Content uploaded by Octávio S Paulo
Author content
All content in this area was uploaded by Octávio S Paulo
Content may be subject to copyright.
Bioacoustics
The International Journal of Animal Sound and its Recording, 2008, Vol. 18, pp. 17–34
0952-4622/08 $10
© 2008 AB Academic Publishers
THE EVOLUTION OF CICADA SONGS
CONTRASTED WITH THE RELATIONSHIPS
INFERRED FROM MITOCHONDRIAL DNA
(INSECTA, HEMIPTERA)
PAULO J. FONSECA11*, ESTER A. SERRÃO2, FRANCISCO PINA-MARTINS1,
PEDRO SILVA3, SARA MIRA2, JOSÉ A. QUARTAU1, OCTÁVIO S. PAULO1
AND LEONOR CANCELA2
1Centro de Biologia Ambiental/Departamento de Biologia Animal, Faculdade de
Ciências da Universidade de Lisboa, P-1749-016 Lisboa, Portugal
2CCMAR, CIMAR-Laboratório Associado, F.C.M.A., Universidade do Algarve,
Gambelas, P-8005-139 Faro, Portugal
3Centro de Genética e Biologia Molecular/Departamento de Biologia Vegetal,
Faculdade de Ciências da Universidade de Lisboa, P-1749-016 Lisboa, Portugal
ABSTRACT
The molecular phylogeny of nine Palaearctic species of cicadas (Hemiptera, Cicadoidea)
was inferred using two mitochondrial DNA genes, Cytochrome Oxidase I and II. The
two main groups detected, namely species within Tettigetta and Tympanistalna, as
well as the two species investigated in the genus Cicada, are robustly supported
across the analytical methods. The structure of the song syllables, generated during
single tymbal cycles of males of the analysed group of species is remarkably consistent
in these two phyletic lines. This reects the morphology and the mechanics of the
tymbal. However the higher level song patterns, which depend on the activity of the
central nervous system and have evolved to advertise receptive mates, do not seem
to be consistent with either the inferred molecular topology or the basic tymbal cycle.
The observed similarities between the molecular phylogeny and the basic tymbal
cycles seem to reect the basic conservative nature of the tymbal structure, while
the discrepancy between the former and the calling song pattern is probably related
to the high plasticity of the pattern generator in the central nervous system and
dependent on species-specic selection.
Keywords: insect calling songs, sound production, molecular phylogenies, cicadas,
cytochrome oxydase
INTRODUCTION
Acoustic communication has evolved in Vertebrates and Arthropods.
Among the arthropods, certain insects have extensively exploited this
communication channel. Acoustic signals are especially frequent in
1*Corresponding author: Email: pjfonseca@fc.ul.pt
18
the Orthoptera and the cicadas, where they evolved as the basis for
mate nding and species recognition. Cicadas exhibit songs that can
be very complex in their amplitude patterns (e.g. Popov 1975; Young
& Josephson 1983; Fonseca 1991; Sueur et al. 2004) and, especially
in tropical species, frequency modulation can be extensively used
(Gogala 1995). Thus, cicadas can be viewed as excellent models to
study the evolution of acoustic signals. In spite of this, there is still
much to be learned on the subject.
In cicadas, only the males produce loud acoustic signals by means
of a tymbal mechanism. There are two components in the production
of such a sound signal: the rst one is the sound producing apparatus
itself, and the second one is the control of this apparatus through the
nervous system. The rst component, the sound producing apparatus,
is a stereotyped tymbal mechanism which is located in the rst and
second abdominal segments (Pringle 1954). The tymbal muscle pulls
on a tymbal plate, which then modies the convex tymbal membrane
loading each of a series of sclerotized ribs. These ribs may buckle
inward in synchrony, or sequentially, and from one to a series of
sound pulses may be produced per muscle contraction (Young &
Bennet-Clark 1995; Fonseca & Bennet-Clark 1998).
The elasticity of the tymbal, which contains resilin, assures
that its resting state is restored upon relaxation of the tymbal
muscle, which may or may not be accompanied by sound. Moreover,
the overall shape of the tymbal – and so the sound pulses – may be
modied by a tensor muscle acting on the tymbal frame (Fonseca &
Hennig 1996). A period of contraction and subsequent relaxation of the
tymbal muscles, i.e. a single cycle of the sound producing apparatus,
generates the sound pulses that constitute the basic element of a
song, the syllable.
The second component of the sound producing system is the
nervous activity which controls the tymbal and tensor muscles. The
coordination of the several elements of the sound producing system
can generate diverse sounds by assembling the basic syllables in
higher order and more complex song elements.
Our main objective is to examine how the different patterns
of cicada calling songs correlate with the molecular phylogenetic
relationships among the species investigated. To accomplish this
we infer the phylogenetic relationships among a sample of nine
species of Palaearctic cicadas using sequences from two fragments of
mitochondrial genes, Cytochrome Oxidase I and II, and comparing the
topologies obtained with the well known patterns of the songs and the
morphological and physiological parameters of the sound producing
system. These two genes were chosen because they are widely
used for this level of phylogenetic analysis and have already been
successfully used to study relationships among cicadas (e.g. Cooley et
al. 2001; Buckley et al. 2001a,b). We are aware that the number of
19
taxa analyzed in this work is small when compared with the number
of cicadas known. However, there were not many more species we
could use or from which the physiological data is available.
MATERIAL AND METHODS
Sampling
Nine species of cicadas present in Portugal, belonging to six genera
and two families, were selected for this study. The Cicadidae were
Cicada barbara Boulard 1982, Cicada orni Linnaeus 1758 and Lyristes
plebejus (Scopoli, 1763), while within the Tibicinidae we used Tibicina
garricola Boulard 1983, Tympanistalna gastrica (Stål, 1854), Tettigetta
argentata (Olivier, 1790), Tettigetta estrellae Boulard 1982, Tettigetta
josei Boulard 1982, and Tettigetta mariae Quartau and Boulard 1995.
The species Philaenus spumarius (L.) of the family Cercopidae, a
related group of the Cicadoidea, was selected as the outgroup. We
used two sequences, one, produced in our lab and another from the
GenBank (Accession Number AY630340)
All specimens were collected in central and southern Portugal
in June and July. The type of sample material used varied from
freshly collected insects, to insects preserved in ethanol or in a high
salt buffer, to dried insects kept in entomological boxes. DNA was
generally extracted from the tymbal muscles, but for smaller insects
the whole individual was extracted.
DNA extraction, amplication and sequencing
DNA extraction
Whole or parts of insects were ground in microcentrifuge tubes with
liquid N, and then incubated at 55° C in extraction buffer (100 mM
NaCl, 50 mM Tris-HCl pH 8.0, 1 µM EDTA, 0.5% SDS, 200 µg/ml
proteinase K) with frequent agitation, followed by a purication step
with phenol and chloroform:isoamyl alcohol and precipitation with 3
M sodium acetate and 100% ethanol. The pellet was washed with
70% ethanol, resuspended in TE, RNA was digested, and RNAse was
removed with chloroform:isoamyl alcohol followed by an overnight
precipitation with 100 % ethanol and sodium acetate. DNA quality and
quantity in the resuspended pellet were assessed on agarose gels. An
alternative DNA extraction method was used for some of the samples:
ground in 5% chelex and 15-20 µl proteinase K (20µg/ml), incubated
at 55-57° C for 3-4 h, followed by 15-20 min at 100° C. However, the
rst method yielded DNA of better quality and quantity.
20
PCR Amplication
Cytochrome oxidase I (COI) and II (COII) from each sample
was symmetrically amplied by PCR, with the following
primers. COI: C1-J-2195 – TTGATTTTTTGGTCATCCAGAAGT
and TL2-N-3014 – TCCAATGCACTAATCTGCCATATTA; COII:
C2-J-3400 – ATTGGACATCAATGATATTGA and C2-N-3661 –
CCACAAATTTCTGAACATTGACCA (Simon et al., 1994). Amplication
conditions were: 1 X buffer II, 1.9 mM MgCl2, 0.2 mM/dNTP, 0.3
µM/primer, 1.25 U of AmpliTaq (Perkin Elmer, Norwalk, CT) and 2
ng of genomic DNA per 25 µL reaction. The PCR parameters were:
94°C for 5 min, 35 cycles of 94°C for 1 min, 48°C for 1 min, and 72°C
for 1 min, followed by a nal extension step at 72°C for 5 min. PCR
products were electrophoresed in agarose gels, the target bands were
cut under UV light, and the DNA was puried with a gel extraction
kit (Qiagen, GmbH, Germany).
Cloning and sequencing
The PCR products were directly ligated into pGEM-T easy vector
(Promega, Madison, Wi) and transformed into supercompetent cells
(Xl1-Blue MRF’, Stratagene, La Jolla, CA). Plasmid DNA from each
recombinant colony was extracted using alkaline lysis minipreps. The
COI and COII of 1-2 individuals of each species were sequenced in
both directions using cycle sequencing with dye-labelled terminators
(Perkin Elmer) on an ABI 373 automated sequencer or in ABI 310
(Applied Biosystems). For both COI and COII it was possible to fully
sequence the genes in both directions using universal primers in the
vector. Sequences were deposited with Genbank (Accession Numbers
for COI: EU401964-EU401973; COII: EU401954-EU401963).
Sequence alignment and phylogenetic analyses
DNA sequences were initially aligned using CLUSTAL X v1.8
(Thompson et al., 1997) and the resulting alignment was then
inspected and manually changed to eliminate a few frame shifting
gaps. Three sets of analyses were carried out, the rst with combined
sequences from CO I and CO II fragments, and the second with only
CO I or CO II sequences.
Phylogenetic analysis was performed using PAUP*4.0.b4a
(Swofford 2000). Modeltest 3.0 software (Posada & Crandall 1998)
associated with PAUP* was used to select the most appropriate
evolutionary model for the different data sets. The most appropriate
model was then used to calculate the maximum likelihood phylogenetic
tree (Felsenstein 1988). The optimal tree was found by a heuristic
21
search with tree-bisection-reconnection as the branch-swapping
algorithm. Initial trees were obtained via stepwise addition with 100
replicates of random addition sequence. The neighbour-joining tree
(Saitou & Nei 1987) for the dataset was also calculated with the
previously selected maximum likelihood distance. This calculation
involved the following parameters: a shape parameter of the gamma
distribution used to set the relative size of four rate categories, the
proportion of invariable sites, and the proportion of the two types of
transition and transversions (or the transition / transversion ratio in
simpler models). Unweighted parsimony analyses were also carried
out on PAUP. For unweighted maximum parsimony, the optimal
tree was found by a heuristic search with tree-bisection-reconnection
as the branch-swapping algorithm. Initial trees were obtained via
stepwise addition with 100 replicates of random addition sequence.
The g1-statistic was calculated from the frequency distribution
of lengths of a thousand random trees and the tree length of the
optimal tree compared with this distribution (Hillis & Huelsenbeck
1992). Ensemble indices – consistency index (Kluge & Farris 1969),
retention index (Farris 1969) and homoplasy index (Archie 1989)
– were calculated to describe the amount of homoplasy. Gaps were
treated as a fth character state for the parsimony analysis or as
missing data for other analyses.
In all, three forms of analyses bootstrapping with 1000
pseudoreplicates was performed to evaluate the robustness of the
nodes, with the trees obtained in the same way of the original
inference one.
To assess if the sequences evolved in a clock-like way, a
likelihood ratio test was performed. The log likelihood value of a tree
with the same topology and evolutionary model was calculated with
and without enforcing a molecular clock. Twice the difference between
calculated values was then compared with a c2 distribution with n-2
degrees of freedom, where n is the number of sequences used in the
analysis (Huelsenbeck & Crandall 1997).
Finally Bayesian analyses were also undertaken using MrBayes
v3.0b4 (Ronquist & Huelsenbeck 2003).The posterior probabilities of
the phylogenetic trees were estimated by a Metropolis-Coupled, Markov
Chain Monte Carlo sampling algorithm (MCMCMC). The Markov
Chain Monte Carlo (MCMC) procedure ensures that trees are sampled
in proportion to their probability of occurrence under the given model
of gene-sequence evolution while the MCMCMC approach ensures
that the Markov chain did not become trapped in local optima. The
conditions for the Bayesian analysis were previously set up in order
that the likelihood scores of the trees would reach stationarity. Using
a general-time-reversible model of sequence evolution with a gamma
distribution for the among sites rate variation, a total of 1.5×106
generations were sampled every 100 generations with a “burn-in”
22
of the rst 7.5% of the trees. Clade credibility values were obtained
from a 50% majority-rule consensus tree of the remaining trees.
Two runs using different random starting seed were used to assess
congruence of the likelihood values (Huelsenbeck et al. 2002). The
prior model was estimated with MrModeltest v2.2 (Nylander 2004).
The Shimodaira-Hasegawa test (Shimodaira & Hasegawa 1999) was
implemented, when necessary, to test alternative tree topologies.
Sound recording and analysis
Calling songs were recorded at 19 cms-1 with a Uher 4200 report tape
recorder through an AKG D202 or a Sennheiser MKE 2 microphone.
These audio assemblages have a frequency response within ±3 dB in
the range 1-18 kHz. Sound analysis was made with software written
by the authors (e.g. Fonseca 1991).
Evaluation of physiological parameters
The tymbal muscle phase and changes in the leading muscle during
singing were evaluated with electromyograms (EMG’s) of the tymbal
muscles obtained from spontaneously singing cicadas. For details on
the EMG recordings see Fonseca (1996). The tensor muscle effect
on the sound amplitude was evaluated with simultaneous electrical
stimulation of the tymbal motoneuron and the tensor nerve. For a
description of the method see Fonseca & Hennig (1996).
Morphology
Morphological and anatomical observations of dry and ethanol
preserved insect specimens were made using a stereomicroscope (Wild
M5A) equipped with a camera lucida that was used to make the
drawings. We paid special attention to structures related to the sound
producing apparatus such as the structure of the ribbed tymbals with
the presence or absence of a coupling bar, the presence of tymbal
covers and the appearance and thickness of the abdominal wall.
RESULTS
Sequence alignment and variation
The dataset comprises eleven aligned sequences including two
outgroup sequences. All sets of analyses were performed for the COI
23
and COII dataset and for the combined dataset of COI and COII.
Only the results for the concatenated dataset are shown since each
dataset give essentially congruent results but with less resolution and
robustness. In a total of 1184 nucleotide sites, 659 are constant in all
sequences, and 341 are parsimony-informative sites. The fragment of
COI has 878 nucleotide sites and COII has 306. The alignments are
available from the authors.
For the datasets there was a strong bias towards AT, the
average for the combined dataset is A – 30.54% and T – 41.09%. In
all cases the sequences passed a Chi-square test of heterogeneity in
nucleotide frequencies (for the combined dataset: Chi-square = 15.35,
df=30, P>0.98). Uncorrected pair wise distances range between 2.1%
to 25.0% for the combined dataset.
Phylogenetic analyses
Maximum parsimony analysis, with gaps treated as missing data or
as a fth character, produced two equally parsimonious trees with
length of 1021 and three with size 1147 respectively. The consistency,
the retention and the homoplasy indices were respectively 0.704,
0.607, 0.296 for the rst analysis (gaps treated as missing data) and
0.730, 0.641, 0.270 for the latter (gaps treated as fth character). The
random generation of trees produced a highly skewed distribution
for the two analyses, suggesting phylogenetic informative sequences
(g1=-0.766; p<0.01 when gaps are treated as missing data g1=-0.875;
p<0.01 when gaps are treated as fth character).
The selected evolutionary model for the concatenated dataset
was the TVM with a proportion of invariable sites and a gamma
correction (TVM+I+Γ) with a G-A and C-T transition rate of 22.2452,
and a A-C transversion rate of 2.9196, A-T rate of 6.1279, C-G rate
of 6.1334 the proportion of invariable sites was 0.3743 and the
shape parameter of the gamma distribution was 1.1088. For the
combined dataset, the likelihood ratio test between trees with the
same evolutionary model but with the molecular clock enforced or
not, show statistical signicant difference (2δ=19.47, df=9) and the
molecular clock hypothesis was rejected at 95% level.
Tree topologies
The dataset produced overall similar topologies that were robust to
the different phylogenetic inference methods which were used in the
analyses.
Two well-supported clades are shown in all the analyses (Figure
1). The rst one is formed by the Tibicinidae species Tympanistalna
24
gastrica, Tettigetta josei, Tettigetta estrellae, Tettigetta mariae and
Tettigetta argentata, all within the tribe Cicadettini. This clade
shows a remarkable internal stability across inference techniques and
datasets. The second one is formed by only two species, the Cicadidae
Cicada barbara and Cicada orni, both within the tribe Cicadini.
Somewhat unexpectedly, the tibicinid species Tibicina garricola
never clustered with the remaining Tibicinidae as would be expected
Figure 1. Phylogram obtained by Bayesian inference of 1184 base pairs of
the Mitochondrial DNA, Cytochrome Oxidase subunits I and II. Numbers
above the branches are inferred number of substitutions per site and the
numbers below branches are the bootstrap branch support values obtained
from 1000 pseudoreplicates from Maximum Likelihood analysis, Neighbour-
Joining trees with the corrected maximum likelihood distance, parsimony and
clade credibility values from Bayesian phylogenetic inference, respectively.
Adjacent to each terminal node (i.e. each species) there is a scheme with the
IN/OUT pulses of the Tymbal cycle and the respective Tymbal morphology.
The time scale of the pulses is in milliseconds (ms) and the respective scale is
shown. The scale of the morphological schemes represents 1 mm. The tymbal
morphology is characterised by a) the tymbal plate, b) the longitudinal ribs
and c) the tensor sclerite. The presence of a dorsal coupling bar connecting
the longitudinal ribs is shown (see solid arrows).
25
according to traditional taxonomy, in spite of belonging to a different
tribe (Tibicinini). Similarly, the third cicadid, Lyristes plebejus, within
the tribe Lyristini, also never clusters with the two other Cicadidae,
always clustering instead with Tibicina garricola, although with
low robustness values (Figure 1). These two species seem to have
an unresolved position in the tree either forming a polytomy with
the two other clades (bayesian) or cluster with the other Tibicinidae
species (maximum likelihood, neighbour-joining, parsimony) or with
the other Cicadidae species (parsimony) or even appear in basal
position in relation with the two other clades (parsimony).
A Shimodaira-Hasegawa test showed no statistical signicant
difference between the optimal tree (Figure 1) and several alternative
suboptimal trees with the topology constrained according to different
hypothesis namely with a different basal clade (p=0.330; p=0.432) or
non-monophyly of the three main ingroup clades (p=0.330; p=0.432;
p=1.000).
Sound producing apparatus and song characteristics
The characteristics of the sound producing apparatus and the pattern
of the calling songs of the nine species are summarized in Table 1
(for clarity in the text each parameter indicated in Table 1 is shown
in bold).
The tymbal mechanism of sound production consists of a ribbed
tymbal that is pulled by the tymbal muscle that, upon contraction,
leads to the buckling of the tymbal with subsequent sound generation.
Sound pulses may be produced during the inward (IN) movement of
the tymbal as well as when the tymbal pops out (OUT) due to elastic
forces (Figure1). In the Tettigetta and Tympanistalna the tymbal
generates two sound pulses in each tymbal cycle, or syllable, one at
the inward and the other upon the outward movement of the tymbal.
In Tibicina there is a series of well delimited pulses generated by the
collapse of the same number of ribs during the IN, while the OUT is
almost silent. In the Cicadidae the inward buckling is accompanied
by two (Lyristes) or three (Cicada) pulses more or less fused together
in a complex wave while the OUT produces one sound pulse.
The number and position of the tymbal ribs is different among
species. Unlike the species included in the family Cicadidae, the
Tettigetta spp. and Tympanistalna gastrica exhibit a tymbal coupling
bar dorsally joining the long ribs of their tymbals (see arrows in
Figure1) at their dorsal edges. This coupling bar causes the tymbal
to collapse as a whole producing a single IN sound pulse. The other
Tibicinidae, Tibicina garricola, also has the longitudinal long ribs
connected dorsally to a rib-like bar, but here the ribs are spaced out
dorsally (cf. Figure1). Moreover, while the three Cicadidae, Lyristes
26
TABLE 1
Selected song, physiological and morphological parameters of the cicadas analysed in this study.
The representatives of the two families (Tibicinidae and Cicadidae) are separated, and the species
are clustered by genera and presented in the same order as in Figure 1. Identical values for the
parameters analysed are shown in identical grey shading. A question mark (?) indicates lack of data.
Morphological parameters
Song parameters
Physiological parameters
Family
Species
Tymbal
coupling
bar
Tymbal
covers
Thickness
of
abdominal
wall
Tymbal
cycle
Frequency
peak of
calling
song (kHz)
Gross
temporal
pattern of
calling song
Tymbal
muscle phase
Changes
in the
leading
muscle
Tensor
muscle effect
on sound
pulses
Tettigetta
argentata
Yes
No
Thick
1 IN / 1
OUT
13-15
Simple
phrases
0.1
No
Strong
amplitude
increase
Tettigetta
mariae
Yes
No
Thick
1 IN / 1
OUT
13-14
Simple
phrases
?
?
Strong
amplitude
increase
Tettigetta
estrellae
Yes
No
Thick
1 IN / 1
OUT
11-14
Complex
phrases
< 0.1
?
Strong
amplitude
increase
Tettigetta josei
Yes
No
Thick
1 IN / 1
OUT
15-17
Complex
phrases
0.03 (close to
synchrony)
No
Strong
amplitude
increase
Tympanistalna
gastrica
Yes
No
Thin and
translucent
1 IN / 1
OUT
10-13
Complex
phrases
0.15
No
Strong
amplitude
decrease
Tibicinidae
Tibicina
garricola
Yes, but
ribs well
separated
dorsally
No
Thick
3-7 IN / 1
quiet
OUT
8-10
Continuous
song
0.3-0.5
Yes Not clear
Lyristes plebejus
No
Yes
Thick
2 fused
pulses IN
/ 1 OUT
5-6
Complex
phrases
0.4 -0.5
? ?
Cicada orni
No
Yes
Thin and
translucent
3 fused
pulses IN
/ 1 OUT
4-5
Simple
discontinuous
song
0.5
Yes Not clear
Cicadidae
Cicada barbara
No
Yes
Thin and
translucent
3 fused
pulses IN
/ 1 OUT
5-7
Continuous
song
0.5
Yes
Not clear
Table 1
27
plebejus, Cicada orni and Cicada barbara, possess tymbal covers,
i.e. integument expansions that project anteriorly and cover in part
(Cicada) or completely (Lyristes) the tymbals, these structures are
not present in the other species. The thickness of the abdominal
wall also varies. It is thick in the majority of the cicadas, but in
Tympanistalna gastrica and in the two species of the genus Cicada
the wall of the abdomen is thin and translucent.
The vibrations induced in the tymbal by the pull of the muscle,
which correspond to the resonant properties of this structure backed
by an internal air chamber, are then radiated to the air by a number
of structures. These include the main generator, the tymbal, that
usually contributes the spectral peak to the song, but sound is
conveyed also through the tympana. The abdominal wall may also
play an important role, especially if it is thin. The species studied
can be grouped according to the frequency peaks of their calling
songs. The Tettigetta species generate sounds with higher frequency
components (10-17 KHz), followed by Tympanistalna gastrica (10-13
kHz), and by Tibicina garricola (8-10 KHz). In contrast, the Cicadidae
produce lower pitched sounds 4-7 KHz).
The nervous system dictates which tymbal is activated rst
and the phase between the paired tymbal muscles. Changes in the
leading muscle were not seen in the Tettigetta and Tympanistalna
but were observed in Tibicina and in the Cicadidae (Lyristes and
Cicada spp.). A similar pattern was seen in the tymbal muscle
phase. The Tettigetta species and Tympanistalna show smaller
phases between the activation of the tymbal muscles (from almost
synchrony in Tettigetta josei to about 0.15 in Tympanistalna gastrica)
while in the other species the tymbals alternate with phases ranging
from 0.3 to 0.5.
Finally, the tensor muscle effect on the sound pulses can
be different in the several species. While in Tettigetta its contraction
strongly increased the sound amplitude, in Tympanistalna gastrica
the opposite effect was obtained with a marked decrease of the sound
amplitude. The effect was not clear in Tibicina and the two Cicada
species, and it was not assessed in Lyristes plebejus.
The songs can be classied according to their gross temporal
pattern. In continuous songs observed in Cicada barbara and Tibicina
garricola the syllables are simply repeated in a continuous sequence
for many seconds to several minutes. In contrast, in Cicada orni, which
exhibits a simple discontinuous song, this sequence is interrupted by
regular pauses. In other species the sound sequence is organized in
characteristic phrases, i.e. long sound units composed of a certain
sequence of groups of syllables and pauses, which are repeated over
time. Tettigetta argentata and Tettigetta mariae have simple phrases
with groups of similar syllables along the phrase. In contrast, in the
complex phrases of Tettigetta estrellae, Tettigetta josei, Tympanistalna
28
gastrica and Lyristes plebejus there are sequences of different groups
of syllables along the phrase (see Fonseca 1991 and Sueur et al. 2004
for a detailed description of the songs).
The observed Gross temporal pattern of the calling songs,
which measures the contribution of the second component of the sound
producing system, shows a remarkably different pattern from the one
inferred by the phylogenetic results. Simple and complex phrases
appear in both groups, and continuous or discontinuous songs are
produced by species in the several groups.
DISCUSSION
Although the topologies produced by different inference methods
were not always identical, there was no strong support for conicting
arrangements. One possible reason for this unstable situation could
be related to the asymmetry of the species sample by using single
species from both of the genera Tibicina and Lyristes. A more
balanced and larger taxonomic coverage and other genes might help
in solving the current ambiguities. However, the well supported clades
of the phylogenetic analyses are in agreement with the traditional
classications based mostly on morphological characters (Webb 1979;
Boulard 1982,1987). Inside the well-supported Tibicinidae clade (Figure
1) the relations among the species are identical across the different
inference methods. In relation to the Cicadidae species, this set of
results support what others already have shown with morphology and
genetics: that Cicada barbara and Cicada orni are closely related and
well dened species (Quartau 1988; Pinto et al. 1998; Quartau et al.
2001).
The phylogeny obtained with this dataset, in spite of some
ambiguities, was used as a framework for the comparisons of the
sound apparatus and acoustic patterns of the several species.
The calling songs of cicadas exhibit high variability, ranging
from continuous to discontinuous songs with different degrees of
amplitude modulation, to songs with a complex pattern repeated
over time ( e.g. Popov 1975; Young & Josephson 1983; Fonseca 1991;
Sueur et al. 2004). Our main objective was to nd out how exactly
the molecular phylogeny of the cicadas correlates with the observed
pattern of the songs.
The rst component of the sound producing system, the sound
producing apparatus itself, is responsible for the basic sound elements
produced during one tymbal cycle. These are similar in the Tettigetta
spp. and Tympanistalna gastrica, which generate two loud pulses,
one during the inward buckling of the tymbal (IN pulse), and another
resulting from its outward movement (OUT pulse), a tymbal cycle
pattern shared with many Cicadetta species (e.g. Popov 1975; Gogala
29
et al. 1996). These species cluster in a well-supported phylogenetic
clade. The Cicadidae group show a different tymbal cycle, with a
complex IN pulse resulting from two (Lyristes) or three (Cicada)
partially fused pulses followed by one loud OUT pulse (Popov 1975;
Fonseca 1991), a pattern also seen in other genera of this group like
the North-American Tibicen (e.g. Hennig et al. 1994). In the phylogram
the two Cicada species are grouped together, while Lyristes plebejus
has an unresolved status. Tibicina garricola shows a distinct syllable
producing up to seven pulses during the inward tymbal movement,
while the OUT pulse is almost silent, a pattern similar to other
Tibicina (e.g. Sueur & Aubin 2003) and to the American Magicicada
(Moore & Sawyer 1966) and Okanagana species (Stölting et al. 2004).
This species also has an unresolved relationship with the well-dened
clades of the phylogenetic trees.
All the species producing 1 IN –1 OUT pulse have a coupling
bar located dorsally to the longitudinal tymbal ribs, coming close to
the tymbal plate. This bar, probably responsible for the buckling of
the tymbal as a whole (Fonseca & Bennet-Clark 1998), is not present
in any of the other species except Tibicina garricola. Consequently,
this organization of the tymbal and the sound produced (syllable)
is exclusive of the robust clade that assembles Tettigetta and
Tympanistalna. In Tibicina, although all the ribs are dorsally linked
by a sclerotized rib-like structure, the arrangement of the ribs and the
tymbal is very different. Here the long ribs are well separated in all
their length and are not apposed as in Tettigetta and Tympanistalna,
and the resulting IN produces not a single but instead several pulses,
one per each rib collapsed.
Despite the similarities of the basic sound production found
in the group of species with a coupling bar, the effect of the tensor
muscle, responsible for the amplitude modulation of the songs,
introduces another source of variability. Its effect is opposite in the
Tettigetta spp. and in Tympanistalna gastrica (Fonseca & Hennig
1996). In the former it increases the sound amplitude while in the
latter the contraction of the tensor muscle generates a decrease.
This opposite effect seems to corroborate the phylogenetic results,
as these two genera seem to be well-dened sister taxa. Moreover,
while the abdominal wall is very thin in Tympanistalna gastrica,
playing a signicant role in sound radiation (Fonseca & Popov 1994),
in Tettigetta it is much thicker and therefore does not work as a
radiator.
In the Tympanistalna and Tettigetta clade, not only do the
tymbal buckle as a whole, but the left-right tymbal cycle shows only
a small phase difference (< 0.2 of a cycle), while in the other four
species the tymbals almost alternate (phase differences about 0.5). It
is especially noteworthy that Tibicina garricola, which is traditionally
classied with the Tibicinidae, appears outside the group including
30
other traditional members of this group in this phylogenetic analysis.
Tibicina also shows a tymbal cycle phase difference inconsistent with
the other members of the traditional classication. The same pattern
can be observed in the activation of the muscles, since changes in
the leading muscle during the calling song were never observed in
Tympanistalna and Tettigetta, while they have been found in the
Cicada and Tibicina (Fonseca 1996). Finally, all three Cicadidae
species display tymbal covers, which no other species show, suggesting
that this character could have evolved in the lineage leading to this
group.
The phylogenetic relations seem also to be reected in the
spectral peak of the calling songs, which is greatly dependent on the
mechanics of the tymbal (Fonseca & Popov 1994) and, according to
Bennet-Clark & Young (1994), also on body size. The calling songs
with higher frequency peaks are associated with the Tettigetta /
Tympanistalna clade, also the smaller species in our sample, while
the Cicadidae species (Cicada and Lyristes) exhibit lower frequencies,
but here the much larger L. plebejus has a frequency peak similar to
Cicada spp. Tibicina garricola, a species with a body length similar
to Cicada has an intermediate spectral peak, i.e. higher then Cicada
and Lyristes but lower than Tettigetta and Tympanistalna, as do other
members of the genus Tibicina (Sueur & Aubin 2003) and the possibly
related American species of Okanagana (e.g. Stölting et al. 2004).
While the syllable depends on the structure and thus the
mechanics of the tymbal, the second component, the role of the nervous
system which generates the calling song pattern by controlling the
sound producing apparatus, is a major determinant of the species
specicity of the songs. This is obtained as a result of the activity
of a very plastic nervous system that assembles the bilateral tymbal
cycles into complex and diversied temporal sequences. At this level
there are less biomechanical constraints to the evolution of the songs,
more in line with the constraints of the vocal system of vertebrates
(e.g. frogs, Ryan & Rand 2003).
In contrast with the close phylogenetic relations shown by the
species with similar tymbal cycles, as expected, this second level in the
assemblage of the songs does not seem to agree with the phylogenetic
trees. Similar types of songs can be found in distant positions and,
conversely, sister species can have rather different song patterns.
Another example of such plasticity in a closely related group can
be found in New Zealand Maoricicada species (Buckley et al. 2006).
Nevertheless, examples exist of close related species (e.g. Tibicina
spp., Sueur & Aubin 2003; Sueur et al. 2007) that exhibit similar
calling songs even when sympatric. This, however, only reinforces
the plasticity of the nervous system and behaviour, since in those
cases other cues should be used by females, at least at close range,
to discriminate males of their own species.
31
The same was found in other groups where characteristic songs
are considered premating signals mediating reproductive isolation.
For instance, In the Drosophila willistoni species complex Gleason
& Ritchie (1998) found no correlation between song and genetic
divergence, and phylogenies derived from song patterns did not
reect molecular phylogenies. Likewise, the phylogenies of frogs in
the Physalaemus pustulosus species group obtained from molecular
sequences, allozymes and morphological characters were signicantly
different from the tree derived from characters of the advertisement
calls (Cannatella et al. 1998).
This diversity in the patterns exhibited by calling songs is
likely the result of selection on these congregating signals where
related species overlap in time and space. A good example can be
seen in the sympatric sister species of Cicada. While Cicada barbara
produces a continuous signal, Cicada orni has a simple discontinuous
song. Similarly, Tettigetta mariae has a simple discontinuous song,
while Tettigetta estrellae shows a complex pattern composed of short
elements, and Tettigetta josei has a complex pattern of long phrases
(Fonseca 1991). Thus, as often seen, closely related sympatric species
exhibit considerable diversity in their signals (e.g. Popov et al. 1974;
Ritchie & Gleason 1995; Gerhardt & Huber 2002). Conversely, Lyristes
plebejus exhibits a complex pattern of the phrases of its song, along
with the complexity of the calling songs of some of its phylogenetically
unrelated Tettigetta species (cf. table 1).
The diversity of the calling songs seems to be very homoplasious
(Gleason & Ritchie 1998), while the basic song characteristics which
depend on the structural elements, such as the tymbal are more
conservative and robust during evolution and consequently seem to
have a closer correspondence to the phylogeny of genera and higher
groups.
This set of results illustrates the complex evolution of the
cicada sounds in relation to molecular phylogeny, clearly with two
components. One component, more basic, is strongly related with a
more conservative evolutionary pattern, while the other, at higher
level, is closely related with the intrinsic neural plasticity and has
been probably selected under different regimes that may have been
acting in the mating signals, namely adaptation to the environment
for improved signalling to the receiver, avoidance of heterospecic
encounters and sexual selection, among other selective forces. The
complex pattern shown in this study is expected to turn into a
more robust one with enlarged datasets and better knowledge of the
phylogenies of the cicadas and of the mechanisms involved in their
song production. A much larger taxa dataset may reveal additional
relationships in the evolution of cicadas and their song signals.
32
ACKNOWLEDGMENTS
This work was carried out with support from the Centro de Biologia
Ambiental, CCMAR, and a grant by FCT, Portugal and FEDER
(POCTI/BSE/48365/2002). We thank two anonymous referees for their
comments.
REFERENCES
Archie, J.W. (1989). Homoplasy excess ratios – new indexes for measuring levels of
homoplasy in phylogenetic systematics and critique of the consistency index. Syst.
Zool., 38, 253-269.
Bennet-Clark, H. & Young, D. (1994). The scaling of song frequency in cicadas.
J. Exp. Biol., 191, 291-294.
Boulard, M. (1982). Les cigales du Portugal, contribution à leur étude (Homoptera,
Cicadidae). Annls. Soc. Ent. Fr. (NS), 18, 181-198.
Boulard, M. (1987). Taxonomie et nomenclature superieures des Cicadoidea, histoire,
problèmes et solutions. École Pratique des Hautes Études, Travaux du Laboratoide
d’Entomologie, Biologie et Evolution des Insectes Hemipteroidea, Paris. 1, 1-89
Buckley, T.R., Cordeiro, M., Marshall, D.C. & Simon,C. (2006). Differentiating between
hypothesis of lineage sorting and introgression in New Zealand alpine cicadas
(Maoricicada Dugdale). Syst. Biol., 55 (3), 411-425.
Buckley, T.R., Simon,C., & Chambers, G.K. (2001a). Phylogeography of the New
Zealand cicada Maoricicada campbelli based on mitochondrial DNA sequences:
ancient clades associated with cenozoic environmental change. Evolution, 55,
1395-1407.
Buckley, T.R., Simon,C., Shimodaira, H., & Chambers, G.K. (2001b). Evaluating
hypotheses on the origin and evolution of the New Zealand alpine cicadas
(Maoricicada) using multiple-comparison tests of tree topology. Mol. Biol and Evol.,
18, 223-234.
Cannatella, D.C., Hillis, D.M., Chippindale, P.T., Weigt, L., Rand, A.S., & Ryan, M.J.
(1998). Phylogeny of frogs of the Physalaemus pustulosus species group, with an
examination of data incongruence. Syst. Biol., 47, 311-335.
Cooley, J.R., Simon, C., Marshall, D.C., Slon, K., & Ehrhardt, C. (2001). Allochronic
speciation, secondary contact, and reproductive character displacement in periodical
cicadas (Hemiptera: Magicicada spp.): genetic, morphological, and behavioural
evidence. Mol. Ecol., 10, 661-671.
Farris, J.S. (1969). A successive approximations approach to character weighting.
Syst. Zool., 18, 374-385.
Felsenstein, J. (1988). Phylogenies from molecular sequences: inference and reliability.
Annu. Rev. Genet., 22, 521-565.
Fonseca, P.J. (1991). Characteristics of the acoustic signals in nine species
of cicadas (Homoptera, Cicadidae). Bioacoustics, 3, 173-182.
Fonseca, P.J. (1996). Sound production in cicadas: timbal muscle activity during
calling song and protest song. Bioacoustics, 7, 13-31.
Fonseca, P.J. & Bennet-Clark, H.C. (1998). Asymmetry of tymbal action and structure
in a cicada: a possible role in the production of complex songs. J. Exp. Biol., 201,
717-730.
Fonseca, P.J. & Hennig, R. (1996). Phasic action of the tensor muscle modulates the
calling song in cicadas. J. Exp. Biol., 199, 1535-1544.
Fonseca, P.J. & Popov, A.V. (1994). Sound radiation in a cicada: the role of different
structures. J. Comp. Physiol., 175, 349-361.
33
Gerhardt, H.C. & Huber, H. (2002). Acoustic Communication in Insects and Anurans,
Common Problems and Diverse Solutions. Chicago, London: The University of
Chicago Press.
Gleason, J.M. & Ritchie, M.G. (1998). Evolution of courtship song and reproductive
isolation in the Drosophila willistoni species complex: Do sexual signals diverge the
most quickly? Evolution, 52, 1493-1500.
Gogala, M. (1995). Songs of four cicada species from Thailand. Bioacoustics, 6,
101-116.
Gogala, M., Popov, A.V., & Ribariè, D. (1996). Bioacoustics of singing cicadas of
the western palearctic: Cicadetta tibialis (Panzer) (Cicadoidea: Tibicinidae). Acta
Entomologica Slovenica, 4, 46-62.
Hennig, R.M., Weber, T., Moore, T.E., Huber, F., Kleindienst, H.U., & Popov, A.V.
(1994). Function of the Tensor Muscle in the Cicada Tibicen Linnei. J. Exp. Biol.,
187, 33-44.
Hillis, D.M. & Huelsenbeck, J.P. (1992). Signal, noise, and reliability in molecular
phylogenetic analyses. J. Hered., 83, 189-195.
Huelsenbeck, J.P. & Crandall, K.A. (1997). Phylogeny estimation and hypothesis
testing using maximum likelihood. Annu. Rev. Ecol. Syst., 28, 437-466.
Huelsenbeck, J.P., Larget, B., Miller, R.E., & Ronquist, F. (2002). Potential applications
and pitfalls of Bayesian inference of phylogeny. Syst. Biol., 51, 673-688.
Kluge, A.G. & Farris, J.S. (1969). Quantitative phyletics and the evolution of anurans.
Syst. Zool., 18, 1-32.
Moore, T.E. & Sawyer, R.T. (1966). Mechanism of Cicada Timbal Action (Insecta –
Homoptera - Cicadidae). Am. Zool., 6, 509.
Nylander, J.A.A. (2004). MrModeltest v2. Program distributed by the author. Uppsala
University: Evolutionary Biology Centre.
Pinto, G.A., Quartau, J.A., Morgan, J.C., & Hemingway, J. (1998). Preliminary data
on the sequencing of a fragment of cytochrome B gene of mitochondrial DNA
in Cicada orni L. (Homoptera: Cicadidae) in Portugal. Bol. Soc. Port. Ent., 188,
57-66.
Popov, A.V. (1975). The structure of the tymbals and the characteristics of the sound
signals in singing cicadas (Homoptera, Cicadidae) in the southern regions of the
USSR. Entomol. Obozr., 54, 258-291.
Popov, A.V., Shuvalov, V.F., Svetlogorskaya, I.D., & Markovich, A.M. (1974).
Acoustic behaviour and auditory system in insects. In Mechanoreception (Ed. by J.
Schwartzkopff), pp. 281-306. Abh. Rheinisch-Westfäl.Akad. Wiss.
Posada, D. & Crandall, K.A. (1998). MODELTEST: testing the model of DNA
substitution. Bioinformatics, 14, 817-818.
Pringle, J.W.S. (1954). A physiological analysis of cicada song. J. Exp. Biol., 31,
525-560.
Quartau, J.A. (1988). A numerical taxonomic analysis of interspecic morphological
differences in two closely related species of Cicada (Homoptera, Cicadidae) in
Portugal. Great Basin Naturalist Memoirs, 12, 171-181.
Quartau, J.A., Ribeiro, M., Simões, P.C. & Coelho, M.M. (2001) Genetic divergence
among populations of two closely related species of Cicada Linnaeus (Hemiptera:
Cicadoidea) in Portugal. Insect Syst. Evol, 32, 99-106.
Ritchie, M.G. & Gleason, J.M. (1995). Rapid Evolution of Courtship Song Pattern in
Drosophila-Willistoni Sibling Species. J. Evol. Biol., 8, 463-479.
Ronquist, F. & Huelsenbeck, J.P. (2003). MrBayes 3: Bayesian phylogenetic inference
under mixed models. Bioinformatics, 19, 1572-1574.
Ryan, M.J. & Rand, A.S. (2003). Sexual selection in female perceptual space: How
female tungara frogs perceive and respond to complex population variation in
acoustic mating signals. Evolution, 57, 2608-2618.
Saitou, N. & Nei, M. (1987). The neighbor-joining method: A new method for
reconstructing phylogenetic trees. Mol. Biol. Evol., 4, 406-425.
34
Shimodaira, H. & Hasegawa, M. (1999). Multiple Comparisons of Log-Likelihoods with
Applications to Phylogenetic Inference. Mol. Biol. Evol., 16, 1114-1116.
Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., & Flook, P. (1994). Evolution,
weighting, and phylogenetic utility of mitochondrial gene sequences and a
compilation of conserved polymerase chain reaction primers. Ann. Ent. Soc. Am.,
87, 651-701.
Stölting, H., Moore, T.E., & Lakes-Harlan, R. (2004). Acoustic communication in
Okanagana rimosa (Say) (Homoptera : Cicadidae). Zoology, 107, 243-257.
Sueur, J. & Aubin, T. (2003). Specicity of cicada calling songs in the genus Tibicina
(Hemiptera : Cicadidae). Syst. Entomol., 28, 481-492.
Sueur, J., Puissant, S., Simoes, P.C., Seabra, S., Boulard, M., & Quartau, J.A. (2004).
Cicadas from Portugal: revised list of species with eco-ethological data (Hemiptera:
Cicadidae). Insect Syst. Evol., 35, 177-187.
Sueur, J., Vanderpool, D., Simon, C., Ouvrard, D. & Bourgoin, T. (2007). Molecular
phylogeny of the genus Tibicina (Hemiptera: Cicadidae): rapid radiation and
acoustic behaviour. Biol. J. Linn. Soc., 91, 611-626.
Swofford, D.L. (2000). PAUP*. Phylogenetic Analysis Using Parsimony (*and other
methods). v. 4.0. Sunderland, MA: Sinauer Associates.
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., & Higgins, D.G. (1997).
The CLUSTAL_X windows interface: exible strategies for multiple sequence
alignment aided by quality analysis tools. Nucleic Acids Res., 25, 4876-4882.
Webb, M.D. (1979). Revision of Rambur’s Homoptera species from the types in the
British Museum. Annls. Soc. Ent. Fr. (NS), 15, 227-240.
Young, D. & Bennet-Clark, H.C. (1995). The role of the tymbal in cicada sound
production. J. Exp. Biol., 198, 1001-1019.
Young, D. & Josephson, R.K. (1983). Mechanisms of Sound-Production and Muscle-
Contraction Kinetics in Cicadas. J. Comp. Physiol., 152, 183-195.
Received 28 June 2007, revised 18 October 2007 and accepted 30 October 2007.
