Extreme convergence in stick insect evolution: Phylogenetic placement of the Lord Howe Island tree lobster

Abstract

The 'tree lobsters' are an enigmatic group of robust, ground-dwelling stick insects (order Phasmatodea) from the subfamily Eurycanthinae, distributed in New Guinea, New Caledonia and associated islands. Its most famous member is the Lord Howe Island stick insect Dryococelus australis (Montrouzier), which was believed to have become extinct but was rediscovered in 2001 and is considered to be one of the rarest insects in the world. To resolve the evolutionary position of Dryococelus, we constructed a phylogeny from approximately 2.4 kb of mitochondrial and nuclear sequence data from representatives of all major phasmatodean lineages. Our data placed Dryococelus and the New Caledonian tree lobsters outside the New Guinean Eurycanthinae as members of an unrelated Australasian stick insect clade, the Lanceocercata. These results suggest a convergent origin of the 'tree lobster' body form. Our reanalysis of tree lobster characters provides additional support for our hypothesis of convergent evolution. We conclude that the phenotypic traits leading to the traditional classification are convergent adaptations to ground-living behaviour. Our molecular dating analyses indicate an ancient divergence (more than 22 Myr ago) between Dryococelus and its Australian relatives. Hence, Dryococelus represents a long-standing separate evolutionary lineage within the stick insects and must be regarded as a key taxon to protect with respect to phasmatodean diversity.

Full text

Extreme convergence in stick insect evolution:
phylogenetic placement of the Lord Howe
Island tree lobster
Thomas R. Buckley
1,
*, Dilini Attanayake
1
and Sven Bradler
2
1
Landcare Research, Private Bag 92170, Auckland 1142, New Zealand
2
Johann-Friedrich-Blumenbach-Institut fu
¨r Zoologie und Anthropologie, Georg-August-Universita
¨tGo
¨ttingen,
Berliner Strasse 28, 37073 Go
¨ttingen, Germany
The ‘tree lobsters’ are an enigmatic group of robust, ground-dwelling stick insects (order Phasmatodea)
from the subfamily Eurycanthinae, distributed in New Guinea, New Caledonia and associated islands. Its
most famous member is the Lord Howe Island stick insect Dryococelus australis (Montrouzier), which was
believed to have become extinct but was rediscovered in 2001 and is considered to be one of the rarest
insects in the world. To resolve the evolutionary position of Dryococelus, we constructed a phylogeny from
approximately 2.4 kb of mitochondrial and nuclear sequence data from representatives of all major
phasmatodean lineages. Our data placed Dryococelus and the New Caledonian tree lobsters outside
the New Guinean Eurycanthinae as members of an unrelated Australasian stick insect clade, the
Lanceocercata. These results suggest a convergent origin of the ‘tree lobster’ body form. Our reanalysis of
tree lobster characters provides additional support for our hypothesis of convergent evolution. We
conclude that the phenotypic traits leading to the traditional classification are convergent adaptations to
ground-living behaviour. Our molecular dating analyses indicate an ancient divergence (more than 22 Myr
ago) between Dryococelus and its Australian relatives. Hence, Dryococelus represents a long-standing
separate evolutionary lineage within the stick insects and must be regarded as a key taxon to protect with
respect to phasmatodean diversity.
Keywords: convergent evolution; conservation genetics; Phasmatodea; Dryococelus; Eurycanthinae;
Lanceocercata
1. INTRODUCTION
The rediscovery of an organism long thought extinct is a
very rare and fortunate event. The discovery of a small
population of ‘tree lobsters’ on a rocky offshore islet in
the South Pacific Ocean in 2001 was sensational news
for conservationists ( Priddel et al. 2003;Pain 2006;
Robertson 2006). The ‘tree lobsters’ or ‘land lobsters’
(Rentz 1996) are distinct ground-dwelling ecomorphs of
stick insects (insect order Phasmatodea) that are instantly
recognizable due to a unique combination of morpho-
logical and behavioural characters: tree lobsters are
flightless, with a dorsoventrally flattened body, a robust,
stocky habitus (body form) and square-edged thoracic
segments, often exhibiting an elongated secondary ovipo-
sitor in the female and enlarged, powerfully armed hind
legs in the male (figure 1a–f ;Gurney 1947;Hsiung 1987).
In contrast to the majority of stick insects, which are
solitary canopy-dwellers that drop or flick their eggs to the
ground, tree lobsters aggregate in large numbers in cavities
near the ground and deposit their eggs into the soil
(Lea 1916;Bedford 1976;Honan 2008; S. Bradler 2007,
personal observation).
The most famous ‘tree lobster’ is the Lord Howe
Island stick insect, Dryococelus australis (Montrouzier)
(figure 1a,b). This species is a large (up to 130 mm)
ground-dwelling stick insect that was formerly common
throughout Lord Howe Island. The introduction of rats
via a shipwreck in 1918 led to the extinction of the species
on Lord Howe Island by the 1960s at the latest (Priddel
et al. 2003). Long considered to be extinct ( Paramonov
1963;Key 1991;Rentz 1996), a population of the species
was rediscovered in 2001 on Balls Pyramid, a very small,
200 metre wide rock pyramid approximately 25 km from
Lord Howe Island (Priddel et al. 2003). The population
size appeared not to exceed two dozen individuals,
indicating that the species is indeed one of the rarest
insects in the world (Robertson 2006). Since its
rediscovery, a captive population has been established at
Melbourne Zoo and plans to reintroduce the species to
Lord Howe Island after eradication of the rats has been
prepared (Pain 2006;Honan 2008).
Traditionally, the tree lobsters pertain to the subfamily
Eurycanthinae, which has its greatest diversity in New
Guinea and associated islands(Gu
¨nther 1953);for example,
the well-known genera Eurycantha (figure 1e,f)and
Thaumatobactron. These taxa are considered to bethe closest
relatives of Dr yococelus (Gurney 1947;Hennemann & Conle
2006;Honan 2008), thus implying a dispersal of tree
lobsters from New Guinea to Lord Howe Island along the
former Lord Howe Rise. However, the Eurycanthinae is
also recorded from New Caledonia, where it is represented
by Canachus (figure 1c,d) among other less studied genera.
Proc. R. Soc. B (2009) 276, 1055–1062
doi:10.1098/rspb.2008.1552
Published online 16 December 2008
Electronic supplementary material is available at http://dx.doi.org/10.
1098/rspb.2008.1552 or via http://journals.royalsociety.org.
*Author for correspondence ( buckleyt@landcareresearch.co.nz).
Received 27 October 2008
Accepted 26 November 2008 1055 This journal is q2008 The Royal Society

The tree lobsters are also recorded from a few other
islands in the southwest Pacific and off the northern tip of
Australia. The subfamily Eurycanthinae also comprises
genera that are not considered tree lobsters, such as the
bush-dwelling and slender stick insects Asprenas from New
Caledonia and Neopromachus from New Guinea. Other
arboreal taxa such as Cnipsus from New Caledonia have
been placed in the Eurycanthinae by some (Gu
¨nther
1953), but not all authors (Zompro 2001).
Because a considerable investment in the survival of
the species is being made, it is crucial to determine the
evolutionary heritage and phylogenetic distinctiveness of
Dryococelus (Vane-Wright et al.1991;Faith 1992;Mooers
2007). It is currently unknown how phylogenetically
distinct Dryococelus is with regard to the remaining tree
lobsters and what its biogeographic history is. For example,
because Lord Howe Island only emerged 6.4– 6.9 Myr ago
(McDougall et al. 1981), it is possible that Dryococelus is
only very recently derived from the New Guinean or New
Caledonian lineages of tree lobsters. Alternatively, it may
have a much more ancient evolutionary history, which
would enhance its conservation value.
To examine the phylogenetic position of Dryococelus
among stick insects, to reconstruct its biogeographic
history and to assess its conservation value, we obtained
nuclear and mitochondrial DNA sequence data from
almost all major lineages of phasmatodean diversity. In
particular, we sampled all key eurycanthine genera in
addition to a broad selection of other stick insect taxa from
the Pacific and Australasian regions. We also applied a
Bayesian relaxed clock with fossil calibrations to infer the
evolutionary age of Dryococelus and provide the first
divergence time estimates across the Phasmatodea.
2. MATERIAL AND METHODS
(a)Taxon sampling
Our sampling was designed to include all major lineages
of Phasmatodea and also to maximize the sampling of
genera in the Australasian region. We sampled 16 of the
18 traditional euphasmatodean subfamilies, following the
classification of Gu
¨nther (1953) for the reasons outlined by
Klug & Bradler (2006), with amendments from Zompro
(2001). Generic nomenclature follows Otte & Brock (2005)
and Brock & Hasenpusch (2007), except in the cases where
our phylogenetic analysis suggests newly proposed genera to
be non-monophyletic. We have sampled all four major genera
of New Guinean Eurycanthinae, in addition to Dr yococelus
and most genera of New Caledonian tree lobsters. The
phylogenetic tree was rooted using Timema, which has been
shown to be the sister group to the Euphasmatodea in
previous phylogenetic studies ( Whiting et al. 2003;Bradler in
press). Authorities for all taxonomic names are given in the
electronic supplementary material, table 1.
(b)DNA sequence data collection
DNA extractions were performed from muscle tissue using
Aqua pure genomic DNA tissue kit (Bio-Rad, USA)
following the manufacturer’s instructions. We used PCR to
obtain sequence data from two non-contiguous mito-
chondrial and two nuclear genes. The mitochondrial data
included 753 bp from the cytochrome oxidase subunit I
(COI) gene and 689 bp from the cytochrome oxidase subunit
II (COII) gene. The nuclear DNA sequence data included
350 bp from the Histone subunit 3 (H3) gene and 707 bp
from the large subunit rRNA (28S) gene. The COI and COII
genes were amplified using the primers C1-J-2195CTL2-N-
3014 and TL2-J-3034CTK-N-3785, respectively (Simon
et al. 1994). The 28S gene was amplified using the primers
28S-356C28S-1009 from Buckley et al. (2008) and the H3
gene was amplified using the primers H3F (ATGGCTCG
TACCAAGCAGAC) and H3R (ATATCCTTRGGCAT
RATRGTGAC) from Colgan et al. (1998). PCR cycling
conditions were 948C for 1 min, followed by 35–40 cycles
of 948C for 1 min, 53–578C for 1 min and 728C for 1.5 min,
and 1 cycle for 728C for 10 min. DNA products were purified
for sequencing using MinElute 96 UF PCR purification
kit (Qiagen, USA). Purified PCR products were sequenced
using BIGDYE TERMINATOR v. 3.1 cycle sequencing kit
(Applied Biosystems, USA). Cycle sequencing products
were cleaned by 96 well plate ethanol precipitation
and analysed on ABI 3100 Avant Genetic Analyzer
(Applied Biosystems).
(a)(b)(c)(d)(e)(f)(g)
10 mm 10 mm 10 mm 10 mm
Figure 1. Photo composition of different ‘tree lobsters’ compared with a winged, canopy-dwelling stick insect. (a) Male and
(b) female of D. australis,(c) male and (d) female of Canachus alligator,(e) male and ( f) female of Eurycantha horrida, and
(g) male of Phasma gigas.
1056 T. R. Buckley et al. Convergent evolution in tree lobsters
Proc. R. Soc. B (2009)

(c)Alignment, phylogenetic analysis and
molecular dating
The COI and H3 genes were length-invariable and the COII
gene contained only a small region of length variation.
Alignment of this latter gene was achieved using CLUSTALX
(Larkin et al. 2007). The 28S gene was more variable and
contained a number of regions of length variation. The
CLUSTALX alignment was refined using the 28S rRNA
secondary structure information from the Chrysomelidae
(Coleoptera) model (Gillespie et al. 2004). Helices were
identified using this model and used to delimit the boundary
of regions for exclusion (typically unpaired regions).
We estimated phylogenetic relationships using BEAST
v. 1.4.8 (Drummond & Rambaut 2007) under an uncorre-
lated lognormal model (UCLD; Drummond et al. 2006). We
partitioned the data into codon positions for the mito-
chondrial DNA and assigned the H3 and 28S sites to a
different partition each to yield a five-partition model. For
each partition we used the AIC to select the best fit model as
implemented in PAUP

v. 4.0.b10 (Swofford 1998) and
MODELTEST v. 3.6 (Posada & Crandall 1998). We used a Yule
prior for tree shape and exponential priors for the substitution
model and relaxed clock parameters. Priors for the model
parameters were mZ100.0, GTR rate parametersZ100.0,
transition/transversion rate ratioZ100.0, partition rate multi-
pliersZ100, ashape parameterZ1.0, UCLD meanZ1.0,
UCLD standard deviationZ1.0, Yule birth rateZ1.0, mean
substitution rateZ1.0, coefficient of variationZ1.0 and
covarianceZ1.0. We iteratively optimized the Markov chain
Monte Carlo (MCMC) operators by performing short runs
(10!10
6
cycles) and then adjusting the operators as
suggested by BEAST and gradually increasing the run length.
When the MCMC operators were set at optimal levels, as
indicated by the BEAST output, we ran five r uns at 40!10
6
cycles, sampling every 1000th step in the chain, and used
TRACER (Drummond & Rambaut 2007) to monitor conver-
gence, select the burn-in and calculate effective sample sizes.
All runs that were consistent with convergence were
concatenated and used to estimate the posterior distributions
of topology and divergence time.
Phasmatodea fossils are exceedingly rare and currently
only a few fossils can be unambiguously placed on an extant
lineage (Wedmann et al. 2007). We used the occurrence of
fossil euphasmatodean eggs in mid-Cretaceous Burmese
amber (Rasnitsyn & Ross 2000) to place a prior distribution
on the age of the root. These fossils have been dated to
between 95 and 110 Myr ago (Grimaldi & Engel 2005),
therefore we assumed that the divergence between Timema
and the Euphasmatodea occurred more than 95 Myr ago. We
also used the presence of a leaf insect fossil dated at 47 Myr
ago (Wedmann et al. 2007) to set the minimum age on the
divergence of the leaf insects from their closest relatives. For
the root and leaf insect calibration points, we used
exponential distributions with a mean of 1.0 and 5.0,
respectively, and offset these distributions by 95 and 47 Myr
ago, respectively. Although these fossil calibrations may
represent substantial underestimates of the age of the lineages
they are placed on, we have no information to constrain the
upper age of those lineages. Therefore, despite the tight prior
distribution assumed, the resulting dates are interpreted as
lower limits.
The monophyly of the Eurycanthinae was tested by
constructing a maximum-likelihood tree in PAUP

v. 4.0
and comparing the likelihood of this tree to that of a tree
where the Eurycanthinae were constrained to be mono-
phyletic. The significance of the likelihood ratio was tested
using the Shimodaira–Hasegawa (S–H) test (Shimodaira &
Hasegawa 1999) as implemented in PAUP

v. 4.0.b10. For
the bootstrap step of the S –H test, we used 10 000 replicates
of the RELL approximation.
3. RESULTS
(a)Phylogenetic relationships among the
tree lobsters
We obtained 2.4 kb of nuclear and mitochondrial
sequence data from 78 euphasmatodean individuals and
1 individual of the out-group Timema. All newly obtained
sequences have been submitted to GenBank under
accession numbers FJ474100–FJ474403. We observed
excellent support (1.0 Bayesian posterior probability) for
the monophyly of Diapheromerinae, Pseudophasmatinae
sensu stricto (excluding Agathemera and the Heteronemia
group, e.g. Spinonemia), Phylliinae, Aschiphasmatinae,
Necrosciinae, Lonchodinae, Cladomorphinae, Obrimini
and Lanceocercata in agreement with the previous
phylogenetic studies (figure 2;Whiting et al. 2003;Bradler
in press). In general, there are very few relationships
between traditional subfamilies that are well supported,
the exceptions being EurycanthomorphaCLonchodinae
(83% posterior probability), Bacillinae (Xylica)CAschi-
phasmatinae (98% posterior probability) and Palophinae
(Bactrododema)CCladomorphinae (73% posterior prob-
ability), the first also recovered in a previous molecular
phylogenetic study (Whiting et al. 2003). We find support
for neither the monophyly of Phasmatinae, nor its
subgroup Pharnaciini (the latter represented by Pharnacia
and the MacrophasmaCPhasmotaenia clade). The UCLD
model inferred the root of the tree to lie on the branch
between Timema and the Euphasmatodea with a posterior
probability of 1.0, as also supported by previous molecular
(Whiting et al. 2003)andmorphological(Bradler
in press) data.
The Lanceocercata contain a wide array of Australasian
phasmids conventionally thought to be unrelated to one
another, comprising species of Tropidoderinae, Xeroder-
inae, Platycraninae, Phasmatini, Acanthoxylini and
Pachymorphini (Bradler 2001,in press). Within the
Lanceocercata, Dimorphodes (subfamily Xeroderinae)
constitutes the sister group to all other members (100%
posterior probability), which include the remaining
Xeroderinae and the polyphyletic Tropidoderinae, and
Phasmatini as previously suggested ( Whiting et al. 2003).
Perhaps the most surprising result is that the subfamily
Eurycanthinae, containing the tree lobsters, are also
polyphyletic, forming five separate lineages. Members of
the virtual tree lobsters are found in three unrelated
regions of the tree (grey boxes in figure 2), two of them
(TrapezaspisCCanachusCMicrocanachus and Dryococelus)
nested within Lanceocercata. Using an S–H test, we were
able to reject monophyly of the Eurycanthinae with a
P-value of 0.00001. The members of the Eurycanthinae
outside the Lanceocercata clade comprise Eurycantha and
related genera from New Guinea, the Eurycanthomorpha
sensu Bradler (2002,in press). Within Lanceocercata, the
New Caledonian eurycanthines form a clade with various
other taxa from New Caledonia and New Zealand (100%
posterior probability), including the enigmatic Cnipsus,
Convergent evolution in tree lobsters T. R. Buckley et al. 1057
Proc. R. Soc. B (2009)

Megacrania phelaus
Timema
Extatosoma tiaratum bufonium
Achrioptera punctipes
Rhaphiderus scabrosus
Tropidoderus childrenii
Phyllium giganteum
Pachymorpha sp.
Abrosoma festinatum
Phasma gigas
Vetilia thoon
Cnipsus rachis
Pharnacia ponderosa
Anchiale sp.
Erinaceophasma vepres
Carlius fecundus
Xylica oedematosa
Clonaria conformans
Macrophasma biroi
Agamemnon cornutus
Heteropteryx dilatata
Eurycnema goliath
Pseudosermyle phalangiphora
Zehntneria mystica
Canachus sp.
Clitarchus hookeri
Thaumatobactron guentheri
Chondrostethus woodfordi
Pterobrimus depressus
Extatosoma tiaratum tiaratum
Canachus alligator
Peruphasma schultei
Xeroderus sp.
Ctenomorpha marginipennis
Acanthoxyla geisovii
Phyllium celebicum
Rhynchacris ornata
Dimorphodes sp.
Megacrania batesii
Orxines
Spinonemia chilensis
Monandroptera acanthomera (2)
Asprenas impennis
Labidiophasma rouxi
Agathemera sp.
Dryococelus australis
Neopromachus doreyanus
Pseudophasma velutinum
Diapheromera femorata
Sipyloidea sipylus
Trapezaspis sp.
Sceptrophasma hispidulum
Eurycnema osiris
Monandroptera acanthomera (1)
Necrosciinae sp.
Dinophasma saginatum
Hyrtacus sp.
Chitoniscus brachysoma
Acrophylla wuelfingi
Malandania pulchra
Phyllium siccifolium
Bacteria ferula
Trachyaretaon brueckneri
Bactrododema sp.
Eurycantha calcarata
Microcanachus matileorum
Phasmotaenia sp.
Leosthenes sp.
Podacanthus wilkinsoni
Dimorphodes mancus
Anchiale briareus
Acrophylla titan
Ramulus thaii
Anisomorpha buprestoides
Pterinoxylus crassus
Chitoniscus feejeeanus
Carausius morosus
Heteropteryginae
Pachymorphinae
Phasmatinae
Tropidoderinae
Xeroderinae
Eurycanthinae
non-supported subfamilies:
Euphasmatodea
Eurycanthomorpha
Diapheromerinae
Cladomorphinae
Lonchodinae
Necrosciinae
Pseudophasmatinae
sensu stricto
Aschiphasmatinae
Obrimini
Phylliinae
90 80 70 60 50 40 30 20 10 0
millions of years ago
Oxyartes labellatus
Phaenopharos khaoyaiensis
Lanceocercata
New Caledonia+
New Zealand clade
98 100
100
100
100
100 100
100
100
100
100
100
100
100
100
100
100
100 100
100
100
100
100
100
100
100
100
100
83 100
100
83 98
89 90
9390
71
59
74
67
55
56
78 57
68
96
84
89
73
97
95
96
89
71
62
76
88
71 99
93
99
77
71
74 96
99
Figure 2. Bayesian phylogenetic tree showing relationships among euphasmatodean taxa and placement of the ‘tree lobster’
ecomorphs. Branch lengths are drawn proportional to time and values above branches are Bayesian posterior probabilities. Non-
monophyletic subfamilies are indicated by coloured branches according to the inset key. The tree lobster taxa are highlighted
grey. Monophyletic taxa are indicated by vertical bars.
1058 T. R. Buckley et al. Convergent evolution in tree lobsters
Proc. R. Soc. B (2009)

which has been assigned to different subfamilies in the past
(Redtenbacher 1908;Gu
¨nther 1953;Zompro 2001). The
Lord Howe Island tree lobster Dryococelus appears to be
unrelated to the New Guinean and New Caledonian
Eurycanthinae, forming a rather isolated lineage, but
related to the widespread Australian genus Eurycnema
(76% posterior probability).
Using the two calibration points discussed above,
we obtained a lower limit on the age of the extant
Euphasmatodea radiation of 51.9 Myr ago (95% posterior
intervals 47.0–58.7). Most of the subfamilies diverged
over a period of ca 20 Myr ago, indicating that the
Euphasmatodea underwent a rapid radiation.
(b)Analysis of morphological characters
Our findings contradict the view that Dryococelus is
closely related to the New Guinean or New Caledonian
tree lobsters as has been suggested for almost 150 years
by all previous authors ( Westwood 1859;Redtenbacher
1908;Gurney 1947;Gu
¨nther 1953;Key 1991;Zompro
2001;Hennemann & Conle 2006;Brock & Hasenpusch
2007;Honan 2008). The results of our phylogenetic
analysis demand a reassessment of the morphological
evidence. One putative synapomorphic character of
Dryococelus and the New Guinean Eurycantha and
Thaumatobactron are the conspicuously enlarged and
strongly armed hind legs of the males. Both Dryococelus
and Eurycantha, for example, have a large defensive spine
on ventral surface of the hind femora (dsp in figure 3f,g).
However, closer inspection of this armature shows that it
appears to be non-homologous between these genera. In
Dryococelus, the prominent ventral spine is formed by the
ventroexternal carina (vec in figure 3 f), whereas in
Eurycantha and Thaumatobactron a similarly prominent
spine is produced by the midventral carina (mvc in
figure 3g;Gurney 1947) and in Canachus this spine
is absent.
In the neck region of all Eurycanthomorpha, e.g.
Thaumatobactron, the cervix is protected by a prominent
sclerite, the gula (shaded blue in figure 3a), whereas
Canachus exhibits no gular sclerotization whatsoever
(figure 3b), and Dryococelus specimens possess small
gular sclerites in the otherwise membranous cervical
region (figure 3c).
mxp
lp
gu
lac
cvx
prb
sbm
(a)
(f)(h)(g)(i)(j)
(b)(c)(d)
(e)
mxp
lp
lac
cvx
prb
sbm
mxp
lp
lac
cvx
prb
sbm
glr
2mm
t10
tho
t10
tho
t9
par
epi
cer
gon
gpl
gp8
gp9
t10
par
epi
cer
gp8
gp9
gpl
t9
t10
ros
par
epi
cer
gon
gp8
t9
t10
ros
gp9
10mm
fe
mvc
vec
fe vec
mvc
dsp dsp
3mm 3mm
10mm
3mm
2mm2mm
1mm
1mm
gon
Figure 3. Morphological details of D. australis compared with key taxa of the remaining ‘tree lobster’ clades. Ventral view of the
male head region of (a)Thaumatobactron guentheri,(b)C. alligator and (c)D. australis. Male clasper (tergum 10) in posterior view
of (d)D. australis and (e)C. alligator. Ventral view of left hind leg of ( f) male D. australis and (g) male Eurycantha calcarata.
Ventral view of female genitalia (sternum 8 removed) of ( h)D. australis,(i)C. alligator and ( j)Eur ycantha insularis. cer, cercus;
cvx, cervix; dsp, defensive spine; epi, epiproct; fe, femur; glr, gularia; gon, gonangulum; gp8–9, gonapophysis 8–9; gpl,
gonoplac; gu, gula; lac, laterocervicalia; lp, labial palpus; mvc, midventral carina; mxp, maxillar palpus; par, paraproct; prb,
probasisternite; ros, rostrum; sbm, submentum; t9–10, tergum 9–10; tho, thorn pad; vec, ventroexternal carina. The arrows in
(a–c) indicate the position of the posterior tentorial pits. Gular sclerotization shaded blue, gonoplacs shaded red.
Convergent evolution in tree lobsters T. R. Buckley et al. 1059
Proc. R. Soc. B (2009)

In the female ovipositor of Dryococelus, all three pairs of
valves are well developed, with the gonoplac forming the
largest sheath (shaded red in figure 3h), which appears to
represent the primitive condition among neopteran insects
(Kristensen 1975) as well as in Phasmatodea ( Tilgner
et al. 1999). The ovipositor of Canachus exhibits only
remnants of the gonoplac (figure 3i) and in Eurycantha the
gonoplac is completely reduced (figure 3j). In females of
Eurycantha and Canachus, the abdominal tergum 10 is
elongated forming a rostrum as part of a secondary
ovipositor for depositing eggs in soil and other substrates
(figure 3i,j). A secondary ovipositor is absent in
Dryococelus, traditionally interpreted as the result of a
secondary reduction (Zompro 2001). Our reconstruction,
however, suggests that a secondary ovipositor was not
present in the ancestral lineage of Dryococelus, suggesting
its absence in Dryococelus to constitute the primary
condition. Thus, the female genitalia are quite different
among each of the three tree lobster lineages, and there
appear to be no obvious synapomorphies to link
Dryococelus and the New Guinean and New Caledonian
tree lobsters. This variation in female terminalia is
reflected in the differences in the genital armature of the
male claspers, where Dryococelus and Canachus resemble
typical Lanceocercata (tho in figure 3d,e;cf.Bradler
in press). Finally, the New Caledonian eurycanthines
(Canachus and the related Asprenas) have small wing
rudiments, whereas the Eurycanthomorpha and Dryoco-
celus are fully apterous.
In summary, Dryococelus lacks all of the apomorphic
characters that define the New Guinean Eurycantho-
morpha, including the presence of a gula, reduced
gonoplacs in the female ovipositor and the presence of
a secondary ovipositor for egg deposition (Bradler 2002,
in press). These relatively inconspicuous but significant
anatomical differences between the different tree lobster
lineages mirror their separate phylogenetic placement.
Nevertheless, the overall strong similarity in general body
form of the different tree lobsters led to their improper and
hitherto unquestioned classification.
4. DISCUSSION
(a)Diversification and convergent evolution in
stick insects
Similarities between species can arise in two fundamen-
tally different ways. Either each species has retained a
comparable trait from their common ancestor, or each has
acquired it independently (Hall 1994). Although the first
possibility might seem far more likely, convergence is in
fact a common phenomenon ( Morris 2003), often found
as a consequence of adaptive radiations in separate
evolutionary lineages (Givnish 1997). Our data strongly
suggest that the tree lobster body form evolved indepen-
dently on the three different landmasses of New Guinea,
New Caledonia and Lord Howe Island. Each lineage of
these ground-dwelling ecomorphs have probably des-
cended from arboreal ancestors as they are all either
nested within arboreal clades (Dryococelus and Canachus)
or sister-group to them ( New Guinean tree lobsters).
Their overall uniformity in body form and behaviour is
probably the product of similar selective pressure associ-
ated with adaptations to ground-dwelling life. Individuals
of Dryococelus and Eurycantha congregate in large
numbers and close spatial proximity in tree hollows and
cavities during the day ( Lea 1916;Gurney 1947;Bedford
1976). Females of all tree lobster genera deposit their eggs
into the soil (Hsiung 1987;Pain 2006;Honan 2008;
personal observations), with females of Canachus and
Eurycantha even using a similarly developed secondary
ovipositor (Bradler 2002;figure 3i,j). Individuals of all
three clades exhibit a robust habitus with dorsoventrally
flattened body and sturdy legs. The greatly enlarged and
armed hind legs of some males probably evolved as a
response to ground-hunting predators and might also be
used against other males ( Lea 1916;Gurney 1947;
Bedford 1976;Hsiung 1987;Honan 2008), although if
this is the case this defence mechanism did not save
Dryococelus from rats (the rats were apparently eating the
nymphs, not adults).
Among Phasmatodea the Lanceocercata exhibit
morphological and ecological parallelisms comparable
with those found between placental mammals and
marsupials (Springer et al. 1997) and between afrotherian
and laurasiatherian mammals (Madsen et al.2001).
Examples of extensive and multiple convergences have
also been demonstrated between lineages of African
cichlid fishes (Meyer et al. 1990;Brakefield 2006), lizards
on Caribbean islands (Losos et al. 1998) and between
several bird families (Fain & Houde 2004). Such
phenotypic similarity between unrelated species is
probably generated by extrinsic selective pressure, as
well as by intrinsic factors such as shared trajectories in the
underlying developmental architecture (Brakefield 2006),
which might provide constraints on the direction of stick
insect evolution. Our results indicate that the Australasian
Lanceocercata and the remaining Euphasmatodea under-
went parallel adaptive radiations that resulted, in addition
to the tree lobsters, in further astounding examples of
convergence. The Lanceocercata also comprise giant-
winged stick insects of the canopy, such as the Australian
Acrophylla, exceeding 26 cm in body length, which is
paralleled by the equally large, morphologically and
ecologically similar African Bactrododema ( Palophinae).
The leaf-imitating forms include Malandania and Tropi-
doderus in Lanceocercata and true leaf insects in the
Phylliinae. Gracile flyers are also found within Lanceo-
cercata (e.g. Carlius), which are strikingly similar to the
forms from the Necrosciinae (e.g. Sipyloidea). In addition,
the small wingless Australian Lanceocercata with exceed-
ingly short antennae (e.g. Pachymorpha)arehighly
reminiscent of the Afro-Oriental Gratidiini (e.g. Sceptro-
phasma,Clonaria,Gratidia). Examples among diminutive
spiny trunk-dwellers include Cnipsus in Lanceocercata and
Neopromachus in Eurycanthomorpha.
Our data show that the Lanceocercata radiated not only
in Australia and New Guinea but also from as far west as
the Mascarene Archipelago in the Indian Ocean and as far
east as New Zealand and New Caledonia. Given the low
dispersal abilities of phasmatodeans overseas ( Nakata
1961), tectonic movements might have been of major
importance in shaping their historical distribution.
Because we have no information from which to derive
the upper limits on any of the divergences times within the
Phasmatodea, all our reported dates must be interpreted
as minimum estimates, and future discoveries of fossil
phasmatodeans pertaining to extant crown groups could
result in significantly older age estimates of their most
1060 T. R. Buckley et al. Convergent evolution in tree lobsters
Proc. R. Soc. B (2009)

recent common ancestor. Our divergence time estimates
using a Bayesian relaxed clock (Drummond et al. 2006)
suggest that the Lanceocercata began to diversify at least
32 Myr ago (29.4–37.5 Myr ago), in the Oligocene. The
radiation of the subfamilies and other major clades
occurred over a period of 20 million years (and much
less if the Diapheromerinae and Phylliinae are excluded)
and may represent a classic case of rapid ancient radiation.
Our divergence time estimates suggest that Dryococelus
shared a most recent common ancestor with its closest
Australian relative, the genus Eurycnema, at least 22 Myr
ago (15.9–26.2 Myr ago), which contrasts with the
emergence of Lord Howe Island 6.4–6.9 Myr ago as the
result of volcanic activity along the Lord Howe Rise
(McDougall et al. 1981). Unless future sampling of other
Lanceocercata taxa reveals more closely related extant
relatives then Dryococelus must have existed elsewhere
prior to the formation of Lord Howe Island, possibly to
the north on the now-submerged seamounts known as the
Lord Howe seamount chain, or to the northwest along
the submerged Tasmantid Guyots ( McDougall et al.
1981). Interestingly, the oldest of the submerged islands
in the Lord Howe seamount chain, Nova Bank, is roughly
estimated to be 23 Myr old (McDougall et al. 1981),
which accords well with our minimum divergence estimate
of Dryococelus from its mainland Australian relatives. The
Lord Howe Island tree lobster may have evolved on now-
drowned islands far to the north of Lord Howe and
progressively dispersed down the island chain, leaving its
ancestral populations to become extinct as their islands
eroded away.
(b)Conservation genetics of D. australis
Numerous authors have suggested that the conservation
of phylogenetic diversity is preferable over the protection
of pure species richness, taking into account that
different species can vary drastically in their evolutionary
isolation and heritage ( Va n e - W r i g h t et al. 1991;
Faith 1992;Mooers 2007). The isolated phylogenetic
position and great age of Dryococelus among the Lanceo-
cercata highlights the importance of its conservation
with respect to phasmatodean diversity. Considerable
effort has already been undertaken to conserve the small
remaining population of Dryococelus (Honan 2008). The
results of this study indicate that this investment is
decidedly warranted.
Specimens and assistance with fieldwork were provided by
H. Blafard, M. Brinkert, P. Brock, S. Brown, J.-J. Cassan,
S. Cazeres, D. Clark, J. Colville, P. Cranston, N. Evenhuis,
K. Hill, R. Hoare, M. Humphrey, T. Jewell, B. Kneubu
¨hler,
R. Leschen, D. Marshall, J. Midgley, C. Mille
´, P. Miller,
G. Monteith, M. Moulds, D. Otte, D. Paulaud, K. Rabaey,
R. Simoens, K. Vandervennet and K. Will. Thanks
to T. Reischig and P. Schwendinger for taking the photos.
U. Aspo¨ ck, S. Randolf and P. Schwendinger loaned speci-
mens. Permits for New Caledonia were issued by D. Paulaud
and J.-J. Cassan with assistance from C. Mille
´. R. Leschen,
A. Stumpner and two anonymous reviewers provided
comments on the manuscript. Funding was provided by
the National Geographic Society (7906-05), the Royal
Society of New Zealand Marsden Fund ( LCR302), the
Foundation for Research, Science and Technology through
the Defining New Zealand’s Land Bioti OBI and an Ernst
Mayr Travel Grant in Animal Systematics.
REFERENCES
Bedford, G. O. 1976 Defensive behaviour of the New Guinea
stick insect Eurycantha ( Phasmatodea, Phasmatidae,
Eurycanthinae). Proc. Linn. Soc. NSW 100, 218–222.
Bradler, S. 2001 The Australian stick insects, a monophyletic
group within the Phasmatodea? Zoology 104(Suppl. IV
(DZG 94.1)), 69.
Bradler, S. 2002 Evolution of secondary ovipositors in
stick insects (Insecta: Phasmatodea). Zoology 105(Suppl. V
(DZG 95.1)), 60.
Bradler, S. In press. Phylogeny of the stick and leaf insects
(Insecta: Phasmatodea). Species Phylogenet. Evol.
Brakefield, P. M. 2006 Evo–devo and constraints on
selection. Trends Ecol. Evol. 21, 362–368. (doi:10.1016/
j.tree.2006.05.001)
Brock, P. D. & Hasenpusch, J. 2007 Studies on the Australian
stick insects (Phasmida), including a checklist of species
and bibliography. Zootaxa 1570, 1–81.
Buckley, T. R., Attanayake, D., Park, D.-C., Ravindran, S.,
Jewell, T. R. & Normark, B. B. 2008 Investigating
hybridization in the parthenogenetic New Zealand stick
insect Acanthoxyla (Phasmatodea) using single-copy
nuclear loci. Mol. Phylogenet. Evol. 46, 335–349. (doi:10.
1016/j.ympev.2008.02.016)
Colgan, D. J., McLauchlan, A., Wilson, G. D. F., Livingston,
S. P., Edgecombe, G. D., Macaranas, J., Cassis, G. &
Gray, M. R. 1998 Histone H3 and U2 snRNA DNA
sequences and arthropod molecular evolution. Aust.
J. Zool. 46, 419–437. (doi:10.1071/ZO98048)
Drummond, A. J. & Rambaut, A. 2007 BEAST: Bayesian
evolutionary analysis by sampling trees. BMC Evol. Biol. 7,
214. (doi:10.1186/1471-2148-7-214)
Drummond, A. J., Ho, S. Y. W., Phillips, M. J. & Rambaut, A.
2006 Relaxed phylogenetics and dating with confidence.
PLoS. Biol. 4, e88. (doi:10.1371/journal.pbio.0040088)
Fain, M. G. & Houde, P. 2004 Parallel radiations in the
primary clades of birds. Evolution 58, 2558–2573. (doi:10.
1554/04-235)
Faith, D. P. 1992 Conservation evaluation and phylogenetic
diversity. Biol. Conserv. 61, 1–10. (doi:10.1016/0006-
3207(92)91201-3)
Gillespie, J., Cannone, J., Gutell, R. & Cognato, A. 2004
A secondary structure model of the 28S rRNA expansion
segments D2 and D3 from rootworms and related leaf beetles
(Coleoptera: Chrysomelidae; Galerucinae). Insect Mol. Biol.
13,495–518.(doi:10.1111/j.0962-1075.2004.00509.x)
Givnish, T. J. 1997 Adaptive radiation and molecular
systematics: issues and approaches. In Molecular evolution
and adaptive radiation (eds T. J. Givnish & K. J. Sytsma),
pp. 1–54. Cambridge, UK: Cambridge University Press.
Grimaldi, D. A. & Engel, M. S. 2005 Evolution of the insects.
Cambridge, UK: Cambridge University Press.
Gu
¨nther, K. 1953 U
¨ber die taxonomische Gliederung und
geographische Verbreitung der Insektenordnung der
Phasmatodea. Beitr. Entomol. 3, 541–563.
Gurney, A. B. 1947 Notes on some remarkable Australasian
walkingsticks, including a synopsis of the genus Extato-
soma (Orthoptera: Phasmatidea). Ann. Entomol. Soc. Am.
40, 373–396.
Hall, B. K. 1994 Homology: the hierarchical basis of comparative
biology. San Diego, CA: Academic Press.
Hennemann, F. & Conle, O. V. 2006 Papuacocelus papuanus
n. gen., n. sp.—a new Eurycanthinae from Papua
New Guinea, with notes on the genus Dryococelus Gurney,
1947 and description of the egg ( Phasmatodea: Phasmatidae:
Eurycanthinae). Zootaxa 1375, 31– 49.
Honan, P. 2008 Notes on the biology, captive management
and conservation status of the Lord Howe Island Stick
Insect (Dryococelus australis) ( Phasmatodea). J. Insect
Conserv. 12, 399–413. (doi:10.1007/s10841-008-9162-5)
Convergent evolution in tree lobsters T. R. Buckley et al. 1061
Proc. R. Soc. B (2009)

Hsiung, C. 1987 Aspects of the biology of the of the
Melanesian stick-insect Eurycantha calcarata Lucas
(Cheleutoptera: Phasmatidae). J. Nat. Hist. 21,
1241–1258. (doi:10.1080/00222938700770761)
Key, K. H. L. 1991 Phasmatodea (stick-insects). In The
insects of Australia, vol. 1 (ed. CSIRO), pp. 394–404.
Ithaca, NY: Cornell University Press.
Klug, R. & Bradler, S. 2006 The pregenital abdominal
musculature in phasmids and its implications for the
basal phylogeny of Phasmatodea (Insecta: Polyneoptera).
Org. Divers. Evol. 6, 171–184. (doi:10.1016/j.ode.2005.
08.004)
Kristensen, N. P. 1975 The phylogeny of hexapod ‘orders’.
A critical review of recent accounts. Zool. Syst. Evol.
Forsch. 13, 1–44.
Larkin, M. A. et al. 2007 CLUSTALW and CLUSTALX v. 2.0.
Bioinformatics 23, 2947–2948. (doi:10.1093/bioinfor-
matics/btm404)
Lea, A. M. 1916 Notes on the Lord Howe phasma, and on
an associated longicorn beetle. Proc. R. Soc. S. Aust. 40,
145–147.
Losos, J. B., Jackman, T. R., Larson, A., de Queiroz, K. &
Rodrı
´guez-Schettino, L. 1998 Contingency and determin-
ism in replicated adaptive radiations of island lizards.
Science 279, 2115–2118. (doi:10.1126/science.279.5359.
2115)
Madsen, O. et al. 2001 Parallel adaptive radiations in two
major clades of placental mammals. Nature 409, 610–614.
(doi:10.1038/35054544)
McDougall, I., Embleton, B. J. J. & Stone, D. B. 1981 Origin
and evolution of Lord Howe Island, southwest Pacific
Ocean. J. Geol. Soc. Aust. 28, 155–176. (doi:10.1080/
00167618108729154)
Meyer, A., Kocher, T. D., Basasibwaki, P. & Wilson, A. C.
1990 Monophyletic origin of Lake Victoria cichlid fishes
suggested by mitochondrial DNA sequences. Nature 347,
550–553. (doi:10.1038/347550a0)
Mooers, A. Ø. 2007 The diversity of biodiversity. Nature 445,
717–718. (doi:10.1038/445717a)
Morris, S. C. 2003 Life’s solution: inevitable humans in a lonely
universe. Cambridge, UK: Cambridge University Press.
Nakata, S. 1961 Some notes on the occurrence of
Phasmatodea in Oceania. Pacific Insects Monogr. 2,
107–121.
Otte, D. & Brock, P. D. 2005 Phasmida species file: a catalog to
the stick insects of the world. Philadelphia, PA: The
Orthopterists Society.
Pain, S. 2006 Return of the giants. New Sci. 191, 42–45.
(doi:10.1016/S0262-4079(06)60113-9)
Paramonov, S. J. 1963 Lord Howe Island, a riddle of the
Pacific, part III. Pac. Sci. 17, 361–373.
Posada, D. & Crandall, K. A. 1998 MODELTEST: testing the
model of DNA substitution. Bioinformatics 14, 817–818.
(doi:10.1093/bioinformatics/14.9.817)
Priddel, D., Carlile, N., Humphrey, M., Fellenberg, S. &
Hiscox, D. 2003 Rediscovery of the ‘extinct’ Lord Howe
Island stick-insect (Dryococelus australis (Montrouzier))
(Phasmatodea) and recommendations for its conserva-
tion. Biodivers. Conserv. 12, 1391–1403. (doi:10.1023/
A:1023625710011)
Rasnitsyn, A. P. & Ross, A. J. 2000 A preliminary list of
arthropod families present in the Burmese amber
collection at The Natural History Museum, London.
Bull. Nat. Hist. Mus. Geol. Ser. 56, 21–24.
Redtenbacher, J. 1908 Die Insektenfamilie der Phasmiden III.
Phasmidae Anareolatae. Leipzig, Germany: Wilhelm
Engelmann.
Rentz, D. C. F. 1996 Grasshopper countr y. Sydney, Australia:
University of NSW Press.
Robertson, H. 2006 Sticking in there. Curr. Biol. 16,
R781–R782. (doi:10.1016/j.cub.2006.08.062)
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., Crespi, B., Liu, H. & Flook, P. K. 1994
Evolution, weighting, and phylogenetic utility of mito-
chondrial gene sequences and a compilation of conserved
PCR primers. Ann. Entomol. Soc. Am. 87, 651–701.
Springer, M. S., Kirsch, J. A. W. & Case, J. A. 1997
The chronicle of marsupial evolution. In Molecular
evolution and adaptive radiation (eds T. J. Givnish &
K. J. Sytsma), pp. 129–161. Cambridge, UK: Cambridge
University Press.
Swofford, D. L. 1998 PAUP

:phylogenetic analysis using
parsimony (

and other methods), v. 4. Sunderland, MA:
Sinauer Associates.
Tilgner, E. H., Kiselyova, T. G. & McHugh, J. V. 1999
A morphological study of Timema cristinae Vickery with
implications for the phylogenetics of Phasmida. Deut.
Entomol. Z. 46, 149–162.
Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991
What to protect?—Systematics and the agony of choice.
Biol. Conserv. 55, 235–254. (doi:10.1016/0006-3207(91)
90030-D)
Wedmann, S., Bradler, S. & Rust, J. 2007 The first fossil leaf
insect: 47 million years of specialized cryptic morphology
and behavior. Proc. Natl Acad. Sci. USA 104, 565–569.
(doi:10.1073/pnas.0606937104)
Westwood, J. O. 1859 Catalogue of the Orthopterous insects in
the collection of the British Museum. Part I: Phasmidae.
London, UK: Taylor & Francis.
Whiting, M. F., Bradler, S. & Maxwell, T. 2003 Loss and
recovery of wings in stick insects. Nature 421, 264–267.
(doi:10.1038/nature01313)
Zompro, O. 2001 A review of Eurycanthinae: Eurycanthini,
with a key to genera, notes on the subfamily and
designation of type species. Phasmid Stud. 10, 19–23.
1062 T. R. Buckley et al. Convergent evolution in tree lobsters
Proc. R. Soc. B (2009)