The phylogeny of monkey beetles based on mitochondrial and

Molecular Phylogenetics and Evolution 60 (2011) 408–415
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
The phylogeny of monkey beetles based on mitochondrial and ribosomal RNA
genes (Coleoptera: Scarabaeidae: Hopliini)
Dirk Ahrens a,b,⇑, Michelle Scott a,c, Alfried P. Vogler a,c
a
Department of Entomology, Natural History Museum, London SW7 5BD, United Kingdom
Zoologisches Forschungsmuseum Alexander Koenig Bonn, Adenauerallee 160, 53113 Bonn, Germany
c
Department of Life Sciences, Silwood Park Campus, Imperial College London, Ascot SL7 5PY, United Kingdom
b
a r t i c l e
i n f o
Article history:
Received 27 December 2010
Revised 5 April 2011
Accepted 18 April 2011
Available online 6 May 2011
Keywords:
Scarabaeidae
Melolonthinae
Classification
Activity patterns
Sexual dimorphism
Species diversification
a b s t r a c t
Monkey beetles (Hopliini) are a large clade of flower and leaf feeding species within the Scarabaeidae
(chafers) with greatest diversity in southern Africa. Their internal relationships and sister group affinities
have not been studied with DNA methods. We used partial gene sequences for 28S rRNA, cytochrome oxidase I (cox1) and 16S rRNA (rrnL) for 158 species, representing most recognized subfamilies of Scarabaeidae, including 46 species of Hopliini. Combined analyses using maximum likelihood and Bayesian
inference under the two preferred alignment parameters recovered the Hopliini as monophyletic. Hopliines were inserted at the base of a clade of Cetoniinae + Rutelinae + Dynastinae, being either recovered as
their immediate sister group, or as part of an expanded set of basal branches that also includes the tribe
Macrodactylini which has been classified as part of the Melolonthinae (may chafers). At the level of subtribes, we found Hopliina paraphyletic with respect to Pachycnemina which also includes the monophyletic clade of Heterochelina and Gymnolomina. Trait mapping under parsimony on the preferred tree
resulted in inferences of three independent origins of sexual dimorphism, which coincided with shifts
to ‘flower-embedding’ pollination. In contrast, night active taxa, which are general phyllophages as other
pleurostict chafers, never show clear sexual dimorphism. South African lineages include several deepbranching lineages. The exceptional morphological and phylogenetic diversity of the South African fauna
may therefore be due to their antiquity, in addition to sexual selection in the day-active lineages. Phylogenetic studies of the endemic South African plant radiations have demonstrated the repeated evolutionary shift to beetle pollination, but it remains to be investigated if this is driven by the hopliine pollinators
present in the bioregion or by a propensity of the local plant lineages favoring this pollination syndrome.
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
Among the scarab beetles (Scarabaeidae), the tribe Hopliini, or
monkey beetles, are a diverse group of phytophages comprising
about 1200 species (Lacroix, 1998). They are widely distributed
throughout the temperate and tropical regions, but missing from
the Neotropics and Australia. Hopliines are particularly diverse in
Southern Africa, which harbors up to half of the world’s diversity
of species and genera, as well as showing great morphological
diversity, while levels of endemism are also very high in this region
(Colville et al., 2002; Colville, 2006). Several species are generalist
herbivores, while others exhibit close associations with flowers
as mating sites and as a source of pollen on which most Hopliini
feed as adults (Steiner, 1998; Colville et al., 2002; Krenn et al.,
⇑ Corresponding author at: Zoologisches Forschungsmuseum Alexander Koenig
Bonn, Adenauerallee 160, 53113 Bonn, Germany. Fax: +49 228 9122212.
E-mail addresses: [email protected], [email protected] (D. Ahrens).
1055-7903/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2011.04.011
2005; Carrillo-Ruiz and Morón, 2006). Due to this life style they
are important pollinators, and several groups of plants have developed specialized features to attract these beetles (Steiner, 1999;
Colville et al., 2002; Goldblatt et al., 2005; Goldblatt and Manning,
2006; van Kleunen et al., 2007). Specifically, they have driven the
evolution of particular flower types in several genera of Iridaceae:
Hopliine-pollinated flowers are often very different in pigmentation pattern, tepal size and orientation, and floral symmetry from
their closest relatives pollinated by bees, moths or hummingbirds
(Goldblatt and Manning, 2006). These beetle pollinated flowers
are typically radially symmetrical, bowl-shaped, and relatively
large, and are thus able to accommodate two or more beetles. They
often show ‘beetle marks’, which are prominent, contrast-rich
markings on the vividly pigmented floral tepals (van Kleunen
et al., 2007). In addition, pheromones contribute to species-specific
aggregation and mate recognition in hopliine beetles (Zhang et al.,
2003).
The unique features of the hopliines raise the question about
their evolutionary origin. Without doubt they are members of
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
the ‘pleurostict’ scarab beetles, a huge radiation of almost exclusively herbivorous species, of which only a small proportion are
flower feeding. However, the precise relationships of Hopliini with
other major scarab lineages, and hence the first appearance of this
specialized life style, remain unclear. Some authors considered
Hopliini as a lineage that is phylogenetically separated from the
main scarab clades of ‘rutelines’ and ‘melolonthines’ based on
striking diagnostic features, specifically the single metatarsal claw
that distinguishes them from all related groups (e.g. Mulsant,
1842; Reitter, 1902; Peringuey, 1902; Janssens, 1949; Medvedev,
1976; Iablokoff-Khnzorian, 1977). Others considered Hopliini as
part of the Melolonthinae, which has also been referred to as Melolonthides (Lacordaire, 1856), Melolonthini (Gemminger and von
Harold, 1869), Melolonthinae (e.g. Dalla-Torre, 1913; Evans,
2003; Smith, 2006), or Melolonthidae (Bates, 1888; Balthasar,
1963). Yet others placed the Hopliini within ‘Rutelidae’ (=Rutelinae) (Burmeister, 1844; Mulsant and Rey, 1871; Paulian, 1959;
Paulian and Baraud, 1982). Alternatively, Lacroix (1998) considered Hopliini as an independent group at the family level, but he
did not comment on its relationships with other groups of Scarabaeoidea. The discussion about the position of Hopliini among either
rutelines or melolonthines focused on the retractable metatarsal
claws found in both Hopliini and Rutelinae (Burmeister, 1844).
Only Hopliini and some species of the distinct tribe Macrodactylini
have a single claw, as opposed to two claws in all other groups
(Katovic, 2008). The close relationship of the latter with Hopliini
was also suggested in a study using various morphological characters (Carrillo-Ruiz and Morón, 2006). However, recent studies have
shown that establishing the systematic position of Hopliini is complicated because the Melolonthinae, the largest group of pleurostict scarabs, has been found to be paraphyletic (Ahrens, 2006;
Ahrens and Vogler, 2008). Consequently, inferring the position of
Hopliini will require the resolution of basal relationships of pleurostict chafers more widely, based on broad taxon sampling beyond what is currently available (e.g. Browne and Scholtz, 1998;
Hunt et al., 2007).
In addition, the internal relationships of Hopliini are of interest
for studying what drives the morphological, behavioral and ecological diversity in this lineage (Midgeley, 1992; Colville et al., 2002;
Colville, 2006). In particular, a phylogenetic framework is needed
to explore the evolution of major morphological tendencies and
life traits (e.g. feeding behavior, diel activity pattern, sexual dimorphism) and the underlying processes of trait selection (Midgeley,
1992) and niche partitioning (Picker and Midgeley, 1996). As a
framework for investigating the evolution of species richness in
the Hopliini and their mutual relations with flowering plants we
study the phylogenetic relationships of the tribe using DNA sequences from one nuclear and two mitochondrial genes. We are
specifically interested in the extent to which the pollinator guilds,
as defined by Picker and Midgeley (1996), are phylogenetically related to lineages with different life styles and how the origin of
flower feeding might be correlated with the occurrence of sexual
dimorphism as a possible mechanism for increased diversification
rates via sexual selection.
2. Material and methods
2.1. Taxon sampling, DNA extraction, and DNA sequencing
A representative sample of 46 species was collected from various biogeographic regions with focus on South Africa and Madagascar representing almost 20% of the 108 genera of Hopliini.
Together, these regions harbor almost 70% of species and >80% of
genera of hopliines in the world. Missing genera from elsewhere
are likely to be phylogenetically close to those included in the
409
study, e. g. Palearctic species not included resemble the large genus
Hoplia. Other pleurostict chafers included here represent all principal lineages (subfamilies) of Scarabaeidae, plus two species each of
Hybosoridae and Glaphyridae, largely following Ahrens and Vogler
(2008) but with the number of Sericini reduced (Supplementary
Table S1). All trees were rooted with Hybosorus illigeri. DNA was
extracted from thoracic leg muscle tissue using Promega WizardSV
extraction plates. Following DNA extraction, beetles were dry
mounted. Vouchers will be deposited at ZFMK Bonn.
Mitochondrial gene regions included cytochrome oxidase subunit 1 (cox1) and 16S ribosomal RNA (rrnL) with the adjacent regions NAD dehydrogenase subunit 1 (nad1) and tRNA leucine
(trnLeu), hereafter referred to as rrnL. PCR and sequencing was performed using primers Pat and Jerry for cox1 and 16Sar and ND1A
for rrnL (Simon et al., 1994). For a few specimens only a shorter
fragment could be obtained using the primer pair 16Sar and
16SB2 (Simon et al., 1994). Nuclear 28S rRNA containing the variable domains D3–D6 was amplified using primers FF and DD
(Monaghan et al., 2007). Sequencing was performed on both
strands using BigDye v. 2.1 and an ABI3730 automated sequencer.
Sequences were edited manually using Sequencher v. 4.8 (Genecodes Corp., Ann Arbor, MI, USA). GenBank accession numbers
are given Supplementary Table S1.
2.2. Phylogenetic analysis using multiple alignment
Evolutionary dynamics of mitochondrial and nuclear rRNA
genes differ (e.g. Vogler and Pearson, 1996; Terry and Whiting,
2005), and hence different alignment parameters and algorithms
were applied to various partitions. As we used the new data for
Hopliini in the context of an existing taxon sampling of pleurostict
scarabs of Ahrens and Vogler (2008), the best parameters
established in that study were applied here. The choice of combination of aligned gene matrices was made according to the
preferred combinations established in Ahrens and Vogler (2008)
based on congruence of phylogenetic signal in a hundred
combinations of different primary alignments made from various
alignment procedures and parameters with various software package. Here we only used the two best performing combinations
using MAFFT (vers. 5.8; Katoh et al., 2002, 2005) and MUSCLE vers.
3.6 (Edgar, 2004) for rrnL (both with default parameters), and
Clustal X (Thompson et al., 1997) for 28S (optimized parameters
gap opening cost 10 and gap extension cost 6.66). These alignments were combined for phylogenetic analysis together with
cox1, referred to as AL1 (cox1 + rrnLMAFFT + 28SClustalX) and AL2
(cox1 + rrnLMuscle + 28SClustalX).
The concatenated data matrices were subjected to parsimony
tree searches using TNT (Goloboff et al., 2004) with 10 ratchet iterations, 10 cycles of tree drifting and three rounds of tree fusing for
each of 200 random addition sequences. Gaps were treated as a
fifth character state. Model-based phylogenetic analyses under
Maximum Likelihood (ML) were performed in PhyML (Guidon
and Gascuel, 2003) using a GTR model (as selected by Modeltest
using AIC; Posada and Crandall, 1998; Akaike, 1974) with all
parameters estimated from the data and four substitution rate categories. Bayesian analysis was conducted using parallel MrBayes
3.2 (Huelsenbeck and Ronquist, 2001), conducting MCMC runs
(Yang and Rannala, 1997) after partitioning (Nylander et al.,
2004; Brandley et al., 2005) the data for rrnL, 28S, and three codon
position of cox1. Based on the preferred alignment combination, we
used a GTR + I + G model as determined by Modeltest with five partitions referring to rrnL, 28S, and the first, second and third codon
postion of cox1. Tracer 1.3 (Rambaut and Drummond, 2003) was
used to graphically determine stationarity and convergence of
runs. Tree searches were conducted for 13 and 25 106 generations, using a random starting tree and two runs of three heated
410
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
and one cold Markov chains (heating of 0.1). Chains were sampled
every 1000 generations and 106 generations were discarded as
burn-in based on the average standard deviation of split frequencies as well as by plotting ln L against generation time. Bayes factors (Kass and Raftery, 1995) were calculated for both the MAFFT
and the MUSCLE alignments in order to determine significant differences in the marginal likelihood of initial unconstrained compared to analyses using a constraint topology for the monophyly
of Hopliini with a single metatarsal claw.
3. Results
Sequences (n = 158) for three loci (cox1, rrnL, 28S) were aligned
and concatenated to result in two alignments produced with
MAFFT (AL1) and MUSCLE (AL2), to produce matrices of 2491
and 2487 nucleotide positions, respectively (Table 1). Parsimony
ratchet searches on the combined matrices using gaps as 5th character state returned two overall shortest trees of 21,621 and 22,037
steps for alignment AL1 and AL2. These analyses performed poorly
with regard to recovery and position of well established monophyletic groups (Supplementary Fig. S3), for example placing the dung
beetles (Scarabaeinae) among the pleurostict clades under AL2,
and the macrodactyline genus Plectris nested within the Hopliini
in the AL1 alignment. Only the AL2 alignment recovered the Hopliini as monophyletic in the strict consensus, as sister to a clade that
comprised Rutelinae and Dynastinae. The poor performance of parsimony was not improved when gaps were treated as missing data,
resulting in two best trees of 20,625 steps (AL2) or in four best
trees of 20,597 steps (AL1) (see Supplementary Fig. S3). While in
AL2 dung beetles (Scarabaeinae) were nested again within pleurosticts as sister to the monophyletic Hopliini (Supplementary
Fig. S3), the strict consensus tree of the AL1 was only poorly resolved but recovered most major clades including the Hopliini as
monophyletic (tree not shown).
Model based analyses performed with PhyML and MrBayes
separating functional data partitions produced trees that were similar for AL1 and AL2 (Tables 1 and 2). All analyses recovered the
Hopliini as monophyletic (Fig. 1, node 1), and while posterior
probabilities for this node was always 1 in the Bayesian analysis,
bootstrap support for the Hopliini was low under ML (18%). The
tree (Fig. 1) also provides an update on the phylogeny of pleurostict scarabs. Most of the major clades including the principal scarab
lineages as well as Hopliini, however, seem to be particularly well
supported mainly by the mtDNA since they did not appear monophyletic in model based searches (PhyML, MrBayes) on the 28S
Table 1
Comparison of alignments and tree scores under various tree search methods. All
likelihood and Bayesian analyses were conducted with the GTR + I + G model. The last
row shows the results from a Bayesian analysis constrained for monophyly of taxa
with a single metatarsal claw.
a
AL1
AL2
Alignment lenth
cox1
rrnL
28S rRNA
826 bp
884 bp
781 bp
826 bp
880 bp
781 bp
Informative characters
cox1
rrnL
28S rRNA
420
451 (486)a
84 (100)
420
456 (477)a
84 (100)
Tree scores
TNT (tree length)
PhyML
MrBayes
MrBayes constr.
21,621
81909.37
78161.03
78166.27
22,037
81858.39
78033.34
78039.82
Values in parentheses when gaps were treated as 5th character.
Table 2
Comparison of tree topologies from the different analyses regarding the phylogenetic
resolution for major clades. Topologies differing from the preferred tree are
highlighted bold (M – monophyletic, P – paraphyletic; – under inclusion of
Diphycerini; ? – uncertain polytomy; § – Cetoniinae + Rutelinae [including ‘Dynastinae’]; § – including the tribes Automolini, Liparetrini, Scitalini, Maechidiini,
Heteronycini).
Pleurosticti
Aph + Scar
Glaphyridae
Hop + Macr
Hop + Plectris
Dyn + Ano
Dyn + Ado
Ser + Abl
Cet + Rut§
Aphodiinae
Dynastinae
Cetoniinae
Melolonthinae
Rutelinae
Scarabaeinae
Adoretini
Anomalini
Cetoniini
Diplotaxini
Enariini
Geniatini
Hopliini
Macrodactylini
Melolonthini
Rhizotrogini
Rutelini
Sericini
SW Melolonthines$
MrBayes
PhyML
AL1
AL2
AL1
AL2
AL1
Parsimony
AL2
AL2 (5th = ?)
M
M
M
–
M
M
–
M
M
M
M
M
P
P
M
M
M
M
M
M
M
M
P
M
P
M
M
M
M
M
M
–
M
M
–
M
M
M
M
M
P
P
M
M
M
M
M
M
M
M
P
M
P?
M
M
M
M
M
M
–
–
M
–
–
–
M
M
P
P
P
M
M
M
M
M
M
M
M
M
M
P
M
M
M
M
M
M
–
–
–
–
–
–
M
M
P
P
P
M
M
M
M
M
M
M
M
M
M
P
M
P
M
M
M
M
–
–
–
M
–
–
M
P
P
P
P
M
M
M
M
M
M
M
P
P
M
P
P
P
M
–
M
M
–
–
–
–
–
–
M
P
P
P
P
M
M
M
M
P
M
M
M
M
M
P
P
P
M
–
M
M
–
–
M
–
–
–
M
M
P
P
P
M
M
M
M
M
M
M
M
P
M
P
P
P
M
rRNA gene alone (Supplementary Fig. S4). At the basal node, a
group of Southern Hemisphere melolonthines (node 5 in Fig. 1)
was found sister to all other Pleurosticti. This group had been recognized in Ahrens and Vogler (2008) as a deeply separated lineage,
as sister to Sericini + Ablaberini (node 6). The Macrodactylini
(nodes 2 and 3), the possible sister group of Hopliini (Carrillo-Ruiz
and Morón, 2006), was not monophyletic. In the Bayesian analysis
the three Isonychus species (node 3 in Fig. 1) were widely separated
from Plectris, the fourth representative of this tribe in this study. In
both PhyML trees Macrodactylini were monophyletic but only under inclusion of the genus Diphycerus which has been treated as
separate tribe since the work of Medvedev (1952) (see Smith,
2006). However, in all analyses both lineages of Macrodactylini
were closely associated to the cetonine–ruteline–dynastine clade
(node 4). Several rutelines and dynastines added here resulted in
a placement of the monophyletic Dynastinae either as sister to
Anomalini in the Bayesian analysis, or as sister to Adoretini + Anomalini (AL2) or to Adoretini (AL1) with PhyML (Supplementary Fig. S1). The Rutelini + Geniatini always diverged at the base of
the dynastine–ruteline clade. The South African genus Isoplia
which strongly resembles the Hopliini in external appearance, in
particular the hairy species of Eriesthis, was always recovered as
sister of Anomalini, albeit with low branch support (Fig. 1 and Supplementary Fig. S1). Hence, these similarities are due to convergence, as initially proposed by Peringuey (1902), who treated this
group as a separate tribe (Isopliini) of the Rutelinae.
The critical question about the sister group of Hopliini remained
incompletely resolved. The Bayesian analyses identified the
macrodactyline genus Plectris (node 2) as sister to the Hopliini,
but the relevance of this finding is unclear because of the nonmonophyly of the Macrodactylini in these trees. Under ML, in the
AL1 tree Macrodactylini + Ablaberini were sister to the Hopliini
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
411
Fig. 1. Majority-rule consensus-tree of the Pleurosticti from Bayesian analysis of the AL1 alignment. Posterior probabilities are given below the branches. Key nodes discussed
in the text are marked by numbers 1–7.
(Supplementary Fig. S1) while in the AL2 tree Macrodactylini was
again sister (under inclusion of Pachydemini, i.e. the genus Buettikeria) to the cetonine–ruteline–dynastine clade which together
were the sister of Hopliini.
Moving now to the relationships within Hopliini (Fig. 2), the
tree topology was very similar among various model-based
searches. We found the Hopliina, one of the currently recognized
subtribes (Smith, 2006), paraphyletic with respect to the Pachycnemina (clade A in Fig. 2), as well other clades such as clade B
(corresponding to Heterochelina Burmeister, 1844) and clade C,
corresponding to the genus Gymnoloma (=Gymnolomina Burmeister, 1844). The latter two groups have already been synonymized
with Hopliina (Smith, 2006). In all model-based analyses Hoplia
was monophyletic, as sister to a clade of night active species from
Madagascar (Amorphochelus, Michaeloplia) and South Africa (Congella, Microplus). Only five taxa changed positions slightly in the
two Bayesian searches (Fig. 2). These were Scelophysa and Echyra
which switched their positions close to the ‘Clania’-Eriesthis and
the Hoplia-Congella clade. In AL1 Eriesthis was paraphyletic in the
MrBayes and ML trees since Pachycnema and Lepithrix were nested
between Eriesthis cf. rhodesiana and all other Eriesthis species,
while in AL2 Eriesthis was monophyletic in a polytomy with
Pachycnema and Lepithrix (Fig. 2, node A).
3.1. Evolutionary trends and biogeography
The analysis established a phylogenetic framework for investigating the evolution of Hopliini and its interaction with flowering
plants. We found that, although the monophyly of Hopliini was
well supported, the tree topology was not consistent with a potential key synapomorphy, the single metatarsal claw, contradicting
Carrillo-Ruiz and Morón (2006) (see Fig. 2). According to their phylogenetic hypothesis, a group of ‘Hopliini’ with two metatarsal
claws, which were not included in their study, were expected to
be the sister of all single-clawed Hopliini. This concerns only two
hopliine genera, Cylichnus Burmeister (1844) and Nanaga Peringuey (1902) that have been classified as Hopliini by Peringuey
(1902). In our analysis the double-clawed genus Cylinchus was
nested deeply within the single-clawed clade and never branched
off at the base. Therefore, we performed the Bayesian MCMC constrained for the monophyly of hopliine taxa with single metatarsal
claw and found the Bayes factors (B10) 1 for AL1 and AL2 (see
412
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
Fig. 2. Majority-rule consensus trees of Hopliini from Bayesian inference for AL1 (left panel) and AL2 (right). Posterior probabilities are given below the branches. Gray and
dark shading indicates taxa with moderate and severe sexual dimorphism, respectively. Branch color on tree on the left indicates one (red) or two (black) two metatarsal
claws. Branch color on tree on left indicates day active (black) and night active (blue) species. Colored dots refer to various zoogeographic regions: red, southern Africa; blue,
Madagascar; yellow, Asia; white, Europe. Arrows mark the taxa whose position differ in both trees. A, B, C refer to clades of Pachycnemina (A), Heterochelina (B) and
Gymnolomina (C). The habitus images from top-left to right bottom side: Scelophysa endroedyyoungai, Congella tesselatula, Echyra umbrina, Clania macgregori, Eriesthis
hypocrita, Paramorphochelus agricola, Ceratoscelis spSA2, Heterochelus vulpinus, Hoplia farinosa, Cylinchus vanus, Pachycnema calviniana, Lepithrix cf. forsteri, Omocrates sp.,
Gymnoloma nigra. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1), i.e. the present data would not reject the hypothesis that
the single metatarsal claw had a single origin. Reversals to a second
claw may not be implausible because rudiments of it are evident in
many species of Hoplia, indicating a labile trait.
Further morphological and ecological features were mapped on
the preferred tree, to assess their role in clade diversification and in
the evolution of the beetle-flower mutualism. Many taxa exhibit
sexual dimorphism (gray shading in the left tree, Fig. 2), whereby
males exhibit strongly developed hind legs and claws (Scelophysina, Pachycnemina) or bear strong spines on the hind legs and differ
in the color of the pygidium (often the only part of the beetle that
is visible when embedded in the flower bottom) from females
(Heterochelina). Based on inferred character changes on the tree,
sexual dimorphism has originated three times given the present
taxon sampling, although only two nodes defining these transitions are strongly supported. The dimorphism is only found in
day active flower visiting taxa (black branches in tree on the right
of Fig. 2), and specifically only among ‘embedding’ pollinators that
persist on a single flower large enough to harbor multiple individuals. In contrast, night active taxa, all of which are general phyllophages, never show clear sexual dimorphism. Due to the poorly
supported basal relationships of Hopliini the ancestral state for diel
activity pattern (day or night active) remains unresolved. Interestingly, sexual dimorphism is found predominantly in South African
lineages although flower feeding may occur also in European and
Asian Hopliini, while night activity occurs in lineages from Asia,
South Africa and Madagascar.
4. Discussion
Using the dataset of Ahrens and Vogler (2008) as a backbone,
and applying the same alignment and tree search procedures
found to be preferable, we now modified the sampling scheme to
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
focus on the systematic position of Hopliini within the major herbivore scarab lineages. Taxon sampling of Hopliini was incomplete
but likely included a large proportion of major clades, as the basis
for a framework of phylogenetic relationships and major evolutionary trends within the tribe. The strategy for alignment was
based on the notion that homology at the nucleotide level can be
established, first, relative to a phylogenetic tree that recovers
‘known’ groups whose monophyly is not in doubt based on existing
knowledge and, second, using an overall measure of character
homology at the nucleotide level (taxonomic congruence and character congruence sensu Wheeler, 1995). Ahrens and Vogler (2008)
combined both measures of congruence into a single value to select the best alignment. All alignments generated in this study
were based on progressive methods of pairwise alignment following a guide tree, implemented initially in the Clustal software and
in more recent methods that use an additional refinement step.
The latter would be expected to improve homology statements
over the simple Clustal procedure by realignment of several closely
related sequences from the initial pairwise alignment, but this was
not the case throughout, as apparently the theoretical improvements of the refinement do not necessarily provide an increase
of homology across the entire data set, given the criteria used (e.
g. the ensemble RI as a measure of total synapomorphy). This
seems to be confirmed by recent studies showing that no single
method was capable of producing high quality alignments for all
types of data (Aniba et al., 2010). Here we accepted the preferred
alignment parameters of the earlier study under the assumption
that in the current data set new taxa are added to existing deep
branches already present on the tree. Hence, by adding these taxa,
homology assignment at the nucleotide level would mainly affect
the sampling density of indels (and nucleotide changes) on the
tree, but their variation is the result of the same mutational steps
and therefore can be approximated by the same alignment parameters. While specific tests might be required to establish this more
firmly, we accepted the preferred settings from the earlier study on
the higher hierarchical level, not least to maintain the basal relationships found there (i.e. the outgroups for the Hopliini).
The test of homology was based on parsimony analysis, and unlike the current implementations of likelihood and Bayesian analyses, parsimony searches permit the use of indel variation for
tree building by coding ‘gaps’ as a fifth character state. Parsimony
analysis therefore is an important heuristic tool for assessing the
data set. However, phylogenetic relationships inferred with this
method were unsatisfactory in particular for deep nodes. Both
types of model-based analyses performed better in this regard,
indicating that the complex character variation is better described
by these models. Problems with the tree are therefore mainly due
to the difficulties with character reconstruction in the parsimony
analysis, in particular affecting basal relationships, rather than
the alignment. However, although the results from an unpartitioned likelihood analysis and a partitioned Bayesian analysis were
largely consistent, both using the GTR model, differences among
the two alignments still show the dominant effect of alignment
choice on the tree topology and the difficulty of selecting one alignment over another.
With regard to the taxonomic conclusions, we confirm the
consistent recovery of a monophyletic cetonine–ruteline–dynastine clade (node 4 in Fig. 1), and the association of Macrodactylini
and Hopliini with that clade. This is broadly congruent with the
topology proposed by Carrillo-Ruiz and Morón (2006) based on
a morphological character set. Whereas the evidence for the
monophyly of Macrodactylini and their phylogenetic position relative to other major groups including the Hopliini remains somewhat unclear, these findings resolve the reasons for the
controversy about the placement of Hopliini in the past. In particular, the discussion about their position among either rutelines or
413
melolonthines focused on the retractable metatarsal claws found
in both Hopliini and Rutelinae (Burmeister, 1844) that are present also in some Macrodactylini (Katovic, 2008). Given Burmeister’s (1844) classification grouping Hopliini within Rutelinae,
this character might be a synapomorphy for the Rutelinae plus
Hopliini, if the single-clawed genus Cylichnus represents the
ancestral state for the Hoplini. The inferred trees (Fig. 1, Supplementary Fig. S1) contradict this possibility, although a topology
required under this scenario that places Cylichnus as sister to all
other Hopliini could not be excluded with high statistical support.
Metatarsal claws are of high adaptive value for clinging to leaves
and other plant parts (see Fig. 1) and seem to be subject to selection for reduction and asymmetry of both claws, leading to convergent evolution. The evolutionary plasticity of this trait is also
evident in some species of Hoplia that only show a trace of a second metatarsal claw, with some species even showing intraspecific polymorphism.
Although the functional significance of the reduction or elimination of the second metatarsal claw is unclear, these changes
are found not only in the obligatory pollen feeders. The latter show
some apparent adaptations for pollination shared among all dayactive and floricolous species. These include dense and long pilosity of the body that makes hopliine beetles more effective pollen
carrier even than bees (Mayer et al., 2006). In addition, hopliines
often exhibit striking coloration conferred by setae that have been
transformed to colored scales. These are likely to be relevant in
mate recognition and also produce mimetic patterns, e.g. to resemble bumblebees. Finally, flower visiting species show great diversity in the morphology of mouth parts (Peringuey, 1902; Nel and
Scholtz, 1990; Carrillo-Ruiz and Morón, 2006), including the presence of either toothed maxilla or untoothed to membranous maxilla. These appear to be correlated with functional aspects of the
feeding on pollen vs. nectar, respectively, and may be important
for partitioning the food resources. Detailed analysis of various
character states and homology assignment of various mouth parts
are required to corroborate the link of mouthpart morphology and
feeding choice.
It is well known that morphological diversity of Hopliini is
greatest in the South African lineages (Peringuey, 1902). This might
be due to a greater lineage age of the flower dwelling hopliines in
South Africa, or to evolutionary pressures that differ from other regions in the world. The latter hypothesis would predict that South
African lineages are more diverse than their sister taxa elsewhere,
but this is not the case for at least one clade composed of herbivorous South African (Congella) and Malagasy taxa that is the sister
to the species rich genus Hoplia (Fig. 2). On the contrary, the South
African fauna appears to be a composite of several basal lineages
whose greater clade age results in higher species richness overall
compared to other regions. However, the support at basal nodes
is weak and the reconstruction of ancestral geographic areas
(South Africa vs. Asia/Holarctic) remains uncertain (Fig. 2). In addition, lineages associated with flower feeding may not have exhibited this life style ancestrally, as it requires co-radiation with
plants. Studies of the evolution of pollinator syndromes in Iridaceae (genus Babiana; Schnitzler et al., in press) demonstrated that
hopliine beetle pollination has arisen multiple times as a recently
derived trait. Moreover, host data of Babiana species (Goldblatt
and Manning, 2007) indicate that various lineages have recruited
pollinators independently from different genera, arguing against
ancestral relationships with pollinators or co-evolution. Vice versa,
closely related lineages of pollinators are associated with several
Babiana species, suggesting that these lineages of hopliines existed
already before a shift to feeding on certain flowers. The age of
lineages therefore seems to be less crucial for the flower-beetle
mutualism and possible effects on diversification rate. It remains
to be investigated if the repeated shifts to beetle pollination is
414
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
driven by the pollinator environment or due to traits of the endemic plant radiations in the South African bioregion.
On the other hand, given multiple origins of strong sexual
dimorphism in lineages of Hopliini that are all very diverse, it is
likely that traits involving sexual selection might have evolved
independently from a specific beetle-flower mutualism. Already a
shift to unspecific pollen feeding on flowers would be sufficient
to drive the evolution of sexual dimorphism due to sexual selection, since the food resource (florescence) also becomes a site for
species aggregation and mating where males compete for females
(Midgeley, 1992). This scenario is consistent with the fact that
many pollen feeding monkey beetles are encountered on a great
variety of host plants. Which of the potential drivers of diversification in South African monkey beetles is the most crucial, needs to
be investigated with a denser taxon and geographical sampling
comparing the influence of sexual selection vs. geographical speciation on the diversification rates.
Acknowledgments
We are grateful to Silvia Fabrizi for help with the laboratory
analysis; to the A. Vogler labgroup for helpful discussions; to Gerhard Burmeister, Benjamin Isambert, Michael Kuhlmann, David
Lees, Sergej Murzin, and Stefan Schmidt, for collecting specimens;
and to James Harrison for help with access to literature. We thank
the anonymous referees who helped to improve the final version of
the manuscript. Laboratory analysis at the NHM was partly supported by NHM grant-in-aid. DA was supported by the German Science Association (DFG/AH175/1-2 and AH/175/2-1).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ympev.2011.04.011.
References
Ahrens, D., 2006. The phylogeny of Sericini and their position within the
Scarabaeidae based on morphological characters (Coleoptera: Scarabaeidae).
Syst. Entomol. 31, 113–144.
Ahrens, D., Vogler, A.P., 2008. Towards the phylogeny of chafers (Sericini): analysis
of alignment-variable sequences and the evolution of segment numbers in the
antennal club. Mol. Phylogenet. Evol. 47, 783–798.
Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans.
Autom. Contr. 19, 716–723.
Aniba, M.R., Poch, O., Thompson, J.D., 2010. Issues in bioinformatics benchmarking:
the case study of multiple sequence alignment. Nucl. Acids Res. 38, 7353–7363.
Balthasar, V., 1963. Monographie der Scarabaeidae und Aphodiidae der
palaearktischen und orientalischen Region. Coleoptera: Lamellicornia. Band 1.
Prag, Verlag der Tschechoslowakischen Akademie der Wissenschaften. 391pp.
Bates, H.W., 1888. Biologia Centralia Americana, Insecta, Coleoptera, Part 2, vol. II.
Taylor and Francis, London, 432 pp.
Brandley, M.C., Schmitz, A., Reeder, T.W., 2005. Partitioned Bayesian analyses,
partition choice, and the phylogenetic relationships of scincid lizards. Syst. Biol.
54, 373–390.
Browne, J., Scholtz, C.H., 1998. Evolution of the scarab hind wing articulation and
wing base: a contribution toward the phylogeny of the Scarabaeidae
(Scarabaeoidea: Coleoptera). Syst. Entomol. 23, 307–326.
Burmeister, H.C.C., 1844. Handbuch der Entomologie, pt. 1, vol. 4. (Coleoptera,
Lamellicornia, Anthobia et Phyllophaga Systellochela). Berlin. 780 pp.
Carrillo-Ruiz, H., Morón, M.Á., 2006. Study on the phylogenetic relationships of the
hoplids (Coleoptera: Scarabaeoidea). Proc. Entomol. Soc. Washington 108, 619–
638.
Colville, J., 2006. A Profile of the Insects of the Kamiesburg Uplands. Naquamaland,
South Africa.
Colville, J., Picker, M.D., Cowling, R.M., 2002. Species turnover of monkey beetles
(Scarabaeidae: Hopliini) along environmental and disturbance gradients in the
Namaqualand region of the succulent Karoo, South Africa. Biodiv. Conserv. 11,
243–264.
Dalla-Torre, K.W., 1913. Coleopterorum Catalogus, Scarabaeidae, Melolonthinae
(IV), vol. XX, pars 45. 450 pp.
Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and
high throughput. Nucl. Acids Res. 32, 1792–1797.
Evans, A.V., 2003. A checklist of the New World chafers (Coleoptera: Scarabaeidae:
Melolonthinae). Zootaxa 211, 1–458.
Gemminger, M., von Harold, E., 1869. Catalogus Coleopterorum, hucusque
descriptorum synonymicus et systematicus, vol. 3. In: Gummi, E.H. (Ed.),
Scarabaeidae. Monachii, pp. 753–958.
Goldblatt, P., Bernhardt, P., Manning, J.C., 2005. Pollination mechanisms in the
African genus Moraea (Iridaceae, Iridoideae): floral divergence and adaptation
for pollinators. Adansonia Ser. 3 (27), 21–46.
Goldblatt, P., Manning, J.C., 2006. Radiation of pollination systems in the Iridaceae of
sub-Saharan Africa. Annu. Bot. 97, 317–344.
Goldblatt, P., Manning, J.C., 2007. Floral biology of Babiana (Iridaceae: Crocoideae):
adaptive floral radiation and pollination. Annu. Miss. Bot. Gard. 94, 709–733.
Goloboff, P., Farris, S., Nixon, K., 2004. TNT (tree analysis using new technology).
Cladistics 20, 84.
Guidon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to estimate
large phylogenies by maximum likelihood. Syst. Biol. 52, 696–704.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny.
Bioinformatics 17, 754–755.
Hunt, T., Bergsten, J., Levkanicova, Z., Papadopoulou, A., John, O.S., Wild, R.,
Hammond, P.M., Ahrens, D., Balke, M., Caterino, M.S., Gómez-Zurita, J., Ribera, I.,
Barraclough, T.G., Bocakova, M., Bocak, L., Vogler, A.P., 2007. A comprehensive
phylogeny of beetles reveals the evolutionary origins of a superradiation.
Science 318, 1913–1916.
Iablokoff-Khnzorian, S.M., 1977. Uber die Phylogenie der Lamellicornia. Ent. Abh.
Staatl. Mus. Tierk. 41, 135–199.
Janssens, A., 1949. Table synoptique et essai de classification pratique des
Coléopte‘res Scarabaeidae. Bull. Inst. Roy. Sci. Nat. 25 (15), 1–30.
Kass, R.E., Raftery, A.E., 1995. Bayes factors. J. Am. Stat. Assoc. 90, 773–795.
Katoh, K., Kuma, K., Toh, H., Miyata, T., 2005. MAFFT version 5: improvement in
accuracy of multiple sequence alignment. Nucl. Acids Res. 33, 511–518.
Katoh, K., Misawa, K., Kuma, K., Miyata, T., 2002. MAFFT: a novel method for rapid
multiple sequence alignment based on fast Fourier transform. Nucl. Acids Res.
30, 3059–3066.
Katovich, K., 2008. A generic-level phylogenetic review of the Macrodactylini
(Coleoptera: Scarabaeidae: Melolonthinae). Insecta Mundi 23, 1–78.
Krenn, H.W., Plant, J.D., Szucsich, N.U., 2005. Mouthparts of flower-visiting insects.
Arthropod Struct. Dev. 34, 1–40.
Lacordaire, T., 1856. Histoire Naturelle des Insectes. Librairie Encyclopedique de
Roret, Paris, 594 pp.
Lacroix, M., 1998. Insectes Coléoptères Hopliidae (2e partie). Faune Madagascar 88,
401–755.
Mayer, C., Soka, G., Picker, M., 2006. The importance of monkey beetle
(Scarabaeidae: Hopliini) pollination for Aioaceae and Asteraceae in grazed
and ungrazed areas at Paulshoek, Succulent Karoo, South Africa. J. Ins. Cons. 10,
323–333.
Medvedev, S.I., 1952. Plastinchatousye (Scarabaeidae), podsem. Melolonthinae, 2.
Fauna SSSR, Shestkokrylye X(2). Zool. Inst. Akad. Nauk SSSR (n.s.) 52, 1–274.
Medvedev, S.I., 1976. Sistematike i Filogenii Plastinchatouyx Zhukov Palearktiki.
Entomol. Ob. 55, 400–409.
Midgeley, J., 1992. Why do some hopliinid beetles have large hind-legs? J. Ent. Soc.
Sth. Afr. 55, 157–159.
Monaghan, M.T., Inward, D.G., Hunt, T., Vogler, A.P., 2007. A molecular
phylogenetic analysis of the Scarabaeinae (dung beetles). Mol. Phylogenet.
Evol. 45, 674–692.
Mulsant, E., 1842. Histoire naturelle des coléoptères de France, pt. 2. Lamellicornes,
Paris, 623 pp.
Mulsant, E., Rey, C., 1871. Histoire naturelle des Coleopteres de France.
Lamellicornes et Pectinicornes, Paris, 736 pp.
Nel, A., Scholtz, C.H., 1990. Comparative morphology of the mouthparts of adult
Scarabaeoidea (Coleoptera). Entomol. Mem. 80, 1–84.
Nylander, J.A.A., Ronquist, F., Huelsenbeck, J.P., Nieves-Aldrey, J.L., 2004. Bayesian
phylogenetic analysis of combined data. Syst. Biol. 53, 47–67.
Paulian, R., 1959. Coleopteres scarabéides. Faune de France. Lechevalier, Paris. 2a
edición. 298 pp.
Paulian, R., Baraud, J., 1982. Faune des Coleopteres de France, II. Lucanoidea et
Scarabaeoidea. Lechevalier, Paris, 477 pp.
Peringuey, L., 1902. Descriptive catalogue of the Coleoptera of South Africa. Trans.
South African Philos. Soc. 12, 564–920.
Picker, M.D., Midgeley, J.J., 1996. Pollination by monkey beetles (Coleoptera:
Scarabaeidae: Hopliini): flower and colour preferences. Afr. Entomol. 4, 7–14.
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution.
Bioinform. Appl. Note 14, 817–818.
Rambaut, A., Drummond, A., 2003. Tracer. MCMC Trace Analysis Tool.
Reitter, E., 1902. Bestimmungstabelle der Melolonthiden aus der europäischen
Fauna und den angrenzenden Ländern. Verh. Naturf. Ver. Brünn 41, 27–158.
Schnitzler, J., Dyer, R.J., Barrouclough, T.G., Savolainen, V., in press. The complexity
of plant-pollinator coevolution in a biodiversity hotspot. Evolution.
Simon, C., Frati, F., Beckenbach, A., et al., 1994. Evolution, weighting, and
phylogenetic utility of mitochondrial gene sequences and a compilation of
conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87, 651–
701.
Smith, A.B.T., 2006. A review of the family-group names for the superfamily
Scarabaeoidea (Coleoptera) with corrections to nomenclature and a current
classification. Coleopterists Soc. Monogr. 5, 144–204.
Steiner, K.E., 1998. Beetle pollination of peacock moraeas (Iridaceae) in South Africa.
Plant Syst. Evol. 209, 47–65.
D. Ahrens et al. / Molecular Phylogenetics and Evolution 60 (2011) 408–415
Steiner, K.E., 1999. The evolution of beetle pollination in a South African orchid. Am.
J. Bot. 85, 1180–1193.
Terry, D.M., Whiting, M.F., 2005. Comparison of two alignment techniques within a
single complex data set: POY versus clustal. Cladistics 21, 272–281.
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The
ClustalX windows interface: flexible strategies for multiple sequence alignment
aided by quality analysis tools. Nucl. Acids Res. 25, 4876–4882.
van Kleunen, M., Nänni, I., Donaldson, J.S., Manning, J.C., 2007. The role of beetle
marks and flower colour on visitation by monkey beetles (Hopliini) in the
Greater Cape Floral Region, South Africa. Ann. Bot., 1–7.
415
Vogler, A.P., Pearson, D.L., 1996. A molecular phylogeny of the Tiger beetles
(Cicindelidae): congruence of mitochondrial and nuclear rDNA data sets. Mol.
Phylogenet. Evol. 6, 321–338.
Wheeler, W.C., 1995. Sequence alignment, parameter sensitivity, and the
phylogenetic analysis of molecular data. Syst. Biol. 44, 321–331.
Yang, Z., Rannala, B., 1997. Bayesian phylogenetic inference using DNA sequences: a
Markov chain Monte Carlo method. Mol. Biol. Evol. 14, 717–724.
Zhang, A., Robbins, P.S., Averill, A.L., Weber, D.C., Linn Jr., C.E., Roelofs, W.L., Villani,
M.G., 2003. Identification of the female-produced sex pheromone of the scarab
beetle, Hoplia equina. J. Chem. Ecol. 29, 1635–1642.