selz et al., 2014), and by the transfer of selectively
favoured alleles under the influence of natural or sexual selection (Pfennig & Pfennig, 2012). Filtered introgression, in which some genes are exchanged and
others are not, has been reported in some animals
(Lerner et al., 2013; Parchman et al., 2013; Liu et al.,
2014; Pereira, Barreto & Burton, 2014) as well as
plant taxa (Martinsen et al., 2001; Kim et al., 2008).
Introgression may give rise to new lineages that have
the potential to become new species (Schwarzbach &
Rieseberg, 2002; Grant & Grant, 2009, 2014a; Brelsford, Mil
a & Irwin, 2011; Hermansen et al., 2011; Mil
a
et al., 2011; Amarel et al., 2014), and may be sufficiently profound to create a swarm of recombinant
lines that form the starting point of an adaptive radiation (Seehausen, 2004; Albertson & Kocher, 2005;
Hasselman et al., 2014).
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
DARWIN’S FINCHES
Given its prevalence and importance, hybridization
has moved to the central stage of research into the
causes of biological diversity (reviewed by Schwenk
et al., 2008; Rheindt & Edwards, 2011; Abbott et al.,
2013). Here we examine the consequences in a wellstudied group of Darwin’s finches.
DARWIN’S FINCHES
Fourteen species of Darwin’s finches occur on the
Gal
apagos islands, having diversified from a common ancestor over the last 1–2 Myr. Analyses of
finch genomes show they have been exchanging
genes by hybridizing throughout most of their history (Lamichhaney et al., 2015). Rare hybridization
appears to be currently widespread among several
sympatric pairs of closely related species throughout
the archipelago (Grant, Grant & Petren, 2005). A
long-term study on the small island of Daphne
Major has provided detailed evidence of contemporary hybridization between species of similar size,
on the order of 1–2% per breeding episode, typically
as a result of learning the song of another species
(Grant et al., 2004; Grant & Grant, 2010, 2014a).
The medium ground finch Geospiza fortis Gould
hybridizes with the resident cactus finch G. scandens (Gould) and rare immigrants of the small
ground finch G. fuliginosa Gould. None of these
three species hybridizes on Daphne with the much
larger G. magnirostris Gould (~30 g). There is no
evidence of genetic incompatibilities among this
young group of species, which is estimated to have
originated in the last 200 000 years (Lamichhaney
et al., 2015).
The beaks of G. fortis on Daphne Major Island
became progressively more pointed over a period of
three decades, and those of G. scandens became progressively blunter (Grant & Grant, 2014a). These
two trends were apparently caused by an exchange
of genes between G. fortis and G. scandens, and
between G. fortis and G. fuliginosa, which was facilitated by a change in feeding conditions associated
with an El Ni~
no event in 1982–1983. We proposed
that a shift towards a relative abundance of smaller
seeds favoured the dry-season survival of hybrids.
The hypothesis of introgression to explain the morphological pattern was based on observations of persistent hybridization at low frequency (Grant &
Grant, 1992; Grant, 1993), relatively high survival
and breeding success of the hybrids after 1982, measurements of fully grown hybrid offspring of interspecific pairs (Grant & Grant, 1994), limited direct
evidence of exchange of microsatellite alleles (Grant
& Grant, 2014a) and a quantified shift in seed profiles (Grant & Grant, 1993).
813
The hypothesis of introgression between G. fortis
and G. scandens was tested by five predictions, and
supported by the results (Grant & Grant, 2014a).
Support was stronger for G. scandens than for
G. fortis. There are two reasons for this. First, effects
of G. fuliginosa were omitted from the analyses
because they do not have an independent breeding
population on the island. Second G. fortis, unlike
G. scandens, experienced strong directional selection
on beak size in 2004–2005 associated with a prolonged drought that began in 2003: small size was
selectively favoured (Grant & Grant, 2006). Therefore, morphological change over the years was a step
function of time produced by both selection and
introgression. The independent contribution of introgression to the change in G. fortis was not analysed.
The focus of this paper is the effect of introgression
of G. fuliginosa genes on G. fortis morphology (body
size and beak dimensions). We test two predictions
of the introgression hypothesis. Given the high additive genetic variation underlying beak traits (Grant
& Grant, 1994), we predict that (1) genetically identified hybrids are morphologically intermediate
between the parental species groups, and (2) the
morphological trajectory of G. fortis (including the
hybrids) is determined by the relative frequencies of
the respective hybrids in the absence of selection. To
quantify the separate phenotypic effects on G. fortis
of introgression from G. fuliginosa (a smaller species) and G. scandens (a larger species) we compare
genetically identified hybrids with G. fortis in the
period of weak or no selection on G. fortis (1987–
2004; Table 1). We also quantify the consequences of
selection on G. fortis. We compare G. fortis morphology with the two groups of hybrids and G. fuliginosa
before and after selection in 2004–2005. We show
that in the period after selection G. fortis became
indistinguishable from G. fuliginosa in some traits
as a result of selection and continuing introgression.
These traits are biologically significant because they
function in species recognition and mate choice
(Grant & Grant, 2008, 2014a).
Table 1. Natural selection experienced by G. fortis after
1986
1991–1992
2003–2004
2004–2005
2009–2010
Body size
Beak size
Large + 0.08
Large + 0.23
Small
0.31
Small
Beak shape
0.48
Pointed
0.14
Numbers are statistically significant selection coefficients
(standardized directional selection differentials). After
Grant & Grant (2014a).
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
814
P. R. GRANT and B. R. GRANT
MATERIAL AND METHODS
Field methods have been described in several previous publications (Grant, 1993; Grant & Grant, 2008,
2014a). Beginning in 1973 and continuing every
year, we captured birds in mist-nets, weighed them
(g) with a Pesola balance, measured (mm) the length
of wing, tarsus and beak, as well as the depth and
width of the beak, with calipers, dividers and a
ruler, and banded them uniquely with one numbered
metal and three coloured PVC leg bands prior to
release. Nestlings were banded at day 8 and measured when captured in nets in subsequent years
when fully grown. From 1988 to 2012 we took a
small drop of blood by venipuncture from every bird,
transferred it to EDTA-soaked filter paper and
stored the papers in a jar of drierite (silica gel). For
genetic characterization of individual finches we
used the 14 autosomal microsatellite loci isolated by
Petren (1998).
Two principal components analyses were performed
on correlation matrices from the total data set to
reduce the inter-correlated variables to synthetic and
interpretable size and shape variables. Weight, wing
and tarsus length were used for an analysis of body
size (PC1), and the three beak measurements were
used for a separate analysis of beak size (PC1) and
beak shape (PC2). The percentage variance explained
by PC1 body is 83.8, and all variables loaded approximately equally. The percentage variance explained in
the beak analysis was 74.3 for PC1 and an additional
22.1 for PC2. All variables loaded strongly on PC1.
For PC2 the eigenvectors were negative and approximately equal for beak depth and width, and larger
and positive for beak length. PC2 is therefore interpreted as an index of pointedness (or bluntness) of
the beak (Grant & Grant, 2014a).
Assignment tests with the program STRUCTURE
(Pritchard, Stephens & Donnelly, 2000; Pritchard,
Wen & Falush, 2007) were performed on 2105
finches initially classified by morphology into three
groups (G. fortis, G. fuliginosa and G. scandens): see
Grant & Grant (2010) for procedures and model specifications. We used an ancestry model and set the
number of previous generations to two because it
performed better than a model with only one or three
previous generations. In this analysis an individual
may be genetically identified with an estimated probability of belonging to another species (generation 0),
having a parent (generation 1) or having a grandparent (generation 2) from another species. We used a
no-admixture model in preference to an admixture
model because it performed better with individuals
known from pedigree analysis in the early part of
the study to be hybrids (Grant & Grant, 1994, 2010).
Divergence of allele frequencies among populations
(net nucleotide distance) was computed by the program.
Of the 2105 finches, 1194 were assigned by
microsatellite genotype to G. fortis, 541 to G. scandens and 67 to G. fuliginosa with probabilities of
≥ 0.9 (mostly ≥ 0.98). The remaining 303 individuals
(14.4%) did not meet the 90% criterion. They had
mixed assignments and are therefore considered to
be hybrids or backcrosses. Backcross assignments
predominated. However, we make no distinction
between F1 hybrids and first-generation backcrosses
because a simulation showed that the difference
between them cannot be reliably detected with the 14
microsatellite loci (Grant & Grant, 2010). Secondand later generation backcrosses are not identified by
this method and are included in the respective species groups. For 286 of the 303 individuals the assignment probabilities to two species summed to 0.9 or
more. They are treated as members of the G. fortis
group as potential breeders and listed as G. fortis 9 G. scandens if the assignment probability is
greater to G. fortis than to G. scandens; as part of
the G. scandens group and listed as G. scandens 9 G. fortis if the assignment probability to
G. scandens is higher than to G. fortis; and as G. fortis 9 G. fuliginosa for a mixture of G. fortis and
G. fuliginosa. Geospiza fuliginosa exists as a result of
rare recurrent immigration, it does not have an independent breeding population on the island, and hence
there is no backcrossing to this species. For the
remaining 17 individuals (< 1%) assignment probabilities exceeded 0.1 for each of the three species. These
individuals are treated as G. fortis 9 G. scandens
hybrids because G. scandens and G. fuliginosa have
never been observed to hybridize. They are the result
of hybridization of G. scandens with G. fortis 9 G. fuliginosa F1 hybrids or backcrosses to
G. fortis (Grant & Grant, 2010).
One-way ANOVAs were used to test for
heterogeneity among samples of measurements,
and Tukey–Kramer highest significant difference
(HSD) tests were used for assessing differences in
paired comparisons of species and hybrid groups.
Non-parametric tests (Wilcoxon Matched Pairs
Signed Ranks test and Median test) were used for
testing differences between samples of frequency
data. All statistical tests were performed in JMP
version 10, and illustrations were prepared in SigmaPlot.
RESULTS
H YBRID
INTERMEDIACY
The first prediction, that genetically identified hybrids
are morphologically intermediate between the paren-
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
815
DARWIN’S FINCHES
tal species groups, is upheld by the data in Table 2
and statistical tests with PC scores. There is a perfect
rank order correlation in average weight, wing and
beak length from the group of smallest average size
(G. fuliginosa) to the largest (G. scandens). The linear
trend is illustrated with beak length in Figure 1.
Geospiza fortis 9 G. fuliginosa hybrids and the two
parental species are significantly different in body size
(PC1, F2,1385 = 36.36, P < 0.0001), beak size (PC1,
P < 0.0001)
and
beak
shape
F2,1385 = 68.19,
(F2,1385 = 51.34, P < 0.0001), although the specific difference in body size between G. fortis 9 G. fuliginosa
and G. fortis is not significant (Tukey’s HSD test,
P = 0.9835). There is similar heterogeneity among
G. fortis, G. scandens and their hybrids in all three
traits (all F2,1797 > 100, P < 0.0001), although G. fortis and the G. fortis 9 G. scandens hybrids do not differ in beak size (Tukey’s HSD test, P = 0.2359). The
two groups of hybrids are not exactly intermediate
between the parental species because most are likely
to be backcrosses to G. fortis (see Methods). These
results confirm earlier findings of hybrid intermediacy
from a morphological analysis without genetic identification of species and hybrids (Grant & Grant, 1994).
M ORPHOLOGICAL
TRAJECTORIES THROUGH TIME
Over the period 1987–2004, body size of G. fortis (including unidentified hybrids) remained the same
whereas beaks became smaller on average and more
pointed (Fig. 2). These trends continued after selection occurred in 2004–2005 (Table 1). Geospiza fuligi-
nosa did not change in any of the traits either before
or after selection (N = 32 in each period; all
P > 0.05).
F REQUENCIES
OF HYBRIDS THROUGH TIME
Across the study period as a whole (1981–2012),
G. fortis 9 G. fuliginosa hybrids (118) outnumbered
G. fortis 9 G. scandens hybrids (63) in the samples
by almost two to one. In the period 1987–2004, there
were 80 G. fortis 9 G. fuliginosa hybrids and 30
G. fortis 9 G. scandens hybrids. In eight out of the
nine years during this period when the combined
samples of hybrids and G. fortis exceeded 25, the
G. fortis 9 G. fuliginosa frequencies (mean = 0.105)
were greater than the G. fortis 9 G. scandens frequencies
(mean = 0.034;
Wilcoxon
S = 20.5,
P = 0.0117). In the 5 years after selection with the
same minimum sample sizes the frequencies of
G. fortis 9 G. scandens hybrids (0.082, N = 31) were
almost the same as G. fortis 9 G. fuliginosa hybrids
(0.089, N = 35).
E FFECTS
OF SELECTION IN
2004–2005
As a result of selection G. fortis became more similar
to the smaller species (G. fuliginosa) and less similar
to the larger species (G. scandens). Before selection
G. fortis were clearly distinct from G. fuliginosa in
their larger and blunter beaks, as well as in average
body size (Table 3). After selection and with continuing introgression G. fortis were no longer distinguishable from G. fuliginosa in the two size traits,
Table 2. Morphological features of species and hybrids
Species and hybrids
G. fuliginosa (N = 65)
G. fortis 9 G. fuliginosa
(N = 129)
G. fortis (N = 1194)
G. fortis 9 G. scandens
(N = 65)
G. scandens 9 G. fortis
(N = 111)
G. scandens (N = 543)
Mean
SE
Mean
SE
Mean
SE
Mean
SE
Mean
SE
Mean
SE
Weight
Wing
Tarsus
Beak
length
Beak
depth
Beak
width
13.68
0.37
15.50
0.23
15.65
0.05
16.46
0.28
17.74
0.24
19.12
0.09
63.91
0.52
66.70
0.26
66.93
0.07
68.14
0.37
69.12
0.30
70.86
0.12
18.57
0.16
19.09
0.09
18.98
0.02
19.30
0.13
19.97
0.10
20.56
0.04
9.73
0.19
10.59
0.11
10.67
0.02
11.60
0.16
12.26
0.13
13.65
0.04
7.78
0.17
8.73
0.10
9.00
0.02
8.99
0.09
9.06
0.07
9.02
0.03
7.60
0.14
8.34
0.07
8.55
0.02
8.54
0.07
8.73
0.07
8.61
0.02
PC1 body
1.8541
0.2475
0.6697
0.1340
0.6519
0.0284
0.0846
0.1748
0.7312
0.1470
1.6665
0.0526
PC1
beak
2.2870
0.3145
0.6228
0.1676
0.1995
0.0372
0.0546
0.1653
0.4742
0.1351
0.7345
0.0449
PC2
beak
0.0501
0.0654
0.3256
0.0409
0.4883
0.0101
0.0115
0.0769
0.2744
0.0651
1.1055
0.0217
Weight is in grams, and the remainder in millimetres. Positive PC (principal components) means indicate large body
and beak size (PC1) and pointed beaks (PC2). Note G. fortis 9 G. scandens are genetically more similar to G. fortis than
to G. scandens, and G. scandens 9 G. fortis are genetically more similar to G. scandens than to G. fortis. Data are from
the whole study period 1981–2012.
N, sample size; SE, standard error of the mean.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
816
P. R. GRANT and B. R. GRANT
Figure 1. Mean beak lengths (and 95% CIs) of three
interbreeding species on Daphne Major Island and their
hybrids. Abbreviations: ful, G. fuliginosa; Ff, G. fortis 9
G. fuliginosa; Fort, G. fortis; FS, G. fortis 9 G. scandens;
SF, G. scandens 9 G. fortis; Scan, G. scandens. Data
from Table 2.
but their beaks remained more blunt on average,
although this received weak statistical support. Differences in average beak traits between G. fortis and
G. fortis 9 G. fuliginosa hybrids disappeared.
Before selection G. fortis and G. fortis 9 G. scandens hybrids were indistinguishable in size traits,
but differed in beak shape: G. fortis had blunter
beaks on average. After selection G. fortis had
become much smaller than the hybrids in both size
measures, and their beaks remained blunter than
the beaks of G. fortis 9 G. scandens hybrids.
DISCUSSION
ON DAPHNE MAJOR ISLAND
H YBRIDIZATION
The situation on Daphne is dynamic, and possibly
unique in the Gal
apagos. Hybridization is tri-specific.
Introgression is bidirectional but unequal between
G. fortis and G. scandens, with genes being transmitted more from G. fortis to G. scandens than in
reverse, and unidirectional from G. fuliginosa to
G. fortis (Grant & Grant, 2010). We have previously
shown that G. scandens has become morphologically
more similar to G. fortis. Here we show that G. fortis
has become more similar to G. fuliginosa (Table 3).
These effects are explicable in terms of the morphological differences among the species and the relative
frequencies of hybridization. Geospiza fuliginosa is
smaller than G. fortis in body size and beak size,
and G. scandens is larger than G. fortis in both
traits. Both of these species have more pointed beaks
than G. fortis (Table 2).
As predicted, average beak size of G. fortis
declined gradually when introgression of genes from
G. fuliginosa greatly exceeded introgression from
G. scandens in the period 1987–2004 (Fig. 2). Aver-
age beak size dropped sharply as a result of strong
natural selection in the drought of 2004–2005 (Grant
& Grant, 2006, 2008). The decline in beak size did
not persist to a statistically significant extent during
the subsequent period of continuing introgression
and no selection (2005–2012). A possible reason is
that the effects of G. fuliginosa genes on average
beak size of G. fortis were weakened but not completely nullified by opposite effects of introgression of
G. scandens genes.
Average body size of G. fortis remained approximately constant from 1987 to 2004, followed by a
short-term decrease associated with the 2004–2005
selection event, and no further change. The probable
reason is that genes introgressed from the two species had opposite effects. The combination of occasional selection for large body size (Table 1) and
introgression of G. scandens genes may have been
roughly balanced by introgression of G. fuliginosa
genes. In contrast, genes affecting beak shape that
introgressed from the two species had the same
effect on G. fortis of increasing the degree of pointedness.
The interpretation based on numbers of hybrids
neglects genetic differences between the species. A
greater per-capita genetic effect on G. fortis morphology is to be expected from introgression of
G. scandens genes than from G. fuliginosa genes,
for two reasons. First, G. fortis differs genetically
more from G. scandens than from G. fuliginosa.
The net nucleotide distance at microsatellite loci
between G. scandens and G. fortis is 0.76, in
contrast to only 0.28 between G. fuliginosa and
G. fortis. Second, G. scandens and G. fortis have
different beak shape allometries, whereas G. fortis
and G. fuliginosa allometries are similar (Boag,
1984; Grant & Grant, 1994). Nonetheless G. fortis hybridized much more often with G. fuliginosa
than with G. scandens before the selection
event of 2004–2005, and the net effect of interbreeding was a predominant influence of G. fuliginosa genes over G. scandens genes in the G. fortis
population.
Genetic inferences are grounded in two morphological surrogates, beak size and beak shape. The connection between morphology and genes is
underpinned by high heritable variation and polygenic inheritance of the metric traits, and covariation within the G. fortis and G. scandens populations
(Grant & Grant, 1994). The connection is further
strengthened by identified genes that influence the
development of beak size and shape through the production of signalling molecules (Abzhanov et al.,
2004, 2006; Mallarino et al., 2011), and by the variable effects of ALX1, a regulatory gene affecting beak
shape (Lamichhaney et al., 2015).
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
DARWIN’S FINCHES
817
Figure 2. Regression of principal components scores of G. fortis on years before and after selection in 2004–2005. Each
individual is entered only in the first year of measurement. Body size: before selection F1,799 = 0.4003, P = 0.5271
(b = 0.0042 0.0066); after selection F1,340 = 0.7582, P = 0.3845 (b = 0.0212 0.0243). Beak size: before selection
F1,799 = 8.2738, P = 0.0041 (b = 0.0243 0.0084); after selection F1,340 = 3.8061, P = 0.0519 (b = 0.0558 0.0286).
Beak shape: before selection F1,799 = 4.7363, P = 0.0298 (b = 0.0050 0.0023); after selection F1,340 = 14.9446,
P = 0.0001 (b = 0.0344 0.0089). Beaks became progressively more pointed.
I SLANDS
AND CONTINENTAL HYBRID ZONES
Parapatric, closely related species of birds often
hybridize at their borders, giving rise to broad or
narrow hybrid zones that may be stable (Mettler &
Spellman, 2009) or directionally moving (Rowher,
Bermingham & Wood, 2001; Gill, 2004; Curry, 2005;
Secondi, Faivre & Bensch, 2006). The hybrids are
typically recognized by a combination of plumage
traits that distinguish the parental species. The
study we report here is unusual in two respects.
First, the environment is a small island, there is no
spatial separation of the hybridizing species, and
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
818
P. R. GRANT and B. R. GRANT
Table 3. Morphological comparison of two species and their hybrids before and after selection in 2004–2005
Before selection
PC1 body
Comparison
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
fortis
fuliginosa
fortis 9 G.
fuliginosa
fortis 9 G.
fortis
fortis 9 G.
fuliginosa
fortis 9 G.
fortis 9 G.
fortis 9 G.
fortis
fuliginosa
fuliginosa
scandens
scandens
fuliginosa
scandens
0.4831
2.5547
0.7059
2.5547
0.7059
0.4831
0.0053
2.5547
0.0053
0.7059
0.0053
0.4831
F3,999 = 40.193
After selection
PC1 body
P
< 0.0001
< 0.0001
0.2957
< 0.0001
0.0105
0.0668
PC1 beak
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
fortis
fuliginosa
fortis 9 G.
fuliginosa
fortis 9 G.
fortis
fortis 9 G.
fuliginosa
fortis 9 G.
fortis 9 G.
fortis 9 G.
fortis
fuliginosa
fuliginosa
scandens
scandens
fuliginosa
scandens
0.1172
3.0864
0.5358
3.0864
0.5358
0.1172
0.0444
3.0864
0.0444
0.5358
0.0444
0.1172
F3,999 = 60.790
fortis
fuliginosa
fortis 9 G.
fuliginosa
fortis 9 G.
fortis
fortis 9 G.
fuliginosa
fortis 9 G.
fortis 9 G.
fortis 9 G.
fortis
fuliginosa
fuliginosa
scandens
scandens
fuliginosa
scandens
0.5438
0.0415
0.3834
0.0415
0.3834
0.5438
0.0406
0.0415
0.0406
0.3834
0.0406
0.5438
F3,999 = 48.832
0.9897
0.0230
0.0234
0.0015
0.2844
< 0.0001
PC1 beak
< 0.0001
< 0.0001
0.0002
< 0.0001
0.1651
0.9905
PC2 beak
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
G.
1.0724
1.1318
0.6023
1.1318
0.6023
1.0274
0.1715
1.1318
0.1715
0.6023
0.1715
1.0724
F3,446 = 9.386
P
0.9883
1.4626
0.7852
1.4626
0.7852
0.9883
0.0662
1.4626
0.0662
0.7852
0.0662
0.9883
F3,445 = 8.600
0.1763
0.0939
0.7397
< 0.0001
0.0225
< 0.0001
PC2 beak
< 0.0001
< 0.0001
0.0008
0.7950
< 0.0001
< 0.0001
0.3499
0.1445
0.2177
0.1445
0.2177
0.3499
0.0706
0.1445
0.0706
0.2177
0.0706
0.3499
F3,445 = 12.762
0.0273
0.8555
0.1547
0.1446
0.0117
< 0.0001
Mean PCs (principal components scores) are shown. The two components for beak variation refer to beak size (PC1) and
beak shape (PC2). Positive PC means indicate large body and beak size (PC1) and pointed beaks (PC2). All F tests of
heterogeneity for each group and trait combination (ANOVAs) are significant at P < 0.0001. P values in the table refer
to pairwise comparisons by Tukey–Kramer HSD tests. Sample sizes before and after selection are, respectively, 852 and
342 G. fortis, 33 and 32 G. fuliginosa, 84 and 45 G. fortis 9 G. fuliginosa, and 34 and 31 G. fortis 9 G. scandens.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
DARWIN’S FINCHES
therefore hybrids and the parental species are completely intermingled; there is no hybrid zone. The
small size of the island and low population sizes,
together with the lack of migration, are advantages
in the search for ecological and behavioural causes of
hybridization as well as fitness consequences. A few
other studies share these advantages (Ryan, Moloney
& Hudon, 1994; Wiley et al., 2007; Kleindorfer et al.,
2014). Second, the key morphological traits that separate species vary continuously and not discretely,
unlike the majority of species studied elsewhere. We
have used genetic markers in addition to observations and morphological measurements to identify
hybrids.
Natural and/or sexual selection together with
recurrent hybridization and backcrossing have been
implicated in the maintenance of spatially restricted
hybrid zones in birds (McDonald et al., 2001; Bronson et al., 2005) and other animals (Taylor, Larson &
Harrison, 2015). These processes can be inferred
genomically (Irwin et al., 2009; Toews, Brelsford &
Irwin, 2011; Baldassarre et al., 2014; Poelstra et al.,
2014) but are difficult to study directly because
known individuals disperse out of local study areas,
and without radio-tracking are lost to the observers.
The study on Daphne is able to show the separate
effects of natural selection and introgressive
hybridization on the morphological trajectory of a
population through time. The two effects combined
synergistically (Grant & Grant, 2014b): introgression
from G. fuliginosa to G. fortis led to small average
beak size, and subsequent selection on the same trait
led to a further reduction in average beak size and
continuing introgression. An expected increase in frequency of hybridization with G. fuliginosa was not
detected in the small number of samples after selection. In other situations elsewhere selection and
hybridization can have opposing effects, for example
under the numerous conditions where F1 hybrids
and backcrosses are at a selective disadvantage for
reasons of viability, fertility or mate acquisition
(Price, 2008). In these cases the opposing processes
act essentially as stabilizing selection.
819
the species differ considerably in size. Thus, overall
size, and size and shape of the beak are important
cues used in the choice of mates (Grant & Grant,
2008, 2014a). They constitute a barrier to interbreeding of species.
The study provides insight into the breakdown of the
barrier, resulting in a potential collapse of two species
into one. The number of examples in the literature
where this is known or suspected is increasing, and
almost all involve a human influence (e.g. Taylor et al.,
2006; Vonlanthen et al., 2012). Kleindorfer et al. (2014)
have reported an apparent example from a study of tree
finches (Camarhynchus spp.) on the disturbed
Gal
apagos island of Floreana. Weakening of the barrier
between G. fortis and two congeners is exceptional in
occurring under natural circumstances. The key factor
was a natural change in the environment, specifically a
change in the composition of the vegetation and seed
supply in dry seasons induced by an exceptional amount
of rain in 1983 (Grant & Grant, 1993, 2014a). This
enabled hybrids to survive to the following breeding seasons. Natural fusion of species is to be expected in other
taxa where populations are small, they occupy confining
habitats such as tops of mountains or small ponds, and
the environment is naturally perturbed. The outcome
may be reversed by a reversal of the environmental
state, or it may lead to the formation of a new species
along a novel evolutionary pathway (Larsen, March
anRivadeneira & Baker, 2010). This has also happened on
Daphne with the formation of a new, reproductively isolated, lineage (Grant & Grant, 2014a, b).
ACKNOWLEDGEMENTS
We thank the Gal
apagos National Park Service and
the Charles Darwin Research Station for permits
and logistical support, the National Science Foundation (USA) for financial support, the many field
assistants who have assisted in data collection, and
William Gilbert and two anonymous referees for
helpful comments on the manuscript.
REFERENCES
S PECIATION
IN REVERSE
None of the three species has hybridized with a
fourth species on Daphne, the much larger G. magnirostris. Geospiza fuliginosa and G. scandens, also
of disparate size, have not interbred, although genes
may pass from G. fuliginosa to G. scandens via the
conduit of G. fortis (Grant, 1993; Grant & Grant,
2010): microsatellite profiles of a few G. fortis individuals appear to be a mosaic of genes from all three
species (see Methods). Even learning the song of
another species does not lead to hybridization when
Abbott R, Albach D, Ansell S, Arntzen JW, Baird SJE,
Beirne N, Boughman J, Brelsford A, Buerkle CA,
Buggs R, Butlin RK, Dieckmann U, Eroukhmanoff F,
Grill A, Cahan SH, Hermansen JS, Hewitt G, Hudson
AG, Jiggins C, Jones J, Keller B, Marczewski T, Mallet J, Martinez-Rodriguez P, M€
ost M, Mullen S,
Nicholls R, Nolte AW, Parisod C, Pfennig K, Rice AM,
Ritchie MG, Seifert B, Smadja CM, Stelkens R, Szymura JM, V€
ain€
ol€
a, Wolf JBW, Zinner D. 2013.
Hybridization and speciation. Journal of Evolutionary Biology 26: 229–246.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
820
P. R. GRANT and B. R. GRANT
Abzhanov A, Protas M, Grant BR, Grant PR, Tabin CJ.
2004. Bmp4 and morphological variation of beaks in Darwin’s finches. Science 305: 1462–1465.
Abzhanov A, Kuo WP, Hartmann C, Grant PR, Grant
BR, Tabin CJ. 2006. The calmodulin pathway and evolution of elongated beak morphology in Darwin’s finches. Nature 442: 563–567.
Albertson RC, Kocher TD. 2005. Genetic architecture sets
limits on transgressive segregation in hybrid cichlid fishes.
Evolution 59: 686–690.
Amarel AR, Lovewell G, Coelho MM, Amato G, Rosenbaum HC. 2014. Hybrid speciation in a marine mammal: the
Clymene dolphin (Stenella clymene). PLoS ONE 9: e8645.
Arnold ML. 2006. Evolution through genetic exchange.
Oxford, UK: Oxford University Press.
Baldassarre DT, White TA, Karubian J, Webster MS.
2014. Genomic and morphological analysis of a semipermeable avian hybrid zone suggests asymmetrical introgression
of a sexual signal. Evolution 69: 2644–2657.
Behm JE, Ives AR, Boughman JW. 2010. Breakdown in
postmating isolation and the collapse of a species pair
through hybridization. American Naturalist 175: 11–26.
Boag PT. 1984. Growth and allometry of external morphology in Darwin’s Ground Finches (Geospiza) on Isla Daphne
Major, Gal
apagos. Journal of Zoology London 204: 413–
431.
Brelsford A, Mil
a B, Irwin DE. 2011. Hybrid origin of
Audubon’s Warbler. Molecular Ecology 20: 2380–2389.
Bronson CL, Grubb TC Jr, Sattler GD, Braun MH.
2005. Reproductive success across the black-capped chickadee (Poecile atricapillus) and Carolina chickadee (P. carolinensis) hybrid zone in Ohio. Auk 122: 759–772.
Curry RL. 2005. Hybridization in chickadees: much to learn
from familiar birds. Auk 122: 747–758.
Dobzhansky T. 1941. Genetics and the origin of species, 2nd
edn. New York: Columbia University Press.
Fitzpatrick BM. 2004. Rates of evolution of hybrid inviability in birds and mammals. Evolution 58: 1865–1870.
Garrick RC, Benavides E, Russello MA, Hyseni C,
Edwards DL, Gibbs JP, Tapia W, Ciofi C, Caccone A.
2014. Lineage fusion in Gal
apagos giant tortoises. Molecular Ecology 23: 5276–5290.
Gill FB. 2004. Blue-winged warblers (Vermivora pinus) versus Golden-winged warblers (V. chrysoptera). Auk 121:
1014–1018.
Grant PR. 1993. Hybridization of Darwin’s finches on Isla
Daphne Major, Gal
apagos. Philosophical Transactions of
the Royal Society B 340: 127–139.
Grant PR, Grant BR. 1992. Hybridization of bird species.
Science 256: 193–197.
Grant BR, Grant PR. 1993. Evolution of Darwin’s finches
caused by a rare climatic event. Proceedings of the Royal
Society B 251: 111–117.
Grant PR, Grant BR. 1994. Phenotypic and genetic effects
of hybridization in Darwin’s finches. Evolution 48: 297–
316.
Grant BR, Grant PR. 2006. Evolution of character displacement in Darwin’s finches. Science 313: 224–226.
Grant BR, Grant PR. 2008. Fission and fusion of Darwin’s
finches populations. Philosophical Transactions of the Royal
Society B 363: 2821–2829.
Grant PR, Grant BR. 2009. The secondary contact phase of
allopatric speciation. Proceedings of the National Academy of
Sciences of the United States of America 106: 20141–20148.
Grant PR, Grant BR. 2010. Conspecific versus heterospecific gene exchange between populations of Darwin’s finches.
Philosophical Transactions of the Royal Society B 365:
1065–1076.
Grant PR, Grant BR. 2014a. 40 Years of evolution. Darwin’s finches on Daphne Major Island. Princeton, NJ:
Princeton University Press.
Grant PR, Grant BR. 2014b. Synergism of natural selection
and introgression in the origin of a new species. American
Naturalist 183: 671–681.
Grant PR, Grant BR. 2015. Data From: Introgressive
hybridization and natural selection in Darwin’s finches.
Dryad Digital Repository. doi:10.5061/dryad.vc348.
Grant PR, Grant BR, Markert JA, Keller LF, Petren K.
2004. Convergent evolution of Darwin’s finches caused by
introgressive hybridization and selection. Evolution 58:
1588–1599.
Grant PR, Grant BR, Petren K. 2005. Hybridization in the
recent past. American Naturalist 166: 56–67.
Hasselman DJ, Argo EE, McBride MC, Bentzen P,
Schultz TF, Perez-Umphrey AA, Palkovacs EP. 2014.
Human disturbance causes the formation of a hybrid
swarm between two naturally sympatric fish species. Molecular Ecology 23: 1137–1152.
Hedrick P. 2013. Adaptive introgression in animals: examples and comparison to new mutation and standing variation as sources of adaptive variation. Molecular Ecology 22:
4606–4618.
Hermansen JS, Sæther AA, Elgvin TO, Borge T, Hjelle
E, Sætre G-P. 2011. Hybrid speciation in sparrows. I.
Phenotypic intermediacy, genetic admixture and barriers to
gene flow. Molecular Ecology 20: 3812–3822.
Irwin DE, Brelsford A, Toews DPL, MacDonald C,
Phinney M. 2009. Extensive hybridization in a contact
zone between MacGillivray’s and mourning warblers
(Oporornis tolmiei and O. philadelphia) detected using
molecular and morphometric analyses. Journal of Avian
Biology 40: 539–552.
Kim M, Cui M-L, Cubas P, Gillies A, Lea K, Chapman
MA, Abbott RJ, Coen E. 2008. Regulatory genes control a
key morphological and ecological trait transferred between
species. Science 322: 1116–1119.
Kleindorfer S, O’Connor JA, Dudaniec RY, Myers SA,
Robertson J, Sulloway FJ. 2014. Species collapse via
hybridization in Darwin’s tree finches on Floreana Island.
American Naturalist 183: 325–341.
en MS, Maqbool K,
Lamichhaney S, Berglund J, Alm
Grabherr M, Martinez-Barrio A, Promerov
a M, Rubin
C-J, Wang C, Zamani N, Grant BR, Grant PR, Webster
MT, Andersson L. 2015. Evolution of Darwin’s finches
and their beaks revealed by genome sequencing. Nature
518: 371–375.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
DARWIN’S FINCHES
Larsen PA, March
an-Rivadeneira MR, Baker RJ. 2010.
Natural hybridization generates mammalian lineage
with species characteristics. Proceedings of the National
Academy of Sciences of the United States of America 107:
11447–11452.
Lerner HRL, Andr
es JA, Bogdanowicz SM, Harrison
RG. 2013. Differential introgression in a mosaic hybrid
zone reveals candidate barrier genes. Evolution 67: 3653–
3661.
Liu KJ, Steinberg E, Yozzo A, Song Y, Kohn MH,
Nakhieh L. 2014. Interspecific introgressive origin of genomic diversity in the house mouse. Proceedings of the
National Academy of Sciences of the United States of America 112: 196–201.
Mallarino R, Grant PR, Grant BR, Herrel A, Kuo WP,
Abzhanov A. 2011. Two developmental models establish
3D beak-shape variation in Darwin’s finches. Proceedings of
the National Academy of Sciences of the United States of
America 198: 4057–4062.
Martinsen GD, Whitham TG, Turek RJ, Kiem P. 2001.
Hybrid populations selectively filter gene introgression.
Evolution 55: 1325–1335.
McDonald DB, Clay RP, Brumfield RT, Braun MJ. 2001.
Sexual selection on plumage and behavior in an avian
hybrid zone. Experimental tests of male–male interactions.
Evolution 55: 1443–1451.
Mettler RD, Spellman GM. 2009. A hybrid zone revisited:
molecular and morphological analysis of the maintenance,
movement, and evolution of a Great Plains avian (Cardinalidae: Pheucticus) hybrid zone. Molecular Ecology 18: 3256–
3267.
Mil
a B, Toews DPL, Smith TB, Wayne RK. 2011. A cryptic contact zone between divergent mitochondrial DNA lineages in southwestern North America supports past
introgressive hybridization in the yellow-rumped warbler
complex (Aves: Dendroica coronata). Biological Journal of
the Linnean Society 93: 696–703.
Parchman TL, Gompert Z, Braun MJ, Brumfield RT,
McDonald DB, Uy JAC, Zhang G, Jarvis ED,
Schlinger BA, Buerkle CA. 2013. The genomic consequences of adaptive divergence and reproductive isolation
between species of manakins. Molecular Ecology 22:
3304–3317.
Pereira RJ, Barreto FS, Burton RS. 2014. Ecological novelty by hybridization: experimental evidence for increased
thermal tolerance by transgressive segregation in Tigriopus
californicus. Evolution 68: 204–215.
Petren K. 1998. Microsatellite primers from Geospiza fortis
and cross-species amplification in Darwin’s finches. Molecular Ecology 7: 1782–1784.
Pfennig DW, Pfennig KS. 2012. Evolution’s wedge. Competition and the origins of diversity. Los Angeles, CA: University of California Press.
Poelstra JW, Vijay N, Bossu CM, Lantz H, Ryll B,
M€
uller L, Baglione V, Unneberg P, Wikelski M, Grabherr MG, Wolf JBW. 2014. The genomic landscape underlying phenotypic integrity in the face of gene flow in crows.
Science 344: 1410–1414.
821
Prager EM, Wilson AC. 1980. Phylogenetic relationships
and rates of evolution in birds. In Proceedings of
the XVII International Ornithological Congress. Berlin:
Verlag der Deutsche Ornithologon-Gesellschaft, pp. 1209–
1214.
Price TD. 2008. Speciation in birds. Greenwood, CO: Ben
Roberts.
Price TD, Bouvier MM. 2002. The evolution of F1 postzygotic incompatibilities in birds. Evolution 56: 2083–2089.
Pritchard JK, Stephens M, Donnelly P. 2000. Inferences
of population structure using multilocus genotype data.
Genetics 155: 945–959.
Pritchard JK, Wen X, Falush D. 2007. Documentation for
structure software: version 2.2. Pal Alto, CA: Stanford University.
Rheindt FE, Edwards SV. 2011. Genetic introgression: an
integral but neglected component of speciation. Auk 128:
620–632.
Rowher S, Bermingham E, Wood C. 2001. Plumage and
mitochondrial DNA haplotype variation across a moving
hybrid zone. Evolution 55: 405–422.
Ryan PG, Moloney CL, Hudon J. 1994. Color variation
and hybridization among Nesospiza buntings on Inaccessible Island, Tristan da Cunha. Auk 111: 314–327.
Schwarzbach AE, Rieseberg LH. 2002. Likely multiple
origins of a diploid hybrid sunflower species. Molecular
Ecology 11: 1703–1715.
Schwenk K, Brede N, Streit B. 2008. Introduction. Extent,
processes and evolutionary impact of interspecific
hybridization in animals. Philosophical Transactions of the
Royal Society B 363: 2805–2811.
Secondi J, Faivre B, Bensch S. 2006. Spreading introgression in the wake of a moving contact zone. Molecular Ecology 15: 2463–2475.
Seehausen O. 2004. Hybridisation and adaptive radiation.
Trends in Ecology and Evolution 19: 198–207.
Selz OM, Lucek K, Young KA, Seehausen O. 2014.
Relaxed trait covariance diversity in interspecific
cichlid hybrids predicts morphological diversity in
adaptive radiations. Journal of Evolutionary Biology 27:
11–24.
Taylor EB, Boughman JW, Groenenboom M, Sniatynski
M, Schluter D. 2006. Speciation in reverse: morphological
and genetic evidence of the collapse of a three-spined stickleback (Gasterosteus aculeatus) species pair. Molecular
Ecology 15: 343–355.
Taylor SA, Larson EL, Harrison RG. 2015. Hybrid zones:
windows on climate change. Trends in Ecology & Evolution
30: 398–406.
Toews DPL, Brelsford A, Irwin DE. 2011. Hybridization
between Townsend’s and black-throated green warblers in
an avian suture zone. Journal of Avian Biology 42: 434–
446.
Vonlanthen P, Bittner D, Hudson AG, Young KA,
M€
uller R, Lundsgaard-Hansen B, Roy D, Di Piazza S,
Largiarder CR, Seehausen O. 2012. Eutrophication
causes speciation reversal in whitefish adaptive radiations.
Nature 482: 357–362.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822
822
P. R. GRANT and B. R. GRANT
Webb WC, Marzluff JA, Omland KE. 2011. Random interbreeding between cryptic lineages of the common raven:
evidence for speciation in reverse. Molecular Ecology 20:
2390–2402.
Wiley C, Fogelberg N, Sæther SA, Veen T, Svedin N,
Vogel Kehlenbeck J, Qvarnstr€
om A. 2007. Direct benefits and costs for hybridizing Ficedula flycatchers. Journal
of Evolutionary Biology 20: 854–864.
SHARED DATA
Data deposited in the Dryad repository (Grant & Grant, 2015).
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 117, 812–822